SVN # Date Topics Bug # Commit message
1Web Apps 1.0 initial checkin of working directory
2Update Mahai's name per request
3Updating history entry at top of document
4Update of toDataURL() security requirements.
5response to dbaron's comments entitled 'WA1: Conformance requirements'
6Update <canvas> to be more normative; update patterns to start on coordinate system origin (should match mozilla/safari, and be implementable, unlike before); introduce a clause allowing hardware limitations to guide implementations; use that clause to allow arbitrary nesting depth limitations in html parser. Changes made in parallel, in response to requests from vlad (in bugzilla) and mrbkap (on IRC).
7Typo fixes, wording simplifications, defined error handling for lone LIs, defined handling for value attribute better, added some TODO notes from Henri regarding content model rules
8updated status section and misc conf criteria
9Updates to the Conformance section based on dbaron's feedback
10Need work on this section (encoding detection). Added some issues and XXX comments on the matter.
11Text from Henri for conf checkers and parse errors
12Minor fixes to the parser part.
13Minor typo fixes and notes.
14Bunch of minor fixes in the parsing section
15<meter> and <progress> (sorry for the big checkin, had svn issues)
16A few changes to the parser based in implementation feedback
17Changes in response to bz's implementation feedback for the alternate stylesheet API. Also, changed an occurance of 'marker' to 'bookmark' in one of the algorithms in the parser because the word 'marker' is used for something else already.
18Removed '(optional)' from all attribute definitions while we work out what we want to do there. Minor clarification in the ASSDOM spec.
19More notes on <script defer>, minor wordsmithing in the parsing spec.
20Navigating-across-documents section placeholder, first draft of Custom-protocol-and-content-handlers section, fix a bunch of typos
21misclosed tag
22More security concerns with the new content handler idea
23Storage APIs version 2.0: Now with SSL Support and Only Supporting Lame Text Strings
24Expand the intro to 'Client-side session and persistent storage' a little bit, reword some of the text in the handlers section.
25Fix minor issues raised by jst and biesi in the content handler and storage sections
27typo markup fix; make abortion legal for browsers
28Make alternate stylesheet set names case-sensitive.
29Big changes to the way we do <title> and <script> parsing, and how we handle EOF conditions.
30minor parser changes
31More changes to the parsing model
32Clear up the handling of EOF tokens in body parsing (pass 1)
33added notes to focus section; addressed biesi's comments in the content handler section; fixed a bug in the parser section (<col> elements); added a note about window.getAttention()
34One more EOF handling case
35Remove XMLHttpRequest in favour of the W3C spec for the same API.
36Fix an infinite loop when hitting </td> or </th> in a document, oops.
37Added note to calm implementors that can't work out why whitespace ends up in the wrong place
38Fix ambiguous text in the select insertion mode
39Frameset parsing fixes
40Simplify the AAA while making it actually work...
41Remove some lies! Damn lies! which wasted an hour of my time.
42Address some of biesi's comments for content handling; fix whitespace and comment handling in the parser
43Change the way we classify elements (for no apparent reason as it turns out); fix the phrasing element end tag algorithm (<span><table></span> was broken)
44add isPointInPath()
45Clear up that the content handler thing is only for _potential_ handlers.
46isindex anchor
47rename marked section to markup declaration
48Minor fix to informative text regarding toDataURL
49Some work on script defer/async
50getImageData/drawImageData proposal
51Make <!-- -- --> a parse error
52shapes in canvas - minor edits
53fix a minor typo in the canvas rgba() spec
54Canvas changes -- putImageDara; transform(), and minor changes
55Mention case folding in a couple of places
56security of drawimagedata
57mention more about rounding for the imagedata apis
58Few minor canvas edits
59Minor changes to Audio (repeat->play) and getItem in Storage.
60Sample implementation of custom handlers to explain things better
61Make the description slightly more explicit
62Move paragraphs around so that people who don't read the spec before complaining (e.g. sicking) can get the clues earlier.
63Minor work on scripting, fix example bug in storage section.
64Script defer/async draft work
65More work on <script>
66minor async edits
67Big editorial changes of the introduction section, moving stuff into other parts of the spec as appropriate, adding a dependencies section; also did s/DocumentWindow/DocumentUI/.
68Restructuring to fit the new W3C world; part 1 (WindowHTML in, DocumentUI out); also a few fixes here and there e.g. in the parser.
69Unexpected EOF should report a parse error. Whe should abort the steps in the AAA even if there is no formatting element at all.
70Fix a bug that would make the stack and the list get out of sync...
71IDs for Anne
72<b><p><b></p></b> didn't work.
73Move some sections around (wip)
74More section moves
75Fx the onerror section and move the other browsing context sections out into separate sections
76Move disclosure element; update acks; add parse errors for EOF in title elements, etc
77Define a <details> element to act as a disclosure triangle widget.
78Make all EOF processing go through the data state.
79Make contenteditable case-insensitive
80minor changes to parser spec
81More security problems with storage areas
82Couple more examples in the security section for storage.
83Make option/optgroup not create nodes in body insertion mode
84Second draft of <details>
85More <details> notes.
86Clarify the bogus comment state rules
87make attribute duplication a parse error
88typo fixes
89Fix </html> handling
90AAA fixes: don't bother checking if the node was no longer in the tree now, remove nearest block concept
91WIP for d.w(); make <hx> start tags close <hx> elements
92Oops... (AAA impossible situations: some elements aren't in the list, but aren't blocks...)
93Typos s/let/if/
94Need to also remove the non-formatting nodes, or things go all skewiff
95<wbr> is an empty element.
96How To Read This Specification and other minor edits
97More minor changes to the intro section
98Minor id change s/intro/introduction
99Fix issue in parser with </select>, <a>-implying-</a>; update security section in gen'ed page
100Fix <a><table><a></table>
101Minor text additions for the navigation processing model
102Next pass at doing document.write() plumbing... (more to come)
103it's a switch, not a normal dl
104Markup error
105Consistent markup
106Make the parser set the insertion point correctly and generally make the next step along the way of defining document.write()
107markup fixup
108Clarifications to the insertion point handling. More cross-refs across scripting sections.
109Notes for document.write()
110<details> draft 1 done; removing TBW tag
111More IDs for cross-refs
112Example for entity parsing.
113More fragids.
114What's a duplicate attribute? Exact same name.
115Change the way that paths work to be compatible with safari
116Set the stage for datagrid changes
117more datagrid notes
118Add progress.position for datagrid purposes, tweak some definitions for progress and make more cross-references
119Data Grid Draft 2
120ID for datagrid
121drag and drop baby
122minor things missed when defining drag-and-drop (e.g. adding draggable to the HTMLElement interface IDL)
123text field selections
124make insertion modes links, document.write() notes, rearrange script section, minor edits to the tokeniser, placeholder for attribute minimisation handling
126editorial changes to make hyperlinks work and to fix last screwed up checkin
127start of a d.w() section
128make the table slightly more verbose
129Behold, document.write() in HTML
130more clear issue markers
131minor grammar nazi attack
132crossref error
133notes for innerHTML, cache pining, fix typo 'characters tokens', add issue about merging phases and insertion modes
134add placeholder for where mathml-p might go
135define terminology getting/setting. placeholder for innerHTML in HTML. some wording fixes for RFC2119-compliance in the definition of 'reflect'.
136innerHTML getter draft 1
137minor editorial tweaks to the innerHTML getter
138Behold! The goodness that is the innerHTML setter.
139add another example of bogus sillyness in innerHTML
140move rfc2119 reference, and update to new postprocessor
141Oops, noframes was listed twice. Also, 'Create an element node for' wasn't hyperlinking correctly.
142Make <optgroup> and <option> share the same definition, since they're the same.
143<frame> is an empty element. Oops.
144Oops, broke the reset insertion mode algorithm in the <table> case
145HTMLCollection work
146checkpoint: s/Avalon/WPF/; define "tree order" and link to it; define "ignored" and link to it (using old definition); define "collections" and HTMLCollection and HTMLFormControlsCollection and begin on HTMLOptionsCollection; add note for img.src behaviour
147HTMLOptionsCollection definition
148ECMAScript *ODM* binding; add note about .add/.remove on select elements; DOMTokenString part 1; placeholder for definition of space separated token strings.
149Move markers around and make a note more visible
150remove a few 'the the's.
151add ackn
152document.title part 1
153the html/head/title/body elements; document.title
154add some quotes around 'the body element' in one case where we mean the formal term-ish; add some notes of what needs defining
156add some notes from my research
157add some notes in teh parser text since we have some open issues with unexpected stuff at the start and end of streams
158add an example to the 'a' start tag case to explain the weird requirements
159Replace 'anything else' with something less confusing
160colgroup/column group
161Error Handling for collections; link security exception to a placeholder definition.
162handle HTML elements with no 'name' attribute in HTMLCollection (for templateElements, tBodies, rows, etc)
163document.domain and document.URL definitions (except for the hard parts)
164the body element and document.body
165Define the images, links, forms, and anchors collections.
166document.cookie; placeholder for
167getElementsByName draft 1, add notes for document.close()
168limit getElementsByName() to certain elements
169move some comments around to more appropriate places
170document.getSelection() in the DOM section
171notes in getElementsByName()
172document.commands definition; collapse some space in an IDL block due to comments
173clarify 'root element' since we define that term to mean something slightly different in some cases
174minor editorial tweaks -- capitalisation, etc
175add designMode placeholder section; link to that and execCommand in the document section
176getElementsByClassName() with string and array versions
177getElementsByTagName() now taks an array only.
178ackn update
179Twiddle things around in the Document section, remove useless forward references.
180Move the editing stuff into its own section
181Array isn't typed enough. Let's use another pretend syntax.
182start mentioning 4+ arg work first draft completion.
185document.close placeholder draft 2
187wording tweak in document.close()
188define document.innerHTML, do cleanup from the fallout that causes
189remove 'proposal' markup in canvas section, tighten up some of the canvas normative prose, and poke at the document.write()-in-XML case.
190innerHTML and document.write() for XML, first complete draft; also correct the intro now that we split the APIs into four sections.
191HTMLElement TBW->WIP
192move the dom sections around a bit; markup 'HTML documents' and 'XML documents'; minor related tweaks to text.
193big overhaul of the use of HTML, XHTML, HTML Documents, XML Documents, XHTML, HTML Namespace, and other related terms. (also, fix it so 'document.anchors' is hyperlinked correctly.)
194big, not bug...
195Move sections around in the DOM section again; add a section on APIs in HTML documents (case-sensitivity etc)
196Move HTMLElement definition up to where it should be (part 1)
197HTMLElement cleanup
198Make HTML Elements a term too (hyperlinks some stuff, weee)
199hyperlink 'html elements' some more now that we define it
200make use of HTMLElement for unknown elements a requirement. add placeholders for contenteditable, tabindex in HTMLElement. replace forward reference with an issue asking for general forward reference
201more work on HTMLElement
202global attribute changes (mostly editorial)
203move global attributes up a bit
204editoral: used wrong elements dt/dd; also, didn't define class and contenteditable correctly
205try making the global attributes thing an element dl
206extraneous > in markup (probably from the original, too, go figure)
207HR element (still looking for a decent backronym for that, let me know if you find one...)
208HR done, not TBW anymore; change <i> to <x> for now.
209markup error x/ -> x</
210<t> element, draft the first.
211Ignacio Javier pointed out a typo and confusion (TCP connection security handling paragraph)
212missing quotes make validation hard
213add issue to ask for help on <t>. Also, make a half-hearted attempt at clarifying the timezone definition.
214markup error in example; <t> becomes </t> now
215microsyntaxes for numbers; some cleanup related to the use of terms like 'whatespace'.
216oops, s/gauge/meter/...
217don't require violation of other specs (idness); add the word DOM in certain places for clarity; add some issues -- we're going to need to redefine some attrs in terms of microsyntaxes eventually.
218profile attribute: should use microsyntax. Also, make it use a different hyperlink title for consistency.
219placeholders for URI microsyntaxes
220add reminder to go through and define all attribute value syntaxes
221refs have square brackets
222Move the definition of the datetime attribute up to the microsyntax section
223How to parse datetime='' on <ins> and <del> elements.
224placeholder for 'date and/or time' section
225spacing issues with inequalities
226forgot to mark example as example
227spacing issues with inequalities
228introduced extraneous character at top of file, oops
229date or time (rules for parsing <t>); move some macros around to help with that; add a comment on the use cases for APIs for <t>.
230tokens part 1
231Tokens part 2: define the algorithms; make DOMTokenString use them.
232class attribute redefined in terms of microsyntaxes
233Remove the placeholders for microsyntaxes for URIs for now
234clarify minor things about HTML Documents vs XML Documents
235make document.innerHTML kill any active parser first (on setting)
236behold! placeholders for multimedia elements.
237make some minor edits based on comments by Stephen Ma (networking).
238<iframe> section header
239note bz feedback about innerHTML
240embed, object, param, map, area headers. Make comments and PIs explicitly not affect content model.
241notes to self about usemap, area
242move attributes around so that the order is consistent across elements
243placeholder for canvas
244Believe it or not, microsyntaxes are defined now\! Weee\!
245remove bogus link to 'views'. place [] around a reference without []. Move Link Types section to a new section under Processing Models.
246hreflang might not be present
247The Great Link Abstraction: factoring out common definitions for <link>, <a>, and <area>.
248WIP on <area> and image maps
249dfn img-alt and img-src; mention usemap in img and object definitions
250DOM attributes for embedded content; also, add some 'must's to DOM attribute reflections that were missing them
251dom-a- to dom-img- for the img ones, heh
252quirk is covered by definition of 'reflect'
253make sure className.match(), etc, still works
254note empty sections
255more notes for image maps ismap/usemap processing
256add more stuff to <img> definition, slightly more placeholding for usemap
257behold: the <img> element. Placeholders for some sections that the <img> definition now refers to that don't exist yet. Also, changed media to medium in certain cases.
258remove one mention of CSS pixels and just make it say length, so that we don't have to ref css21 twice in one section, and so that we don't constrain the rendering section
259Constructors on WindowHTML: draft 1; note that IDL does give type info which we mustn't lose; define firing a load/error event; define new Image(); img.complete; define <img src=''> and <img alt=''> better.
260move canvas placeholder up to before area/map
261<title> cleanup now that we have a parser and a DOM
262iframe element definition
263clarify <iframe>
264src image content requirement; img when images are disabled; onload/onerror for iframe; no elements in iframe; remove alt on embed; define embed -- draft 1.
265issue about ordering of events on iframes and content of iframes
266various notes on type='' attributes and on img height/width, plus, work on the definition of embed.
267adding more text to object section
268about face for <embed type=''>. Instead of making it advisory, make it the type decider.
269clarify HTTP error codes with <img>
270we will have to clarify handling of HTTP error codes with <embed>
271should we say that content-sniffing is on for <embed>?
272placeholder section for Content-Type stuff; <object> element draft (incomplete)
273interfaces from plugins on embed, object; errors on object cause fallback (says acid2...); define object processing model, first mostly full draft
274param element spec, and integration with object; allow any attribute on embed; fix s/src/data/ typo in object dfn; clarify that processing of object is 'first match' in the processing model.
275we don't define NPAPI
276<map> definition
277<table> element. factor out requirements for HTMLCollection attributes. Prepare HTMLCollection attribute for out-of-order collections (like table.rows).
278<caption>, <colgroup> (incomplete); move a paragraph down in the structure section to use terms after it defines them; s/zore/zero/
279link to anne's diffing tool
280split reflecting of numbers into reflecting of signed and unsigned numbers; fixed references to toNumber; <colgroup> second part; <col>; table processing model placeholder
281setting span to 0 is bad too.
282thead, tbody, tfoot, tr, and tweaking of some other table wording
283move the TBW marker to the processing model section; correct a minor validity typo
284correct more typos
285don't require a <tfoot>
286<col span> titles were wrong and pointed to colgroup's
287the remaining table elements; placeholders for the DOM interfaces for the table elements
288define DOM interfaces for remaining table elements
289cleanup of some sections that weren't all that (drop <tabbox>, <switch>, move classes='' to processing models; add note about mutually exclusive sections use case to end comments; note some stuff under the rendering section)
290removing <calendar> and <card> since the group still hasn't formally and clearly defined how you're supposed to parse their microformats
291reorder some sections around in the processing model to slightly more closely match the order the elements are defined in
292the start of a tabular data model
293tabular data, row groups and column groups and more crossrefs
294constraints on row groups and column groups
295Corresponding elements (informative so far).
296could have a cell in no row/col groups
297Work in progress for tabular model
298Cross-Document Messaging update (should be pretty much done now)
299um, s/below/above/ -- i moved the section down...
300don't allow <col> under <table>
301oops, unclosed tag
302remove <col> from the list of elements to not ignore in <table>
303more work on tabular data model
304col and colgroup
305minor typos; row groups; make jumps back slightly clearer in table model.
306introduce the concept of table model errors; handle cells that span out of row groups
307more hyperlinks to the tabular data model section; more clearly label the steps in the table model section; make cells flowing out of the table into an error
308use 'current foo' to mean elements, and x-foo/y-foo to mean coordinates; use 'current column' instead of 'current col'; define row processing (still some missing bits); placeholders for some more table stuff
310markup errors; be more explicit about anchoring of groups and cells; introducing the downward growing cells list; factor out ending a row group; make ending a row group use the algorithm for ending a row group; define how to handle rows and cells; make an empty column be an error
311don't make all the cells overlap -- actually increase x-current! also, some comments, and, placeholders for growing cells.
312factor out the steps for growing downward-growing cells
313handle downward-growing cells (those with rowspan=0)
314fix step number in tabular data processing model
315markup error
316fix a bunch of markup errors
317split one item into two items in the algorithm for ending a row group
318placeholder so i don't forget to defined how TDs and THs work together (scope, etc)
319First attempt at a list of link types (they will then be defined in more detail)
320redefine 'up', add archives
321don't ask me what THAT was about
322merge the <a> and <area> columns since they'll always be the same; disallow nofollow and external on <link>
323change 'allowed on' to 'effect on' for link types
324make definitions of linktypes more consistent with each other. define 'alternate'. move comment down
325rewrite the top of the 'alternate' definition to say that it creates a hyperlink
326placeholder for extensibility of rel=''; rel=archives part 1; rel=alternate tweaking to merge the first two paras
327fix markup error
330downgrade some headers
331introduce the concept of synonyms for link types, add rel=bookmark
332try wrapping the example differently; remove comment about rev='' since we aren't introducing that in html5
333minor editorial fixes, and add rel=contact
334make rel=author apply to section, not just page
335rel=external, rel=feed
336Hierarchical link types; remove some comments that have already been dealt with; minor markup fixups
337somehow these words got dropped, oops
341rel=nofollow; mention copyright for rel=license
342target document -> referenced document
344auditory icons, not aural icons
345rel=search, rel=stylesheet placeholder, rel=prefetch
346rel=sidebar, rel=tag
347update comments/issues for link types
348rel=copyright is a synonym for rel=license; define the extension mechanism for 'rel' values.
349types can't specify anything themselves...
350make it even easier to add extensions
351make it possible to distinguish proposals from actual extensions for rel='' link types
352minor editorial tweaks to HTMLDocument and HTMLElement IDLs (ordering, whitespace); getting ready for .focus() and company; complete the definition of 'Rejected' in link types
353move sections about some more to make more sense
354introduce an interaction section for focus/click stuff
355more comments
356interaction, click, focus, etc: first draft
357hasFocus, activeElement, more structure in the activation section
358note elements that are yet to be defined
359<t> becomes <time>
360<x> -> <t>
361xref scripting is enabled / scripting is disabled; define 'text node' to include 'cdata node' and xref a bunch of those and simplify a few sections because of this; minor tweaks with the use of the term 'whitespace'; define <noscript>
362noscript isn't TBW anymore! wee
363the legend element (if anyone has a better idea of how to classify the elements so this doesn't end up in its own little silly 'misc' section, please let me know)
364make 'embedded content' an actual type of content
365the figure element and associated trappings
366add note about win1252 entities
367turns out there's no attribute minimisation in existing parsers. so that makes THAT easy. We'll have to go through and make sure '' is legal for all the boolean attributes, though. Please send e-mails to the list for any such attributes you come across that don't already allow exactly their name and the empty string to mean 'true', and the absence of the attribute to mean 'false'.
368update the TBW markings so we can easily see what's up to date and what isn't
369notes for syntax section
370start tags: draft 1 (wip)
371more work on syntax
372DOCTYPEs and other syntax bits
373doctype tweaks
374minor wording tweak to make things flow better
375tweaks based on feedback from atom folk to help feed autodiscovery; correct links to 'space character'; add dfn for doctype
376behold! attribute syntax.
377rel-alternative is transitive
378five more words on text... more coming soon!
379behold, for inline status annotation
380further work on the syntax section
381let's make that make at least a modicum of sense, shall we
382singleton -> void, for now; better terms are welcome (but 'empty elements' is too confusing)
383further flesh out the syntax section
384minor clarafications to syntax
385<html> tag omission
386head, body tag omission
387li, dt, dd end tag omission
388p element tag omission
389don't put end-tag-implying elements inside p elements
390entities are in the form &#0; not &0;. Oops.
391implied tags in tables
392add blank line for clarity in source
393mark this list as brief since it's not a set of paragraphs
394<thead> can't be followed by <tr>; <tbody> can't be omitted if it comes after an omitted </thead>, </tbody>, or </tfoot>
395appease the XML gods
396space characters in start tags: the return
397refer to later space restrictions for clarity
398various fixes to the tag omission stuff -- thanks simon
399markup error
400elliotte says Unicode uses 'code point' not 'codepoint'
401this is totally unnecessary
402slit on spaces? oops
403as Simon points out, two empty attributes need a space between them
404get rid of the idea of meaningful class names for now
405when can script run; whether script can run
406add the fact that TR can't be in TABLE
407try allowing xmlns='' and see what happens
409explain XML vs HTML
410parser section says 'HTML' must be uppercase. We can discuss if that should change (mail the list), but the syntax section needs to be consistent, at least.
411Misread the spec. Make it clearer
412make it clearer that conformance checkers must use the parser section not the syntax section
413add form controls to the optional end tag section
414make </optgroup> optional and make <optgroup> imply it
415various scripting changes (wip)
416change the paragraphs section to a prose section to be less confusing
418touch up the dialog examples and notes
419confusing indendation begone\!
420move the tabular and form processing model sections closer to their elements, add in some placeholders for form control sections
421move Window into its own section so we can see it in the ToC
422profile out, class and meta wiki registrations in. some predefined classes and metas too.
423profile out, class and meta wiki registrations in. some predefined classes and metas too.
424elements should say which classes apply to them; element names should be alphabetical in the table; search shouldn't apply to article
425move classes up one level, and put metadata attributes back under global attributes
426add IDs to be able to link to them from the wiki
427we have a section on that now
428move security into the scripting section, maybe it'll make more sense there
429try making this easier to follow by some slight changes
430more exposition at the top of the DOM section; minor ordering tweaks to make the top of the DOM section scan better. Nothing normative changed.
431Move Events down to Scripting section for now; move the generic DOM APIs up a bit
432move sections into the scripting section for real now\! (oops, wrong level!); notes for events sections, move stuff around a little...
433move the headings and sections section in processing models up to the headings and sections section in semantics
434slowly adding to the event listener section (work in progress)
435Specify event handler attributes! Just onclick for now. Also, make "HTML element" link to the right place a few times, allow the global attributes to be used, and move the javascript: placeholder down a bit to aid flow.
436xref Object to nothing instead of <object>
437remove redundant syntax requirement in conformance section now that we have a syntax section
438clean up the definition of the ID attribute to handle orphan elements, and to remove conf criteria that really belong in DOM Core, CSS, XPath, etc.
439title attribute shouldn't be used for captions now we have <figure><legend>; make the text of classes slightly simpler; allow class attributes to be used for meaning without registration if there is good reason to do so.
440try having some of the global attributes have their own sections instead of being in a <dl>. Bring the class attribute into that section to see how that looks. remove the redundant class=error section.
441cross-ref xml:lang better
442remove issue that's now been resolved; add idl class to a <pre> block that contains an IDL excerpt
443colour->color; overhaul of the canvas section; move 'sound' to the browser section
444crossreference fixes
445more cross-reference fixing
446make one of the links shorter
447notes regarding sniffing (nothing normative yet)
448fix table processing of bogus content (<a><table><a>xxx case)
449missing 'the'
450remove SCS markers (will be done out-of-band instead)
451oops, lang->xml:lang here.
452clarify 'it' in 'it must be dropped'
453fix validity errors
454remove TBW and WIP markers in favour of external marking script
455intro, not into. also, make intros non-normative.
456inter-element whitespace and comments should not affect algorithms; blockquote and q's interaction with cite; new example for cite; drop <t>; add <i> and <b>; add <div>
457extraneous blank link in example
458finish off the <b> element section (oops)
459move BR to the prose section, it's not a phrase element; it's a prose element.
460typo in acks -- sorry
461copy/paste error
462reconstructing the inlines when there are none would lead to problems.
463missed a step number when i added some steps
464vaguely try to make EOF parse errors clearer
465forgot to remove this note when we supported this
466introduction of WYSIWYG editors and the <font> element (content model to follow)
467transparent content models; div is allowed to contain inline or block now; fix some xrefs
468move transparent section up one
469broaden the definition of transparent content models so that we have 'transparent elements', and make the determination of inline vs block look through transparent elements
470make note of problems with the style='' attribute, and use cases we have to cover
471allow <style> inside <div>, <article>, <section>, <aside>; make <script> stay where it was put, rather than move it to the head (gives compat with IE); add a placeholder 'styling' section; make fixes caused by fallout from those changes
472scoped stylesheets baby
473embed the styling processing model placeholder into the spec a bit more
474fix markup, add open issue in parser
475fix transform() method matrix
476styling processing model (will integrate into Anne's CSSOM spec --
477Closer integration with CSSOM.
478factor out text for all boolean attributes; fix typo in an IDL; define <script> defer and async
479hoist the menu processing model section up to the <menu> element section
480change xref to use autoxreffing instead of an id
481ref from DOM boolean attribute to content boolean attribute
482enumerated attributes (affects dir, contenteditable, draggable); fix typo
483add note about percentages/dimensions; fix the time microsyntax to allow seconds to be optional
484fix note on enumerated attributes, add note for rel=search
485clean up attributes that specify MIME types. add more stuff for rel=alternate.
486define 'language' as a spec term (this will make one thing xref to it)
487minor edits in the <menu> section to make it neater
488fix some xrefs. define value/start better (e.g. using microsyntaxes)
489URI attribute cleanup
490markup error
491make the example XML so it works
492define the values of the scope="" attribute on <th> in terms of enumerated attributes, correct a statement that said that cells could not span colgroups (they can, verified against implementations)
493add two parse errors for EOF in <noscript>/<noframe>/<iframe>/<textarea>; drop leading newlines in <pre> blocks
494How <th> elements work.
495fix the state names to be consistent
496xref fixing...
497very draft description of how navigation should work. more to come, but send comments now please.
498oops, markup error
499moving notes around to get the area ready; removing redundant editorial issues
500more work on history traversal and page navigation
501note about popState during traversal
502javascript: URIs and moving issues around for things that were dealt with in other places
503start to a definition of content-type sniffing (very WIP). add another example of what could end up not affecting the browsing context.
504markup error
505a few more words on content-type sniffing
506wip: content-type sniffing details
507make some non-normative text into a note
508more work on the content-sniffing code
509oops, that made no sense. got to end the steps there
510minor editorial work on the content-sniffing front
511markup error
512how to handle XML files
513how to handle html and text
514scrolling (not that this says much, but that's really UI stuff so let's keep out of the way)
515fix xref (frag id stuff)
516WIP: adding sections for the content-sniffing bits
517how to sniff for text/plain vs app/oc-st content; point to other specs/sections for how to determine the encoding of HTML and text files; minor editorial tweaks
518remove big-issue markup since it
519typo "bytesof"
520sniffing image types; minor commenting tweaks in the text/binary section
521add unknown type sniffing. editorial tweaks to image section and fix for markup errors
522i am a moron
523i am a moron
524add support for PDF and Postscript
525must->should on adding new types, but give rationale for discouraging it
526should <object> trust the Content-Type headers? added as an issue, along with similar ones. fixed editorial 'third-party' issues. reword some things. allow image/svg+xml to be handled by a plugin even though it is XML. other minor markup changes.
527add descriptions to the tables, for bz's benefit :-)
528markup error
529make the section paragraph be a big-issue instead of the section title.
530make the hex uppercase to be consistent with the rest of the spec
531how to detect Atom and RSS vs text/html -- if you have any comments on this PLEASE LET ME KNOW
532add support for RSS 1.0 in draft form (this will need fleshing out at some point)
533fix typos in the original text i wrote here, add omitted ASCII equivalents in places
534typo in the comment
535define copy-and-paste operations; fix grammar of drag-and-drop, fix spelling of canceled
536first draft of text for 'following hyperlinks' (incomplete pending 'target' attribute definition)
537add the text that makes event handler attributes work; also: make 'event handler attribute' an actual term with a DFN element; move the issue about non-JS listeners to lower in the prose
538move Event handling section to the 'other specs' section
539activation behavior; remove redundant comments; define .click() better; remove some mentions of xml-events namespace; and some notes of stuff that DOM3 Events should say; note duplication in such notes.
540remove obsolete notes and issues; change remaining occurances of xml-events namespace to no namespace
541move the activation behavior stuff into the interactive elements section for now
542vaguely define clicking on <a> and <area> elements. mention <base target> so that Anne will stop bugging me about it. :-P
543minor tweaking (alternate stylesheet -> alternative style sheet; define alternative style sheet sets for xreffing)
544notes for image maps
545make more notes on image maps
546Allow keywords to be uppercase, even in XHTML
547notes for image maps
548fix xrefs for 'enumerated attributes'
549move the <figure> element up, and the image map processing model into the embedded content section
550<area shape> attribute syntax and parsing definition
551markup error
552more markup fixes
553update copyright date
554oops, copy/paste error
555DOM interfaces can't have elements, duh
556go Simon! more typos :-P
557hoist up the image map section
558research into 'coords' attribute's % values
559draft comment for how to parse coords attribute
560how to handle lists of integers (e.g. in coords='')
561actually support negative numbers (duh)
562oops, had the 'every other' case before all the other cases
563support decimal points in coords='' (if by 'support' you mean 'ignore like IE does')
564area elements: when can you use coords? how many coordinates must there be? when can you use ping/rel/media/hreflang/type?
565rules for attribute presence/absence on <area>
566define usemap="" (first pass); make some notes for future stuff to come
567framework for image map processing model
568removing 'name' attribute from <map>. Note that for legacy reasons, it's still part of the processing model -- it's just not allowed in content.
569how to read the 'usemap' attribute
570how to handle <map> for text browsers
571user-agent-defined, not user-defined, oops; add a must to the start of the next bit...
572how to go from an <area> to an actual shape. make note of previously unguessed quirks of IE coords parsing. mention that <area> having focus is ambiguous.
573in error 'usemap' doesn't mean image map is empty, it means there's no image map at all.
574notes on what should be added next
575get rid of the notion of valid areas
576stop <sub> elements from creating xrefs
577make text flow better
578Turns out that the inner circle of radial gradients should be filled.
579note security issue with ports on storage APIs
580xref error
581move some notes around; decide on putting server-side-image-maps near following-hyperlinks; define clicking on an <area> element.
582oops, no detail on Event.
583remove note and convert it into a 'note'
584Define what the various coordinates actually _are_.
585image maps are live
586DOMTokenString -> DOMTokenList and make it not inherit from DOMString; add relList and classList; define the kind of content allowed in rel='', and how to go from that to a list of link types.
587server-side image maps. I'm not really sure that the whole hyperlink/image-map/click/.click()/click-event/DOMActivate thing is perfectly set up yet, but for now this is what we have.
588Update the head of the spec to be more friendly.
589make note of a testcase
590behold, designMode
591acks for designMode
593first pass at defining a bunch of events -- any other requests?
594event handlers
595the serialisation "<head><script><p>" is not allowed if the <script> is a sibling of the <p>
596very early work in progress on the encoding detection stuff
597how to handle comments (bogus or otherwise) when scanning for a charset declaration
598oops, <!--> is not a valid comment. My bad.
599oops, < is 0x3C not 0x2C
600more work on the sniffing algorithm
601more work on the character encoding declaration sniffing algorithm: make sure to handle going past the nth byte in the <meta> parse step, fix the markup (missing </ol>), add the prose for handling tags other than <meta>, clarify what 0x3E is in the bogus comment bit, add notes for parsing attributes.
602make sure the spec is explicit about when you just ignore stuff in the encoding decl detection algorithm
6030xe5 was meant to be 0x45
604mainly: how to handle 'charset' and 'content' attributes when looking for a character encoding declaration. also: add more notes for when we make unicode character references more consistent; also, a bunch of markup fixups of little consequence
605turn a redundant requirement into a note
606if the quotes are unmatched, don't do anything. also, don't mention that parsing begins after finding an encoding.
607incomplete description of how to parse attributes in the sniffer code; add a note about what UAs can do to make sniffing more performant.
608finish off the algorithm for detecting the encoding declaration
609notes for the authoring side of character encoding declaration
610backreference charset='' restrictions from the syntax section
611a variety of changes related to character encoding stuff; <meta http-equiv> framework; minor editorial changes
612tweak the way that we describe <meta> to read better
613more work on getting this <meta> element section to flow well
614oops, missing 'the'
615oops, forgot httpEquiv DOM attribute. also, make yet more editorial changes to make this section read better.
616fascist spelling attack
617get rid of vestigial DocumentUI references. Add 'location' to the HTMLDocument idl.
618behold: the world's first actual meta refresh specification.
619capitalisation grammar
620'starting' there reads weird
621Big rework of the <event-source> section based on feedback.
622minor editorial things
623oops, wrong element
624missing word 'attribute'...
625add support for BOMs to event-source
626allow CR and CRLF as well as LF in an event stream
627oops, moved the </pre> end tag to the wrong place
628indent issue
629oops, comments and commands _can_ contain colons, despite what I might have previously written...
630remove 'the' before 'official type' when the latter is a variable name; also, char should be any-char (terminal name didn't match names of productions that use it)
631Fix a few places that were saying we should scan 512 bytes even when fewer than 512 bytes were available. now we do all our type and encoding heuristics on the first 512 bytes only. this might change in the future, input from implementors is very welcome here.
632change note to a question because i'm not sure anymore
633the firing of the 'load' event
635add DOMContentLoaded (opera/mozilla extension)
636how scripts run while parsing (<script async> and appendChild()ed <script src=''> only; <script> and inline <script src=''> are handled during <script> insertion; other things are handled after parsing ends)
637remove DOMContentLoaded based on the thread at
638er, is this right? appendChild()ing an external <script> pauses the tokeniser? need to examine this.
639dramatically simplify how scripts are handled, and simultaneously make it closer to what browsers do today anyway.
640more tweaking around to make scripts execute appropriately
641make those actually normative conformance criteria
642abstract out the message event interface so that it can be used by both cross-document messaging and by server-sent dom event messaging. also, put in a comment for an idea by lachlan of a pragma directive for controlling when to refresh.
643fixed an example's description for event-source
644make it clearer that this is about HTTP requests not actual TCP connections
645load/error events on script elements; making various things delay the load event.
646add a stable ID to this header for external references
647fix xrefs
648oops, bad markup
649note that we should add textLength when doing input/textarea selection stuff
650oops, forgot to say where the event then got dispatched
651typo fix in href
652work around bug in postprocessor
653fix broken xrefs
654a few whatwg.orgs that should be www.whatwg.orgs
655behold: the javascript: URI
656step 1 for and friends
657introduce target='' for <a>, <area>, and, anne will be pleased to hear, <base>
658on <base>, href, not target, is the HTML-only attribute. on <a>, attributes aren't allowed if href='' is missing. fix some hrefs. define 'hyperlink suffix'. define target=''. move definition of ping='' nearer href and target.
659markup error
661if you remove the src= on an event-source, then it should also remove the event source thingy.
662bug in the character encoding detection algorithm
663WIP: change the requirements for scripting threading to be clearer about what exactly it is that is required. Allow UAs to ignore the target argument to if they like. define how to find a browsing context by name. also typo fix.
665support 'replace' for; factor out replacement support; also, add a note that 'url' must be resolved in calls somehow.
666oops typo
667what to do when you have multiple matching browsing contexts
668fix how we talk about objects that implement the [] syntax. grammar fix. support rel=sidebar and <base>; window, frames, self; length, opener, window[]. all very much WIP.
669move Styling processing model nearer to the <style>/<link> element definitions
670move the abstract commands section up to nearer the <command> and <menu> element definitions
671fix header levels
672merge the processing models and browser environment sections, and call it browsing contexts. move sound and repetition templates out to their own section. update the structure-of-this-spec bit to match.
673note new ref for haltingproblem
674new <video> element section. This is SO far from complete it's not even funny.
675experimentally add <meta name=dns content=...>
677make a note of an xref suggestion someone had
678the <video> element and assorted related changes
679keep a note of issues and feature requests that have so far been ignored
680minor changes to the <video> element based on feedback. for a complete description of the changes in this patch, see the e-mail I just sent to the WHATWG list. -Ian
681remove togglePause()
682adding names of people who gave important <video> feedback but that i forgot to add the first time. adding also some notes on features we could add in v2 (or v1.1?)
683new features to consider; change time plan for <video> -- this is just a vague plan, not set in stone or anything
684add 'muted' attribute to <video>
685split <video> into two: <video>, and an abstract Media Element concept.
686the last checkin did a huge set of changes because bert updated his script to fix some of the xref issues -- thanks bert
687codec negotiation for <video> elements: <source> elements; .load() and predownloading on insertion.
688Apple say that the offsets must be floats, so let's become mindshare-compatible with the flash api and use seconds througout
689add <audio>
690xref error
691oops, missed one of the places that should be float
692support for: start/loopstart/loopend/end/loopcount/currentloop; cleanup for DOM things in media.
693grammar typo
694oops, copy/paste error
695rephrase this to be clearer
696Section up the media section
697add a ui='' attribute to toggle ui on/off on <video> and <audio>
698rename ui='' to controls=''
699STOPPED becomes EMPTY
700work in progress -- declarations for states not yet defined
701change buffering state to ready state in the to-be-defined idl notes
702more tweaks to the strawmen
703straw men updates for <video> API
704updated names for <video> state straw men
705updated names for <video> state straw men
706what parts are seekable isn't the same thing as what parts are buffered. add another field to the straw man.
707* Fix the sample script for making the hypothetical tree when creating an outline * Removed notes saying <video> and <audio> were being kept simple for now * Changed to a new set of states for networking et al * Mention that the mapping from time to frames is codec-specific * Remove togglePause() completely * Add UI toggle (controls="") * Introduce error codes for media elements * Removed seek(), instead made position mutable * Added seekable attribute * Added event summary * Fixed IDL of <source> * Fix some step numbers in the content sniffing algorithm * Fail to define "fire a progress event" * Minor editorial work
709s/initial/first/ (spec was mixed before, and 'first' seems easier)
710clarify NaN
711clamp new positions to the current loop limits
712fix legacy of the old state definitions
713support for autoplay of content
714volumechange event
715update secret hidden notes for HTMLMediaElement
717hasAudio/hasVideo notes
718<video> is no longer pure black by default (rendering section will say to use the CSS background, and might default that to black); also, don't imply that video and audio must respectively contain video or audio.
719Allow for non-linear/interactive movies and define how they interact with seeking.
720Add links to the new multipage version of the spec; add support for defaultPlaybackRate and playbackRate to the media element section; add support for ratechange and durationchange; fix some reference section links
721Added cue points to media elements, so you can get callbacks called at specified times in a clip.
722Add class=idl on an IDL block which was missing it.
723Define 'pause playback' for cue points; define what happens when a callback is null.
724Add <source media='media query'>, support a fallback <source>.
725Oops, added a crash bug to the algorithm. Fix that quickly...
726Changed the 'error' attribute from a number to an object, for extensibility.
727Added security section placeholder for the media section.
728standardise on a way to reference DOM3 Events.
729We removed the uuid: namespaces some time ago.
730Removed the entire Audio section in favour of the new HTMLAudioElement. 'new Audio()' still works, though.
731Define the term 'browsing context' and related terms; add placeholders for sections we'll need to define non-Document documents
732Fix a cross-reference; make some of the non-normative statements more obviously into notes.
733Reorder subsections in the Browsing Context section for readabliity.
734Reorganise the Browsing Contexts section some more
735Replace WindowHTML by Window and get rid of dependency on the Window spec
736Try to tie more loose ends togother in the browsing contexts section: define scripting context, security context, global scope
737Minor cleanup of the Window definition
738Minor cleanup of the Window definition
739Allow UAs to not consider two windows related to each other for the purposes of browser context communication.
740Fix the definition of Browsing Context to match reality.
741minor editorial changes to the browsing context section
742Moving 'user prompts' section around so we can deal with auxillary browsing contexts earlier in the prose.
743Defined auxillary browsing contexts; moved some definitions around
744Continue making changes to the browsing context section in an attempt to find a neat way to define this mess.
745Try yet another way of organising this section of the spec.
746Redefine 'top-level browsing context'. Define 'fully active'. Clarifications.
747Rework some of the definitions in the Browsing Contexts section to be clearer and better defined.
748Expand the algorithm for how to pick browsing contexts so it supports popup blocking. Add more hyperlinks in the following APIs section.
749Update the definition to handle the new definitions, also add a bunch of hyperlinks.
750Define the order of frames in window[i]
751Update the session history section to use the new browsing context terms. Nuke the idea of showing non-active documents.
752Move User Prompts down even further.
753Remove obsolete parts of the History interface, clarify what it means to 'traverse the history'
754add hyperlinks
755Remove an infinite loop when navigating with replacement; make Location instantiate per-Document.
756Fix the navigation and history traversal algorithms so that the back button doesn't lock up the browser if the Document was evicted
757Define the URI decomposition attributes.
758markup typos
759Minor tweaks to the URI Component Attribute APIs
760Hyperlinks section: fix a typo, and fix a conf requirement to actually require what it meant to require
761Clarify a step in the navigation algorithm to pre-empt people from saying it cancels itself...
762Clarify the processing model for navigating to text, images, plugins, and UA inline content.
763Write first draft of Content-Type section; remove the paragraph saying inline charset data should be avoided.
764User prompts are now defined (window.alert, window.prompt, window.confirm)
765Fold the concepts of global scope, scripting context, and browsing context into one.
766Move the Thread section up to Browsing Contexts
767remove obsolete notes
768Massive revamp of the scripting security section. Please review this carefully!
769markup typo
770Reference the RFCs for parsing URIs; split a paragraph into two.
771Redefine 'onerror' now that we have event handler attributes
772Add notes for onerror in the ev handler attribute section
773move script errors into the events section since it's now an event
774Make the Window object support events
775fix some xrefs; change some 'XXXX's to 'XXX'; add placeholders for unload/beforeunload events
776Placeholder for beforeunload/unload; interaction of and navigation/session history; define initial state of browsing contexts (about:blank); define secondary browsing contexts
777Specify the placeholder='' attribute, a way to say that an element is irrelevant and should not be shown.
778Mention commit-watchers and twitter in the version history section
779markup error
780placeholder -> irrelevant (placeholder conflicts with a webkit proprietary attribute)
781correct URL for commit-watchers
782Fix a couple of typos
783reflow paragraph
784Fix typo in the sniffer.
785typo fix
786I'm just an idiot, I keep getting this wrong. Thanks go to Wakaba for patiently pointing out my mistakes.
787Be explicit that line breaks in title='' attributes are indeed line breaks.
788postMessage() is on Document today; do we want to move it to Window?
789Make _blank not a valid value for target=''. This might not last...
790Fix the table DOM accessors
791Error handling for people setting table accessors to inappropriate nodes
792Address known <base> element issues.
793Add a link to CSSOM for David
794oops, copy/paste error
795try to make isPointInPath() clearer
796terminology fix
797Make a note of a feature request for canvas patterns
798move the images
799split the header out into a seperate file
800The HTML5 spec is all grown up! A copy is now hosted on the W3C site. The WHATWG and W3C versions are now identical.
801Fix duplicate IDs
802Update the section on proprietary languages given recent developments.
803Canonicalisation of color values in <canvas> should be lowercase.
804drawImage() must raise an exception if the image isn't ready
805Make note of canvas clipping suggestions by Jordan OSETE.
806createPattern() with an incomplete image must also raise an exception.
807Clarify error handling for getImageData/putImageData
808Oops, height and width of <canvas> should be unsigned.
809Minor editorial fixes in the canvas section
810<canvas> gradient updates, cross reference fixes, and editorial fixes
811Define what happens when you set the width/height attributes on <canvas> better.
812note various v3 feature requests
813notes for dashed lines
814Add an example of creating ImageData objects
815Fix some more gradient wording issues.
816Define currentColor resolution; define strokeRect() better in the face of zero dimensions
817move the 'relationship to...' sections into the scope section instead of the conformance section
818Define some error handling for methods and attributes passed NaN, Inf, etc values.
819Define when infinities are allowed in the canvas APIs
820Reference Porter-Duff for compositing operators; Add a requirement that getImageData() not be pre-multiplied (nothing else is).
821Change the rules for how to handle extraeous arguments to methods
822Change the way we allow custom ImageData objects to be created
823Begone, predefined class names.
824Added element.classList.toggle('foo') support.
825DOMTokenList can now be enumerated.
826Define how to distinguish site-wide headers from page headers
827Let's try that again. Distinguishing site-wide headers from page headers:
828Let's try that again. Distinguishing site-wide headers from page headers: now semi-readable and more correct.
829DOMContentLoaded event, draft 1
830Make the class attribute section fit in better
831Define handling of bookmarklets.
832Define 'load' event processing for <iframe> elements ( )
833Fix innerHTML for XML
834it's not the DOM attributes that contain the URIs, it's the content attributes.
835Make the .dir DOM attribute work like in IE.
836Security fix: document.location must not return document.defaultView.location, since the latter might be for another document, and would reveal the user's currently loaded page. Instead, make document.location return the Location object for that document.
837Make the definition of origin for bookmarklets actually make sense.
838By popular request, a SQL database for your enjoyment.
839Make ID have a lowercase 'd' for consistency with other parts of the spec; other minor editorial fixes
840Forgot to define Location in the Window interface
842Define what happens when multiple rows are inserted at once.
843I forgot that attributes have to be lowercased. Fixed.
844Split the requirements on Function/EventListener from the requirement regarding event handler attributes return false to cancel.
845Define and allow framesets to host documents from other domains and then change the frame again
846Clarify the <datagrid> content model, allow <datalist> as well as <select>.
847change the 'popup' value on <menu type=''> to 'context', and rewrite the section to use enumerated attribute value primitives.
848Make scripts not run in non-active documents; add issue regarding this for events.
849add link to the script section from this issue
850Support document.dir (see also the last checkin, which accidentally included some of these changes)
851Define document.lastModified, fix cross-references to document.dir.
852Define scrollIntoView(). Fix a typo ('dir content attribute' not 'dom content attribute').
853Event handlers only cancel when their return value === false.
854note for when we do references (for script types)
855Name update in acks.
856Move the definition of document.dir around.
857Rename datetime to dateTime for the DOM attribute
858Allow async and defer to be specified without 'src'; don't allow them togother.
859activeElement returns the <body> element by default, not <html>, for IE compatibility.
860Expand the reach of browsing context names to reach auxiliary browsing contexts in certain cases musn't effect other domains
862A bunch of editorial changes in the 'traverse history' section, cross-reference fixes, etc.
863browsing context name cross-reference fix
864Typos in IDLs, missing cross-references.
865Oops, scrollIntoView() as defined would be useless... let's fix that...
866Make HTML's document.title play nice with SVG's document.title
867<textarea> elements strip leading newlines, just like <pre> elements
868Move the stability section into the common spec body to make sure it stays in sync across the WHATWG and W3C versions
869Handle entities in the range 128 to 159 (0x80 to 0x9F) as per legacy requirements.
870Oops, entities in the entity section were written wrong. Also, make the table prettier.
871Make this section ID stable.
872Change the notes and terminology around conformance checkers / HTML validators.
873Add advice that Henri wrote for us for users of XHTML.
874Make U+0000 characters a parse error.
875HTML5 parser algorithm change: since CRs can come in as NCRs, we have to handle them in the tree constructor stage.
876So it seems that actually the last checkin might have been wrong. CRs get treated as LFs even as NCRs. What should we do?
877Say what keystroke we suggest for inserting a <br>.
878typos and cross-reference issues
879Change recommendation of what SGML tool users should do
880Parse errors for <li> <dt> <dd> closing other elements.
881Disallow spaces in IDs.
882add note regarding CR entities
883Define processing of <script type> and <script language>. To some extent.
884After a LOT of studying crazy browsers, I've decided to make a tiny change here to the way </font> end tags are handled.
885Oops, the grammar made no sense here. Fixed.
886Support the insane comment stuff in CDATA and RCDATA blocks
887Move requirement around in response to feedback to make it harder to miss.
888The current input character hasn't been consumed yet, so there is no 'current input character'. (twice)
889Add support for &TRADE;.
890Fix uses of the term 'tag name' to be less ambiguous.
891Make innerHTML in XML return a namespace-well-formed internal general parsed entity representation, so that it handles elements with multiple child nodes.
892Adopt nodes when setting innerHTML in XML to avoid NS_ERROR_WRONG_DOCUMENT_ERR.
893Made <script> elements not run when set via innerHTML. Also, editorial fixes.
894Cover the various ways that a DOM might not be serialisable, and make them raise an exception.
895Turns out that U+2329 and U+232A are both non-canonical. Replace them with their canonical decompositions.
896Disallow invalid characters like U+FFFE and U+FFFF when serialising innerHTML in XML.
898Define how entities and semicolons work together. This also removes the parse error for omitting the semicolon for certain entities. We could add it back if people want it back, though.
899Make the networking and parsing sections consistently vague about how to handle bogus bytes.
900Slowly edge towards removing the magic that < characters do inside tags
901Remove support for <script> </script<div>.
902Remove all remaining magic of the < character in tokeniser.
903fix inconsistency in an example; remove an extraneous 'instead'
904Link to the definition of the XML term 'internal general parsed entity'
905more work in trying to make this paragraph clear
906More work on making innerHTML in XML be well defined
907markup typo
908Quirks Mode
909Missing DOCTYPE is a parse error.
910numberic typo
911Oops, accidentially removed these sections in r900.
912Change the quirks mode switches to a list instead of a table (the table was unwieldy). Also, define 'almost stanards' mode.
913it's a brief list (not a complicated one with subparagraphs); mark as such
914Make </ not be a parse error in CDATA blocks, because it otherwise would disallow <script type=text/xml-script> which is quite common and frankly not hugely problematic. This changes continues our road away from XML/SGML compat.
915Change handling of <nobr> in the parser.
916Handle implied end tags with </body>.
917Allow && and other constructions that aren't entities and don't look like entities.
918Handling of </html>, </head>, </body> end tags in the <html> and <head> insertion modes
919Note for things we need to say about CSS later
920Support 'action' and 'prompt' attributes on the obsolete <isindex> element
921Need to mention layout tables being bad.
922Make the innerHTML getter output <pre> and <textarea> contents correctly.
923Remove excessive 'must's in the Document.innerHTML setter; be more explicit about the handling of </foo> in the innerHTML parse case when <foo> is not a PCDATA element (e.g. setting a SCRIPT element's innerHTML to '</script>').
924Support <!--> and <!---> comment forms (non-conforming, at least for now).
925Oops, forgot to update the syntax section with the new rules.
926Made <h1> headers nest. PLEASE GIVE ME FEEDBACK ON THIS CHANGE.
927Fix the parsing of datetime timezones.
928Missing space character.
929more examples of how innerHTML isn't going to always work
930Disallow CDATA and RCDATA from containing end tags that match their end tag (required now that we allow < and </ in general). Allow < in attribute values for consistency with the parsing section.
931Abstract out the innerHTML serialisation algorithm.
932Abstract out the innerHTML parsing algorithm.
933Fix a cross-reference to innerHTML-XML. Change the parser's references to the innerHTML setter to instead reference the HTML fragment parsing algorithm. Clarify that the fragment serialising algorithm serialises the contents of the node, not the node itself. Split out the first step of the fragment parsing algorithm into a note before the steps, and two steps at the start. Rephrase the algorithm so it doesn't say MUST anymore (since it's an abstracted out algorithm, the must is used in the invocation of the algorithm now). Add a note that the reset-insertion-mode algorithm will refer to the context node.
934Rename 'innerHTML case' to 'fragment case'. Remove the handling of innerHTML-HTML from <script>'s rules, as the parser does that for us already.
935<base>, <link>, <meta> elements no longer get moved to the <head> (compatibility with IE). This changes certain parse errors involved with those elements.
936Change the processing of comments in the encoding sniffer
937lastModified: year is 4-digit, not 2-digit. Fix errors in the document's validity. Fix a typo with the ASCII byte code for an exclamation mark. Be explicit that 'up to the formatting element' includes the formatting element.
938Disallow scoped <style> in <head> and non-scoped <style> outside <head>.
939Oops, screwed up the last checkin (didn't include all the changes)
940Allow <noscript> in <head>. Remove most uses of the 'head element pointer'. Abstract out an algorithm for most CDATA and RCDATA elements in the parser. Make <xmp> use the same technique as the other CDATA/RCDATA elements. Fix the bug in the syntax section that allowed <script> and <style> to start BODY elements with an implied <body> even though those elements actually just get put back in the <head> if you have them after a </head>. Fix some issues and add a comment in the spec about how innerHTML on <head> is handled. Some editorial consistency changes. Apologies for mixing in editorial changes with this patch.
941Oops, forgot to mark autoclosing of <noscript> in <head> as a parse error
942&#x0D; now turns into a U+000A LINE FEED character
943Be explicit about what an invalid Unicode character is.
944Strip whitespace outside the root element from the DOM
945allow arbitrary characters to be in attribute names; ban U+0000; reflect the recent rule changes for entities into the writing section.
946Update to the rules for handling of entities (require semicolons, and some changes for parsing entities without semicolons when in attributes).
947Move the 'permitted slash' definition up.
948EOF in DOCTYPE should set quirks mode.
949Change the wording to be more consistent with the rest of the spec to make the spec less surprising.
950make <plaintext>.innerHTML work
951Make sure to reprocess tokens when you have a token that sends you straight to the root element phase without a DOCTYPE.
952Allow <meta http-equiv> in <noscript>; fix the infinite loop with <noscript></body>.
953Write the restrictions on the contents for CDATA and RCDATA sections. Again. This time hopefully it'll stick.
954Make the <meta> rules clearer; fix typos and consistency with the 'escaping text span' stuff
955Change the way encoding detection is handled to support the de facto reparse mechanism. Discourage UTF-32.
956Add Korean equivalent of Channy Yun's name to the acknowledgements.
957Not having any encoding information at all is a bug if you're not using ASCII.
958ban more encodings
959Tweaks to requirements for encoding declarations and support
960Make a new section to put various charset support reqs together; fix a xref
961Make </p> and </br> introduce corresponding empty elements for compatibility with IE
962</p> and </br> implying <p> and <br> are parse errors.
963Let </p> and </br> tags through the first few insertion modes; make </p> be one parse error not two.
964nested-<nobr> should be a parse error, and it should reconstruct the active formatting elements again if necessary.
965As far as I can tell, the table sections should also generate implied end tags. No?
966Oops, <title>s are supposed to stay in the <head>.
967Clarify who is in charge of dropping BOMs. Hint: it's not the air force.
968Make 'new Audio(url)' actually start the load; also fix typos and remove a redundant step in the table algorithm.
969Note that not all known issues are marked.
973markup typo fix
974adding notes for menu construction.
975make note that 'icon' should be honoured for toolbars, make it clearer (i hope) that tool bars are a kind of menu
976the 'show' event should have key information. Mention that, and other event-related stuff, in open issues.
977Make the DragEvents be UIEvents instead of regular Events (not that this has much effect on anything).
978The prose and table contradicted each other.
979Dragging selection might want to do other types than text/plain -- add issue to that effect.
980we should have modifier key data in the DragEvent object
981Oops, this note is no longer accurate.
982Require that <base> elements come before anything that they affect.
983Sniffing for HTML has to detect <script> tags, as it's common for ads to be served without a Content-Type but starting with <script> alone.
984Mention that coordinates are relative to the displayed size, sigh.
985Prepare the image map section so that the spec can support <input type=image usemap> when we add <input>. This is compatible with HTML4, but maybe this shouldn't be supported.
986<map> elements now have a .images attribute that returns the associated images.
987Make id='' required on <map>. Let me know if that's a problem.
988note type of case comparison needed for hash-id-refs, at least for <map>
989Moving the cross-document messaging APIs to Window instead of Document (sorry Opera)
990Note that we need to clarify 'domain' more.
991Clarify that postMessage() is synchronous.
992Make pushState() take a URI and a title for improved awesomeness. Update references to URI/IRI/IPv6 URL RFCs.
993Mention that relative URIs must be resolved in pushState(). We don't yet say how. Resolving URIs in script arguments are a wider issue.
994Let's try adding an event that fires when you change just the fragment identifier, so you can catch that.
995Clear up the navigate-fragid section (nothing important changed really); include acks for onhashchanged.
996Make the Location members not be readonly, since they can be set.
997Drop 'usemap' from <input>. It's (very) rarely used, and when it is used, not supporting it usually results in a better experience. It's not widely supported (only Gecko and Opera support it).
998Fix various typos with irrelevant=''.
999Add .bmp to the list of image file types.
1000Revamp of the <img alt> rules. Also, some editorial notes on Content-Type handling, on the port-scanning issue with <img>/onload, and on a proposal to make the origin tuple include the IP address to avoid DNS rebinding attacks. And a typo fix.
1001add some red boxes to label controversial stuff as known-controversial
1002oops, typo
1003more typo fixes
1004Mention the possibility to omit alt='' attributes in HTML e-mail. Remove an old editorial comment.
1005Ban transparent spacer GIFs.
1006Make and return things.
1007Let's try this for img height/width
1008New alt='' example.
1009Allow a lone height or width attribute, since that could be used to size high res images of unknown specific dimensions.
1010add an issue regarding src='' conf requirements
1011Let's try to make it even clearer that the alt text should be included where possible.
1012missing word
1013Adjust the sniffing rules to take into account data I've been provided regarding the frequency of occurance of strings in typical Web content.
1014simplifying the text and fixing the markup
1015Some security notes on the SQL storage APIs.
1016Make the patterns in the sniffing section support whitespace instead of having asterisks. Also add some notes in other parts of teh spec.
1017note instability of <m>
1018fix the way metadata elements are referred to
1019<datatemplate>, <rule>, <nest>, and the related processing models: first draft (still missing introduction, failure handling for template='' and ref='', detailed rules for handling selectors, any rules at all for handling text expansion)
1020minor typos
1021Update the template section to handle failures during loads; handle dynamic changes of template trees. Still to do: expansion of strings, definition of how to handle selections.
1022Switch to having multiple named and versioned databases per origin. This is a work in progress; more changes (e.g. updating the version) are coming.
1023Introduce transactions and error handling to the SQL feature.
1024Add version support to Database. Actually make transactions commit (oops).
1025fix xref problems with 'expected version'
1026(oops, committed changes from a later edit by mistake)
1027make it clearer what the global transaction is
1028Clarify where unsupported SQL gets caught. Add a comment for errors we might want to report.
1029Rework the way SQL results are returned.
1030we'll have to define what kind of transaction locking to require
1031make a note about executeSql() with multiple statements
1032missing 'and'
1033Moving the Links section down.
1034moving the User Prompts section down.
1035Change the way that changing a database version works.
1036Moving stuff around some more trying to find a good way of ordering this stuff.
1037* Add comment regarding the imperative voice used in algorithms. * Removed references to RFC2732. * Fixed references to <hostname> (changed to <host>, mentioned <ihost> where appropriate). * Add note saying we need to handle IP addresses in definition of origin. * Added section 'Unscripted same-origin checks'. * Moved the 'Offline Web applications' section (and renamed it 'Browser state'). * Added section 'Offline Web applications' for offline application caches. * Made surgical changes to the history traversal to mention offline application caches. * Made surgical changes to the navigation to mention offline application caches. * Made surgical changes to the parser to mention offline application caches. * Fixed cross-references to 'accessible' in the storage section. * Defined '<hostport>' which the RFCs had dropped on the floor. * Fixed typos, added XXX notes to the editor, made certain IDs more consistent.
1038Fix getElementsByClassName() to take a string instead of an array.
1039Also port implicit entries to new caches (they get handled as foreign entries later if necessary).
1040rename 'application' to 'manifest' and define it
1041Clarify 'implicit entries' is only likely to be documents.
1042Block manifests from loading files from other schemes. This blocks, e.g., javascript:, mailto:, data:, and other dodgy schemes, without having to blacklist or whitelist anything, and without having to maintain a list of safe/unsafe protocols.
1043Remove a bogus conf. criteria in the manifests section.
1044Move broad requirements to after the specific requirements in the manifest syntax section.
1045Allow the CACHE MANIFEST line to have trailing spaces in the parser (was already allowed in the syntax).
1046grammar fix
1047added a note about resources in online and explicit sections. fixed the ommission of 'whitelist' earlier, too.
1048Note that a URI can have multiple categories. Move the <scheme> checking to the manifest and add() entry points, instead of the updater.
1049ApplicationCache: Make sure 'length' is well-defined; clarify a note in swapCache() that was supposed to clarify matters.
1050split the concept of 'update ready' from the concept of the cache group being idle/checking/downloading.
1051Define how ? replacement works in sql statements better. Suggestions on making it EVEN better are welcome.
1052Change 'rows' from an Array to a custom object.
1053Add a new SQL error code for 'dataset too large'.
1054mention that top-level browsing contexts bypass the online whitelist
1055Fix currentLoop mathematics.
1056Change loopstart and loopend to default to start and end respectively if not set; clarify that the 'effective foo' values are based on DOM attributes, not content attributes.
1057Explain codecs= values as examples to aid authors. Microsoft, Real, and others: please let me know what vodecs= values your media types (WMV, RM, etc) should have.
1058Redefine how transactions work to be clearer (no normative changes intended).
1059SQL: New error code for failure to obtain a write lock.
1060improve the text about video fallback
1061Make sqlExecute() take an argument array instead of varargs.
1062include a note as to an SQL example to include; add a note about rowsAffected and SELECT statements
1063fix cross-references and a minor markup consistency issue
1064SQL: Make expected version '' allow any version, but actual version '' not work with any random version. Make changeVersion() async.
1065SQL: Add an errorCode for quota exhaustion.
1066Add notes for future feature requests; clarify how UAs can handle not handling <video>.
1067Video: define 'reached' in terms to make cue points be better defined.
1068note v2 feature request, add Brian Campbell to acknowledgements (he suggested making cue points clearer)
1069merge video and media feature requests
1070another idea from brian
1071more suggestions for cue points
1072typo in description of 'loadedfirstframe' event
1073no more 'follow the following'; fix a sentence in the media section that mentioned video explicitly when it should have been generic
1074'properly rendered at all' is somewhat contradictory
1075A series of typo fixes in the <video> element section.
1076make it so that changing 'src' will cause load() to be invoked if it's not already playing
1077Make <audio> say the same as <video> for fallback content; add a note to both saying that the fallback content isn't for a12n uses.
1078A number of changes based on Apple feedback. Typos; clarifications; corrections to contradictions; allow effective start to be based on media data.
1079Moving <meter> around.
1080Add a note to <meter>.
1081Add more examples for <meter>; Mention units.
1082html is hard
1083also, i suck
1084forgot to acknowledge the person who made the <meter> suggestions
1085Make <script async defer> use 'defer', and allow that so that you can use 'async' and fallback to 'defer'
1087Gigantoredesignorrific changes to the SQL APIs. Enjoy!
1088Change 'errorCode' and 'error' to 'code' and 'message', since they're both on an 'error' object now. Also, fix interface names.
1089Add error code 6: key violation / constrain failure
1091markup errors
1092add note about db versions
1093move ObjectArray nearer to its use
1094Allow openDatabase() to have a display name and estimated size.
1095Apparently document.write() concatenates all arguments.
1096Revamp height/width for embedded content elements (img, embed, object, video).
1097Add the poster='' attribute to <video>. Takes a URI.
1098Oops, this example is now redundant (since you could do it with poster='').
1099RFC4281 baby.
1100Define what happens to media elements taken out of the document.
1101Define media element behaviour in the face of history navigation and garbage collection.
1102remove the versions of changeVersion() that have fewer than 4 arguments
1103kill SQLVersionChangeCallback
1104don't commit twice
1105In other news, loopCount became playCount in a formal and touching ceremony today.
1106Cue ranges are better than cue points. They have a whole extra dimension, for one.
1107Just make removal from the document fire .pause().
1108clarify which database we're talking about
1109Define 'lang' processing in XML more tightly; fix a bunch of xrefs ('HTML elements').
1110warn implementors about video.fullscreen() being bad
1111The localisation of error messages isn't a MUST interop requirement.
1112Breadcrumbs. More complicated than necessary.
1113make the global attributes section more consistent
1114Begone, significant inline content!
1115require attributes on <base>; mention that <base> can't be before <html>...; remove the vestiges of per-attribute annotations in the element <dl>s; do some cleanup; require href='' on the <a> that contains an ismap='' image
1116SQL: add a success callback to transaction() and changeVersion(); make the arguments array to executeSql() optional.
1117Cue ranges: made the 'pause' boolean be ignored during seeks, and other minor edits
1118er. mousey down? oops.
1119Introducing the rel=noreferer proposal, and other referer header changes.
1120Define how and work. (This is a poor API. Not sure how to make it better cheaply.)
1121forgot to ack tyler for the rel-noreferer thing
1122Define fragment identifiers, id='', name='', etc.
1123<a media='invalid'> shouldn't default to 'all' (mq defines that anyway)
1124rel doesn't create a hyperlink on <a>/<area>, href does.
1125Make <link> not necessarily be an explicit relationship, since that word is overloaded. Better suggestions welcome. Also, make note of link to OpenSearch.
1126Make rel=next/prev be non-hierarchical, just sequenced.
1127Make rel=first/last be non-hierarchical, just sequenced.
1128Split the first/last/next/last types from the up/index types.
1129Various typos and redundancies about recent link type edits
1130Make it legal to have multiple keywords in rel='', but only for rel=up.
1131someone couldn't work out the antecedent
1132Rename noreferer to noreferrer.
1133'may' makes no sense in this context.
1134'user response positively' is not english.
1135minor pubrules fixes
1136no-op checkin to reset the pubrules status, please ignore
1137explicitly mention that treating 8859-1 as 1252 is a charmod violation
1138Make PDF versions available.
1139Make the W3C copy an 'editor's draft' for real (previous changes in this direction didn't work for various reasons)
1140Massive changes to globalStorage to remove the multi-domain security model and replace it with the much simpler same-origin model.
1141Clarify definitions of what to do with and similar cases when it comes to sessionStorage.
1142Define which exception to throw when out of disk space (will likely change again when we have more exceptions defined).
1143Lift the cat who was amongst the pigeons up and put him back on his pedestal for now. (remove requirement on ogg for now)
1144Add a hint for what to do if you want to give the language of a <code> element's contents.
1145Various editorial changes to the sub/sup section.
1146typos in the <time> element definitions; hyperlink DOMTimeStamp to DOM3 Core
1147add an id for xrefs so that i can annotate that paragraph to say we're removing it...
1148<pre> element: semantics and examples
1149Add examples for ledes.
1150Fix errors in the example I just checked in, and acknowlege the people who (first) suggested it.
1151Update to FPWD status so we can publish.
1152Massive overhaul of the content models of most elements. The Block vs Inline distinction has been replaced by a new set of content categories, block/inline mixing is now allowed, the definition of paragraphs is revamped, and many related issues have been addressed.
1153Fixed spelling of preceded; removed idea about making <body> entirely optional; made <title> required again; clarified expected contexts for <style scoped>; prepared for making <ol> and <ul> allowed in phrasing content using <span> (requires parser changes); made <del> transparent; removed last mention of embedded content in <figure>; fixed copy/paste error in selectedRowIndex name; changed th to only allow phrasing content children; allowed <menu> to be nested; made <div> not transparent
1154Minor editorial fixes: cross references for 'the document's domain', and sectionRowIndex
1155Address Jason White's vote comments
1156Add a subtitle to clarify the scope of the document for people who don't read the spec. (W3C version only.)
1157s/Macromedia/Adobe/, remove some text I thought I'd already removed (discouraging XHTML use)
1158Transformations affect path creation as well as stroking and filling.
1159Clarify Bezier curve requirements; fix typo in noreferrerr.
1160Define what shadows really are in the <canvas> API.
1161Allow shadow blur dimensions to be capped; clarify that clearRect() isn't affected by shadows, opacity, and composition operators; various editorial fixes and clarifications.
1162editorial updates to the shadow algorithm to make it easier to read
1163xref fix
1164Color spaces and <canvas>
1165ImageData: make the object handle having new values set. Clarify what it means to call putImageData() (not drawing). Clarify how to handle non-ImageData data. Make the color spaces compatible with CSS; make toDataURL() not include color space information.
1166canvas: typos, clarifications, fix old examples, add new examples
1167Change copyright headers and other administrivia.
1168fix cross-references to missing references section
1169Woops, typo in the example.
1170Fix a potential security flaw in ping='' --
1171update the acknowledgements to take into account that we're not just using the whatwg list any more -- there are many other sources of input
1172Various changes to how to stroke lines, zero-length line segments, line joins, and line caps.
1173zero length-segments or zero-length segments? editorial clarification
1174add an id for the doctype sniffing
1176editorial changes to IDL, markup
1177Clarify img.complete; define toDataURL() with a 0x0 canvas; fix confusion in the radial gradient description; define patterns with a 0x0 input canvas; minor editorial fixes
1178Minor edits to the rectangle methods of the canvas section.
1179arcTo() fixes to line up with the only sane implementation, WebKit's.
1180<meta http-equiv=content-type> is now defined as conforming and equivalent to <meta charset>
1181switch things around so it's alphabetical
1183content-language is relatively common on henri's validator
1184Make the space in the meta http-equiv=content-type pragma optional, since it's common.
1185note some data from henri's stats
1186more data for http-equiv
1187Define arc() in odd situations better.
1188Canvas: animated images in canvas APIs; error handling for drawImage(); clipping path -> clipping region, and other fixes to clipping regions; clarify stroking and filling some more; remove old -- and wrong -- note about transformations and paths; other minor edits
1189note canvas v3 ideas, note future requirements when we have bindings-for-dom, note explicitly how forward-compatible getContext() must be
1190Remove any exceptions that would be raised by negative width or height values for any rectangles in the canvas API. (also, some minor editorial changes.)
1191Make gradients not be optional, since nobody cared they were optional anyway.
1192Overhaul of canvas security requirements.
1193move the colour spaces section above the security section -- oops
1194Revamp how infinities and NaNs are handled by the canvas API
1195canvas editorial edits -- inf -> infin, infinite line -> half infinite line
1196Changes to <canvas> color correction requirements
1197Revamp of ImageData and related APIs.
1198Update <a ping=''> to address recent concerns.
1199Make failure of setting an item on a Storage object return false, instead of raising an exception. Also, clarify one of the security risks to do with third-party storage access.
1200Redefine how the 'storage' event works. Remove the queueing of events. Add some v2 notes.
1201To help Mozilla transition to the new API, renaming globalStorage to localStorage. The irony is not lost on me.
1202remove 'shall' from the specification
1203Clarify exactly what is meant by some of the line join requirements
1204Simplify the rendering of the Join Triangle.
1205define which arc is the one rendered for 'round' corners
1206There are no outside corners when they're parallel, yet there's still an arc...
1207Redefine 'miter' corners to be more pedantically correct.
1208clarify the note on toDataURL()
1209define stroke() in more detail
1210Make sure to also mention strokeStyle as a cross-domain attack vector.
1211Fix some errors with the syntax and parsing algorithm for offline cache manifests.
1212Correct the last remaining vestige of the old 'application' attribute to 'manifest'.
1213make setItem() go back to raising an exception.
1214Security fix for postMessage(): What is needed is not the domain+uri, but the origin, of the message source
1215oops, extra trailing slash in example
1216oops, forgot to change arguments to initMesssageEvent*()
1217Add a second argument to postMessage(), the origin to which the message must be sent
1218dir='' defaults to ltr.
1219make it clear that the whole thing must be aborted
1220make it clear that the whole thing must be aborted
1221rename the content-type state pragma for <meta http-equiv>
1222Give the user agents more options in how they handle the Refresh pragma
1223Remove duplication of the list of void elements
1224setInterval() with a zero or undefined interval is treated like a setTimeout() by IE.
1225DOCTYPE parsing change: '>' can close a DOCTYPE even in the quoted parts
1226Make the content-type sniffing for browsing context navigation optional when the Content-Type metadata is present.
1227drag-and-drop: add dataTransfer.types; remove the limitation that copy/paste only be for text; remove bogus requirements; add placeholder for intro section
1228pushState(): Fix the problem of state objects not being correctly activated when traversing history. (Defines 'last activated entry', redefines 'activate the state object', changes the traversal algorithm.) Also, editorially: typo fixes; make a note to define about:blank.
1230Improve compatibility of the Location interface when faced with empty strings.
1231Clarify the processing of changing the fragment identifier.
1232Rename <m> to <mark> and dramatically increase the level of detail and number of examples regarding this element.
1233begin noting down what we need in terms of structure for the rendering section
1234Respond to feedback on <cite> and cite='' -- titles of works, citations, and examples thereon
1235Let's try a new strategy for ping=''.
1236Make <address> apply to the entire document when it's used not in a section.
1237* Made it possible to adjust the reconnection timeout. * Made it so that the client tells the server what the last message it received was when it reconnects, so as to allow seamless continuation. * Rewrote the implementation processing requirements. * Simplified the event stream format: - removed the concept of 'commands' - changed the comment syntax from starting with a ';' to starting with a ':' * Removed the possibility to control anything of the event other than the 'data' and 'type' of the event. * Restricted the events that can be fired to all use MessageEvent. * Made 'message' events not bubble. * Made the MIME type of the new event stream format be text/event-stream; removed mention of the old format's type (application/x-dom-event-stream).
1238ack for event-source changes
1239syntax error
1240make <figure>'s caption optional; define <figure> better; add examples
1241SQLTransactionErrorCallback can't return false to commit anymore. If it is called, things always roll back. This avoids the case where such a callback asked for a transaction to be committed, but that failed, and the failure information was lost.
1242Change the meaning of <ol> and <ul> subtly; include some examples.
1243Change the meaning of <dl> subtly; include some examples.
1244Give an example of a reverse-ordered list, and of a list with a caption in a figure.
1245Define the precise meaning of misordered <dl> markup
1246Clarify that <dl> <dt> A <dd> aaa... <dd> bbb... </dl> means aaa and bbb are alternatives, not part of the same definition.
1247copy/paste error
1248Add <ol reversed> by popular request.
1249Fixing typographical errors in examples.
1250allow <blockquote> in <header>/<footer>; add an example for <footer>; allow <footer> anywhere in its section (clarification)
1251Fix cross-refs to sectioning content; remove out-of-date reference to differences between xhtml and html content models for <p>.
1252<aside> examples
1253Add an example for <nav>; fix a note for <dl>.
1254typos in new examples
1255Introduce the concept of 'sectioning root', of which we have four: <blockquote>, <td>, <datagrid>, <figure>; stop <blockquote> from being a sectioning element
1256Rewrite the section on how to generate the outline for a tree. This should in theory not make any significant changes to the normative requirements.
1258Make the 'Determining which heading and section applies to a particular node' subsection redundant; fix up some cross-reference issues that were caused by the recent outline algorithm rewrite.
1259Oops, sectioning root elements don't reset the section level.
1260Changed the sections around to remove overuse of the word 'prose' an the word 'phrase'
1261Try renaming 'prose content' to 'flow content' to aid the (understandably) confused readers.
1262Clarify <header>; fix some places that still considered <blockquote> to be a sectioning element.
1263Make control characters and non-Unicode characters be parse errors, for compatibility with XML.
1264Make using a Win1252-specific byte when the document declared as ISO-8859-1 be a parse error.
1265Make the processing of '<' characters in attributes while doing the encoding-scan parse match the behaviour of the main parser. Also, various minor editorial fixes.
1266Oops, we failed to increment the position in some cases.
1267Clarify requirements for character encoding declarations
1268Clarify the meaning of 'section' in an outline
1269xref typo
1270correct various typos and omissions
1271ISO-8859-11 now treated as Win-874
1272Make UTF-16 turn to UTF-8 if the encoding is detected in an ASCII-compatible manner. Clarify some other things in the encoding detection algorithm.
1273make a note about the lack of trimming on encoding names
1274remove 'BOM' from the table of encoding names. add a note saying that encoding errors are still errors.
1275'character encoding declaration' is now a cross-reference term; made the content='' attribute of <meta> case-insensitive for charset decls. switched utf-8 and win1252 defaults around. other minor editorial jiggling.
1276Clarify the serialisation algorithm by not saying 'child node' so much.
1277Sigh. Apparently U+00A0, out of the million Unicode characters out there, is the only one who needs special treatment. Talk about a drama queen.
1278Make innerHTML on XML documents return a 'document' entity, and raise an exception if it has no element children. Raise exceptions for nodes that contain non-XML characters.
1279missed acknowledging new contributors when i did the character encoding changes recently
1280Further specify how to handle rel=stylesheet with Content-Type: 'null'.
1281Reiterate that HTTP can't just be ignored willy nilly...
1282Make the feed sniffer skip BOMs.
1283More accurately describe the risk of UAs not implementing the sniffing per spec
1284Only the first HTTP Content-Type header is examined.
1285make notes about some sniffing changes we might have to make
1286Changing &lang; and &rang; entities to new codepoints.
1287Leif pointed out some things he's said that directly affected the spec.
1288The sniffing gets worse: Content-Encoding no longer prevents sniffing, and text/plain UTF-8 must get sniffed as well now.
1289The first quirk: ignore Content-Type for rel=stylesheet
1290ProcessingInstructions in innerHTML
1291make it clear you can't have more than one BOM
1292<span ===> is now a parse error.
1293Recent changes have simplified the allowed attribute name syntax.
1294Make entities not get parsed between 'comments' in RCDATA elements.
1295Reduce redundancy in the writing section (editorial)
1296Fix the definition of ambiguous ampersand, and allow quoted attributes to end in ampersands.
1297Fix the definition of what is allowed in a comment to avoid allowing <->--> which wouldn't get parsed as expected. Also <!!--->-->.
1298doesn't much matter if </body> or </html> are followed by a space, given that spaces there will get hoisted up to the body anyway.
1299Ban '=' characters from unquoted attribute values.
1300add warnings regarding the serialisation differences
1301add note about adding entities
1302Add three more public identifiers to trigger quirks mode.
1303Ban attribute names containing single quotes and double quotes, ban unquoted attribute values containing single quotes and double quotes, require spaces between attributes.
1304missing space
1305DOCTYPE parsing: move the setting of the correctness flag to the places that call the bogus doctype state, so that we can later enter that state without triggering the correctness flag change. This change does not change the black box behaviour of the spec, but it can affect existing implementations that are tracking the spec.
1306DOCTYPE parsing: Make trailing garbage after the SYSTEM identifier in a DOCTYPE not trigger quirks mode.
1307DOCTYPE parsing: Rename the 'correctness' flag to 'force-quirks', since it is no longer about correctness. Add a comment to clarify that missing the public and system identifiers is not the same as the empty string.
1308Lay the groundwork for merging phases and insertion modes
1309Split the insertion modes into separate sections
1310deindent the source
1311Move the phase sections lower, move the insertion modes to the same level.
1312Merged phases and insertion modes. Theoretically, this should make absolutely no difference. Please let me know the many ways in which I screwed up.
1313Define 'insertion mode' earlier.
1314Make the processing of correct <html> start tags more explicit. Also, make a text node before <html> give a parse error, which it didn't before.
1315rename 'root element' insertion mode to 'before html'; readd the mistakenly dropped 'before head' insertion mode to the list of insertion modes
1316Remove a redundant entry in the parser; make base, link, and meta elements get popped correctly.
1317Note that <title> elements get parented in the <head> even when found elsewhere.
1318remove redundancy (<div> and <h1> are treated the same)
1319<button> and <select> need to be associated with <form> elements. Also, make the wording for <frameset> more consistent with other wording.
1320Change the algorithm for </form> to work the same as for </div>. Also some scary editoral work that theoretically didn't change anything: simplify the 'generate implied end tags' algorith. Reword some of the algorithms so that they don't sound like you have to look through the stack twice.
1321Simplify error handling in the face of foster parenting: we are no longer worrying about when foster parenting content comes off the stack as far as errors go.
1322clarify the 'reset the insertion mode appropriately' algorithm
1323markup error
1324The 'it' here really wasn't clear about what it was referring to.
1325Try to clarify what inserting something at the end of something means in the outlining algorithm.
1326Make spaces take part in the <table> foster parenting after anything else has taken part in it. This is a high-risk change, let me know if anything breaks or if I missed anything.
1327various clarifications in the after body and after frameset insertion modes.
1328Make the <title> tag simply get put into the current node like <meta> et al. This simplifies the generic (R)CDATA parsing algorithm a bit.
1329remove 'br' end tag processing rules here since it's already taken care of by the previous entry; make the 'anything else' invocation more consistent.
1330Make 'the title element' refer to any title element, not restricted to <head>, since we no longer hoist it up. Also, make <listing> parse like <pre> for compatibility with browsers.
1331Make <input type=hidden> not be taken out of <table> elements unless the <table> element has gotten tainted. (Basically, treat them the space as spaces.)
1332Strip spaces between <html> and <head> start tags.
1333Make consistent use of the 'insert an html element for the token' phrase.
1334Redefine how 'append a character' works, and rename it to 'insert a character'.
1335Let <style> and <script> stay in <table> elements.
1336well that made no sense. Fix the last change to actually do the right thing.
1337Tweak the way the insertion mode sections start.
1338remove a block that is redundant with the 'everything else' block
1339more consistently refer to the context element as such
1340Fix markup errors in the source.
1341New regen pipeline version.
1342Support <select> inside <table>.
1343fix a bug in the AAA to make it work even when doing foster parenting.
1344Make the terminology more consistent for the stack of open elements. (Is there something that grows downwards that we could use as a metaphor here instead of 'stack'?)
1345Clarify what it means to 'process the token as if' into another insertion mode
1346Make the places that switch the insertion mode use more consistent terminology.
1347Make <applet> act like <object> and <marquee>.
1348Revamp how end-of-file tokens work and how </body> is handled in conformance checkers.
1349Allow browsers to work around annoying authors who try to prevent copy/paste.
1350mention an example of a unique-origin data: URI
1351missing quote marks
1352remove the guidance that implies that em/strong are more important than i/b/mark
1353We made it so you could avoid clobbering selections earlier today.
1354Move the Command APIs section down to below the Selection APIs
1356no idea how i flubbed this
1357queryCommandEnabled() and friends -- this really requires heavy testing, please let me know of any mistakes
1358Fix the definition of contentEditable; clarify handling of non-editable sections inside editable sections.
1359Bold/Italic support in contentEditable.
1360Adding isContentEditable
1361Define how event listeners are to be ignored in designMode content.
1362Fix some bugs where an end-of-file token in the fragment case would either never terminate the parser or would terminate with parse errors that shouldn't be
1363The 'in table' fragment case still returned excessive parse errors. Also, fix a bit of the text in contentEditable to be clearer.
1364Make a note about the 'show' event for context menus -- it should include what the menu is being fired for. Also, fix a typo.
1365Make a note about context menus needing to inherit.
1366Define the createLink and unlink commands for execCommand(), and change the infrastructure in the spec to handle these commands.
1367Define insertOrderedList and insertUnorderedList; make note of some commands we'll have to add when we have a presentational story; change the way the command list is marked up.
1368Make execCommand() case-insensitive.
1369Add the insertImage command
1370s/doShowUI/showUI/, by popular request
1371define formatBlock more generally.
1372note for contentEditable v2
1373Abstract out the <td> and <th> attribute definitions.
1374validation typo
1375Add support for headers='' to the table model.
1376Tables: rewrite the table model algorithms to be zero-index-based instead of 1-based; add a bunch of acknowledgements for people who have done research on tables (the fruit of their labour will affect future checkins).
1377Smarter table header cell / data cell association algorithm.
1378Use a real 'minus' sign for negative numbers.
1379Fix some minor issues with the last checkin
1380Move all <tfoot> elements to the bottom of the table.
1381introduce createTBody().
1382markup typo
1383Deal with scope=colgroup cells that span colgroups.
1384Note char='' for future expansion.
1385Make intro sections non-normative; placeholder for a tabluar data intro section.
1386cross-reference error
1387Change the way we detect invalid rows.
1388Make table content models allow zero-or-more instead of one-or-more of most child elements (unambiguously makes the errors into table model errors instead of content model errors).
1389Change the way we define col-span, colgroup-span, colspan, and rowspan to make use of some predefined 'macros'.
1390what we need in the table intro
1391Typos, consistency.
1392capitalisation issue
1393Make <caption> processing more resilient to errors.
1394Redefine some table accessors to handle error conditions better.
1395deleteRow(-1) support
1396<td> can have headers='', not th, oops.
1397Add a link to the WHATWG Issues List to the WHATWG spec's header.
1398Clearer links to the issues list
1399A mechanism for embedding custom non-visible data in an HTML document for scripting purposes
1400Define the syntax for the upcoming SVG+MathML-in-text/html revolution.
1401Define where <svg> and <math> fit into the HTML content model.
1402i left a copy/paste trail
1403Move some of the tree construction state to before the tokeniser.
1404MathML and SVG support in text/html: the parsing infrastructure
1405Fix various editorial problems in the past few checkins.
1406Add support for about two thousand more entities (mostly to support MathML authors).
1407dataSet looks silly and is hard to type. Changed to dataset.
1408Remove a few entities that we don't need (they got added when I was a little overzealous in my merging of the entity sets).
1409Acks for the new entity system
1410Oops, forgot to support <mglyph> and <malignmark> in MathML.
1411The previous checkin wasn't meant to decimate the entities file.
1412Fill in two elements we know won't be allowed in SVG/MathML in text/html. (I'm running a more thorough study to determine what this list will have to be.)
1413Define the fixups for SVG in text/html. Let me know if I missed anything.
1414group the MathML stuff
1415Make it a parse error to have incorrect xmlns='' attributes.
1416Make the SVG and namespaced attribute fixup, and the xmlns value checking, also happen for <svg> and <math> elements.
1417xml:lang support in foreign content
1418Update the list of HTML elements that cause foreign content mode to bail. Let me know if anything should be added -- just show me an organic (non-test) URI with the element causing the problem.
1419How to handle MathML errors.
1420Allow <![CDATA[ .. ]]> in foreign content even in <mo>, <foreignObject>, etc. Just not in HTML elements.
1421ack Neil Soiffer since he was the first one to really propose a sane way to do MathML in text/html (and it is close to what we do now)
1422Acknowledge Jacques Distler, whose MathML blog was examined closely in the design of the MathML support in HTML5.
1423Clarify event handler attributes and how they work.
1424Encourage allowing MathML and SVG to be exported as XML.
1425ack Bruce Miller who first requested that XML export be made available in the UI.
1426Allow xmlns='' on any HTML element that is a child of a node that isn't an HTML element.
1427Add yet more examples of image alt='' text.
1428More alt'' text examples, courtesey of Sander and co.
1429Make the section on alt='' text even more clear that alt='' must not be omitted except in extreme circumstances.
1430Update the spec for consistency with WebIDL.
1431Mark DOMB issues in the source; clarify what setting 'src' on <img> means.
1432The image doesn't convey the cat's name...
1433More work on alternative text.
1434Require authors to include punctuation for <q>. Browsers should not render punctuation marks any more (the rendering section will make that clearer).
1435conformance error
1436Sort the various commands alphabetically.
1437missed one
1438execCommand('insertHTML') support
1439Add example for <dl> that uses <dfn>.
1440latin grammar typo
1441Resync the repository. Something crazy just happened. I'm fixing it.
1442Upstream entities update.
1443Mention the use of classes with <i>.
1444ack r12n for the last checkin.
1445ack r12a for the last two checkins. Sorry for the mess in recent checkin messages and things being mixed in from various places.
1446s/user/use/ typo
1447Remove SVG support from the parser based on a request from the SVG working group.
1448typos and rewrapping of examples
1450Define what happens when you have an unexpandable entity reference; also various minor editorial issues
1451Define document.embeds, document.applets; s/behaviour/behavior/
1452Define deprecated DOM attributes for <body> presentational attributes
1453Define document.plugins === document.embeds
1454mistyped the names
1455Find a better way to link to the 'master' interfaces.
1456Make it more obvious that HTMLElement is the actual interface name for these elements.
1457figure out what is missing still
1458<annevk> Hixie, elementFromPoint is
1459Prettier. So it is claimed, anyway.
1460Define document.charset, .characterSet, .defaultCharset
1461Define document.readyState
1462Define document.scripts
1463<meter> attributes must fit the inequalities.
1464Remove vestigial <label> element.
1465Fix the attributes on <meter> to be floats, not longs; and add mike to the acks (totally unrelated)
1466Remove the automatic cross-referencing for <dfn>.
1467Allow duplicate definitions; require that the definition be with the <dfn>.
1468Change <abbr>. It's only valid for actual expansions now.
1469Minor tweaks to <dfn>'s definition.
1470Introduce <object name> and <iframe name> to allow <a target> to load stuff in them.
1471Allow _blank, but encourage UAs to not open new windows.
1472Apparently everyone in #whatwg except me likes popups!
1473Notes on how to do footnotes.
1474thank the sources of examples
1475mark bad example
1476mark bad examples
1477Remove the term 'valid unicode character' since apparently it doesn't mean much
1478Allow <abbr> without title=''.
1479Include a lastEventId field on the event for <event-source>.
1480Ack for last checkin.
1481Looks like I got this backwards.
1482Make the <dl> conformance criteria less redundant.
1483Various comments on edits and paragraphs and lists.
1484Mention how to handle plurals in abbreviation expansions.
1485markup typo
1486Revamp tabindex=''; fix a bunch of uses of the word 'reflect' to be properly marked up.
1487Make postMessage() asynchronous, require its second argument, make '*' be a generic wildcard value.
1488Er, make the algorithm actually make sense in an async world.
1489trailing <s.
1490showModalDialog(): first draft. Also, some 'the the' editorial corrections.
1491postMessage() should fire on the window, not the document
1492The features argument of showModalDialog() is ignored. Removing it from the spec.
1493Various editorial changes to the local storage section.
1494Add Storage.clear(); define event dispatch for removeItem() and clear()
1495Define onstorage=''.
1496Note about handling nulls in the Storage API.
1497ack for storage.clear()
1499Move 'origin' section up one level, so that it can be made independent of scrpts
1500Revamp the way that the concept of 'origin' is defined; nuke the concept of 'same scheme/host/port'. This shouldn't, in theory, change anything, but let me know if I screwed up...
1501Tighten security a little: <img src='javascript:'> and javascript: in a style sheet should be in a sandbox.
1502Define document.domain and how it works.
1503Make document.domain return the empty string, not null, when there is no origin.
1504Define MessageEvent.origin for postMessage() in the face of IDN.
1505Minor editorial fixups in document.domain-related changes.
1506Support IDNA when setting document.domain.
1507Define the origin for about:blank documents.
1508Implement the descendant-navigation frame policy.
1509Define the origin of javascript: and data: URIs that were the result of an HTTP redirect (and minor editorial fixes)
1510Mention data: URIs explicitly, so we can give them a better origin. Also, safeguard mentions of the data: URIs from becoming xrefs.
1511<font> is gone; style='' is global.
1512don't mind me
1513Fallout from removing the WYSIWYG section.
1514Make things even more explicit.
1515Update the spec status to indicate that the spec is getting more stable (WHATWG only)
1516file:// interop isn't necessary, and UAs want to do clever things with it
1517Define origin for XMLHttpRequest documents -- this may move into the XHR spec at some point.
1518Also clarify the effective script origin for XMLHttpRequest documents.
1519Rendering text to a canvas: the fillText(), strokeText(), and measureText() APIs.
1520Take responsibility for not being any good.
1521Remove doubt that you can't use <q> to quote poetry... Sigh...
1522Allow <time> in <dt> in <dialog>.
1523Clarify what the pairs in a <dialog> are of.
1524Clarify the handling of <object> data='' and type='' and remove bogus requirements.
1525Support classid=''. Also, factor out the common parts of plugin handling for <object>.
1526Add examples for how to handle Java using <object>, add name='' to the list of <object> attributes.
1527Handle errors when handling <object> data better; sniff for the real image type if appropriate.
1528Allow type='' to be used as a hint.
1529Empty classid='' should apparently be ignored.
1530Make type='' take precedence if the server responded application/octet-stream.
1531oops, forgot to give name a DOM attribute
1532making it consistent. dunno whether it's the best value though...
1533Make @font-face fonts available to <canvas>.
1534placeholder for URL section
1535Update the abstract a little; minor markup fix.
1536don't have numbers without the constant names
1537Editorial fixes; make it clear what interacting while 'paused' means.
1538add a note explaining that referrers may be blanked for ssl->http connections
1539Remove <meta name=dns>, since nobody seemed to like it. If you actually like this and want it and will implement it, let me know.
1540Define <meta name=description> and <meta name=application-name>.
1541Allow UAs to raise a security exception for openDatabase() if the domain isn't allowed to have a database.
1542make note that we should kill things when calling
1543make some notes about what needs to happen with events, timers, etc, when a page is navigated.
1544Merge the name/value pair storage and SQL storage sections into one section with a common privacy and security bit, and other editorial fixes.
1545Skip calling a callback if it wasn't specified.
1546fix xrefs for executeSql()
1548make a note of an open issue with executeSql() arguments and types that the backend doesn't support
1549Clarify that only vector fonts are intended to be used.
1550Minor editorial changes to the application cache section to fix mistakes and clarify how things work.
1551xref typo
1552grammar typo due to earlier search and replace
1553Remove the untestable mention of U+0000 in application caches. Make it clear that the cache API returns URIs. Grammar fixes.
1554make a note that we haven't said how to resolve relative urls for the application cache api
1555More issues with URIs -- add more things I need to define, make the application cache api raise errors on invalid uris
1556also note that we have to resolve urls for remove() in appcaches
1557note to self so i don't spend 3 minutes working out why it's not a normative requirement again...
1558<link rel='icon' sizes=''>
1559add an example for rel=icon
1560xref so i can refer to this section
1561markup typo
1562Make parsing lists of numbers have some semblence of an attempt at parsing actual numbers correctly.
1563Bring the coords='' parsing more in line with how IE does it.
1564make it read better
1565Fix the list of numbers algorithm to handle '1,-' and to not be redundant.
1566oops, when we did the whole 'flow' change some old inline elements became allowed where they weren't before, with unforseen results for the datagrid content model
1568Excise the term 'entity' from the specification (as far as the text/html syntax goes). They are now 'character references'. You put them on your resume.
1569minor wording fix
1570grammar typo
1571add a placeholder for lists of elements, attributes, interfaces, events (anything else?)
1572Make boolean attributes case-insensitive
1573minor editorial fixes
1576more typos
1577Explain slightly better how <link> can create multiple links
1578Try to make it more precise when a UA has to download a resource in a <link> with no type=''.
1579er, this made no sense.
1581explain that rel='up up index icon' is two links, not four (or three); fix another typo
1582Require preprocessing of media queries since CSS doesn't recognise the vertical tab character.
1583Clarify where elements can be used; explicitly allow orphan nodes.
1584Reword all conformance requirements never to use the phrase 'must only', which really doesn't mean what you think it means if you apply RFC2119 definitions precisely.
1585Give 'may only' the same treatment as 'must only'.
1586...and 'should only' goes byebye also.
1588Make HTMLAreaElement.shape not accept invalid values.
1589Clear some stuff up, since this wasn't well written before.
1590Big cleanup of the 'reflect' definitions.
1591Make DOMTokenList not put spaces at starts of strings, even for toggle().
1592Make &quot; not be escaped in non-attribute contexts in innerHTML.
1593Make the definition of compatMode less confusing.
1594ack for prev checkin
1595Writing cache manifests: clarify that the BOM is allowed and won't cause ignoring of the file.
1596Make the implicit downloads delay the cache update process. Allow downloads to be reused for cache update. Abstract out the failure mode of the cache update process. Drop fragment identifiers for fallback documents too.
1597Move the history traversal algorithm out of the History object section
1598Application caches: Define what happens when the user cancels the download. Resolve the infinite loop when the server is broken but a fallback resource doesn't have a valid manifest= attribute.
1599Define what to do when you get a new implicit entry to add but the cache is up to date. Rejig some of the behaviour around to be slightly safer (and to prevent the cache update process from being invoked from withint the cache update process).
1600make spelling of labeled consistent.
1601reword 'presentation concerns'
1602Mention that <div> is allowed in Atom as well.
1603Mention <meter> in the <progress> section.
1604Restructure the script execution algorithm to be more clear. Fix a cross-reference problem in the appcache section.
1605Minor editorial fixes in the script section.
1606<script charset> support
1607Remove redundant requirement (this was already in the content model, oops)
1608Oops, the default script type when there's no type='' nor languauge='' on the <script> wasn't defined.
1609mention default values for when type='' and charset='' are omitted.
1610Support language=javascript1.1 to 1.3 as being regular old javascript.
1611Rework the <script> element's semantics to allow it to be abused, er, used, for smuggling proprietary data.
1612Undo the change 'Support language=javascript1.1 to 1.3 as being regular old javascript.' since it turns out the previous text already did that.
1613Fire 'load' events on <script> after running the script, not before.
1614Some introductory comments on scripting for authors.
1615remove wrong note
1616fail is the wrong word
1617Note the request for external captions; move <source> to before the definition of media elements.
1618acks for last checkin
1619Allow <video> elements to show the poster frame instead of the first video frame.
1620ack for last checkin
1621Updates for details of how to use <source>.
1622Change the 'begin' event to 'loadstart'; clarify firing of events for .play()
1623Clarify firing of events for .pause()
1624Removing a video element that has no video shouldn't pause it.
1625off by one error
1626Define the order of events during the seeking algorithm better; add 'seeking' and 'seeked' events to the media element event extravaganza.
1628make a note about what the spec should say about height/width of controls on videos
1629<source pixelratio=''> support (based on feedback from YouTube)
1630video.bufferedBytes and video.totalBytes support (based on feedback from YouTube)
1631misuse of rfc2119
1632video.bufferingThrottled support (based on feedback from Google Video)
1633xref effective script origin; make sure e.s.o. of URI is defined.
1634Mostly editorial: Introduce the concept of 'same origin' as a defined term of art; cross-reference 'origin' throughout; minor disambiguations around the term 'origin'.
1635Defined the 'script execution browsing context' and the 'source browsing context'. and window.showModalDialog() should use the script execution browsing context, not the browsing context of the Window object
1637<iframe seamless> -- clientside inclusion of HTML documents. People have been asking for this for at least ten years.
1638Fix all the typos, cross-reference errors, dandling links, and markup errors of the last checkin...
1639Oops, missed one of the requirements. (credit: roc, iirc)
1640Fix an obvious error with the definition of reversed=''. (credit: mikesmith)
1641Add a warning to make sure the seamless='' attribute doesn't get misimplemented to apply cross-origin (credit: lachy)
1642Reduce the number of people who are credited more than once.
1643* Add <iframe sandbox>, which supports: - isolating frames in their own unique origin, - disabling plugins - disabling forms - disabling scripts - disabling navigating of other browsing contexts - disabling popups and modal dialogs * Make the spec well-defined in the face of browsers that disable scripting on a per-browsing-context basis, and use this for the concept of sandboxing scripts. * Abstract out the concept of scripting being disabled, so that we don't have to refer to designMode all over the place. * Add a note that <noscript> fails pretty badly with designMode. Add and update some XXX notes in various places. * Require that Henri do the impossible by checking that scripts never cause non-conforming states to occur. * Require that sets of space-separated tokens with lists of allowed values not use values outside of the allowed values.
1644Move 'flag' into the names of these flags.
1645experimenting with different ways to separate major blocks in the definitions of some elements, without using subsections
1646Clarify that non-automated conformance checkers could exist and that scripts can cause non-conforming states to occur while they execute.
1647Abstract out the concept of third-party renderers into the term "plugin" and put all its related text in one place. Also, quadruple prose/example/note/warning combo! 20 points.
1648Let's try getting this named character references table to work...
1649Rename the sandbox='' values to be more approachable.
1650xref error
1651Block document.cookies if the page is sandboxed. Add an example of sandboxing.
1652Uh. Forgot the sandbox attribute... I guess that's the biggest security risk with this... (also, seamless DOM attribute was the wrong type on the interface)
1653Oops, I put the bit about cookies under the wrong flag heading.
1654More examples for the new features.
1656typo (credit: avk)
1657i suck
1658Revamp and clean up the concept of script execution contexts.
1659Add height/width to <iframe>.
1660Make reset the character encoding to UTF-16. Make about:blank use UTF-8. (credit: bz; mozilla bug 380383)
1661Encoding aliases for CJK environments.
1662ack for last checkin
1663editorial stuff in the chararcter encoding algorithms (credit: pt)
1664run a spellchecker...
1665Content-Type parsing should stop on a semicolon.
1666Apparently people mislabel Win874-encoded Thai documents as TIS-620 all the time
1667Try to fix the encoding sniffing algorithm...
1668Make entities not be allowed to use non-unicode characters
1669Fold the encoding alias rules into one table, since it turns out the -11 thing was already there and I didn't notice...
1670fix up refs
1671Make encoding name normalisation part of the spec. ASCII spaces and punctuation get ignored. Control characters and characters above U+007F don't. ASCII text is case-insensitive.
1672Outlaw <meta http-equiv=content-type content=t&#x0065;xt/html;charset=utf-8>. (credit: hs)
1673More violations of charmod...
1674Bring things more in line with the requirements of actual web content
1675update <listing> to include a newline on serialisation
1676Add support for sniffing .ico files
1677reduce use of the word 'valid'
1678tweak the wording on the bit about xml and its optional features
1679Improve quirks mode selection algorithm based on data. (credit: sp, pt)
1680Woops, forgot to keep this section in sync. (credit: sp)
1681The syntax section didn't take <h1>..<h6> into account, oops.
1682Spaces are stripped, so it doesn't matter if that's the first thing in an <html> element.
1683c/p error
1684EOF should be explicitly listed in each tokeniser state.
1685Reduce the number of errors in truncated doctypes
1686Bring end tag processing at the top of the document more in line with IE, for more compatible <frameset> handling.
1687ack for last credit; remove a red note from the spec based on testing
1688handlers? what's that? oh you must mean plugin. (credit: ms)
1689Make 'missing' and 'empty' even more different.
1690Make innerHTML on <td> and <th> work.
1691Factor out foster parenting to ensure it is not ambiguous.
1692Allow <noframes> in <head> for compatibility with sites (e.g.
1693Handle end-of-file tokens in foreign content. (credit: avk)
1694Add <h4> to the bail-out tag list. (I can count! I swear!) (credit: sp)
1695Make things more consistent wrt 'create an element for the token' algorithm invocations.
1696Always set the secondary insertion mode when switching to foreign content... (credit: hs)
1697ack the self-closing flag on isindex in order to avoid an extra error and to keep test cases and errors reported in conformance checker UI closer to each other (credit: hs)
1698Make it easier to optimise the parser by adding certain elements to the foreign content escape hatch. (credit: hs)
1699Allow <textarea> to bail out of <select> for compatibility with legacy content. (webkit bug 17421)
1700Require that if a <meta> charset is included, the encoding be ASCII-compatible. Update the list of SVG attributes (commented out) to work around errors in the SVG spec...
1701Shun UTF-32. Make it slightly clearer what 'UTF-16' means.
1702Remove a note that was causing more confusion than it was helping.
1703Make a decision regarding how block level elements interact with <p> elements.
1704<ruby> support. This isn't very compatible with IE, but it seems to be relatively compatible with the Web's usage of ruby.
1705Remove redundant note. (credit: jh)
1706Discuss the difference between rel='' on <link> and <a>.
1707Forgot the image.
1708Bring the definition of rel=nofollow more into line with the way it is used.
1709Make data-='' invalid.
1710typo (credit: pt)
1711Commenting out rel=contact since it clashes with XFN.
1712Make rel=icon use the last icon for compatibility with legacy UAs.
1713Mention that for rel=icon UAs should also use type=''; mention what to do when an icon sucks; mention how to handle missing sizes=''.
1714showNotification() -- first draft.
1715typos with showNotification() (credit: pt)
1716Define .focus(), .blur(), onfocus, and onblur in a way that doesn't require an infinite loop in the face of a hostile author, and that is better defined in terms of multiple iframes, windows, and orphaned elements. Allow UAs to ignore .blur(). Fix up the mess around the body element having focus.
1717s/null/no/ for namespaces; xref typo
1718fix duplicate id (credit: hs/hvn)
1719Remove some simplicity from the draft.
1720scrollIntoView(): turn a must into should, mention horizontal scrolling.
1721hasFocus() is a method, not an attribute. (credit: ap)
1722Make <map name=''> the preferred way of doing things.
1723oops, a vestigial mention
1724onbeforeprint and onafterprint
1725Examples for previous checkin, more notes regarding printing.
1726Many editorial changes around the word 'base', from refering to RFC3986 more directly, to changing 'base ten integer' to 'base-ten integer'
1727Oops, thinko wrt the event names. (credit: avk)
1728Change 'hashed reference' to 'hash-name reference' and update its conformance requirements to require a name attribute.
1729Allow xml:base explicitly.
1730note that we need to resolve the issue of dynamically changing base href
1731add support for the new html elements. fix the handling of optional tags we added recently. also: clarify some notes, remove redundant requirements, clean up some punctuation.
1732typo (credit: de)
1733missing 'include' (credit: de)
1734Define 'computed value' for things not in a tree (mostly affects canvas stuff).
1735put XXXDOCURL markers wherever the term 'the document's address' is present, for easy finding later when i go and define that somehow.
1736bring a bunch of user interaction stuff into one section
1737oops, forgot to update this section while I was at it
1738Remove U+000B LINE TABULATION (VT) from the list of characters treated like a space character in HTML, for consistency with CSS. This is incompatible with IE7, but apparently IE has all kinds of complications here and in any case IE8 doesn't treat U+000B as a space character.
1739define the rounding behaviour of createImageData() to be like that of getID() and putID().
1740Specify as a CanvasPixelArray.
1742Cross-reference 'spaces'.
1743Clarify the meaning of buffered and bufferedBytes. (credit: pj)
1744Make the handling of <video> intrinsic dimensions clearer.
1745Clarify relationships with other W3C specs.
1746Clarify that pixelratio doesn't affect poster frame; add v3 idea.
1747what is poster=''?
1748Clarify this text by having some extra steps and a temporary.
1749examples and xrefs in the outline section
1750duplicate dfn/xref
1751active documents are of browsing contexts, not Window objects
1752Make shadows required.
1753Clean up the security processing model for <canvas>.
1754Add a note about how origin-clean never gets reset.
1755make the intro paragraph easier to understand
1756Clarify arc() for arcs greater than 2pi.
1757Define behaviour for arcTo() with r=0.
1758remove some redundant ImageData stuff
1759Oops, an earlier change forgot to define the start/end points of the arc for whole-circumference arcs.
1760Make the spec more consistent about Infinity and NaN values throwing NOT_SUPPORTED_ERR.
1761keep variable names consistent
1762The example got updated but not its explanatory text...
1763Make it clear there are two putImageData() methods with different amounts of arguments.
1764Try to get closer to browsers. Slightly.
1765I suppose we should mention all the variables...
1766make innerHTML on XML documents not do the xmlns='' thing.
1767Make gradients transform at the right time.
1768Remove redundancy in gradient extrapolation.
1769Clarify the meaning of 'butt'. (re: Canvas line style comments)
1770Define 'miter limit' better. (Re: Canvas line style comments)
1771toDataURL() arguments for image/jpeg. (Re: <canvas> JPEG quality)
1772Make computed values in the Canvas API take effect on setting, not on use. (Re: [whatwg] are relative values of CanvasRenderingContext2D.font live to style changes?)
1773Make poster='' affect the intrinsic dimensions of <video>. (Re: [whatwg] Interpretation of video poster attribute)
1774Jiggle height/width around so that width comes first, and a minor correction... (credit: pt)
1775document.cookie is underdefined
1776Clarify that we are specifying black-box behaviour, so it doesn't matter how the UA achives the goal.
1777notes for urls, and typo fixes
1778A bunch of syntax and parsing fixes mostly to fix omissions relating to the new elements. (bug 5759)
1779minor groundwork for upcoming url overhaul
1780work in progress along the path to writing the URL section
1781Fix an error in the outlining algorithm that broke outlines for documents like <body><h1/><h2/><h3/><h2/>. (credit: jg, gs)
1782more notes for the upcoming URL revolution
1783more notes for the upcoming url rework
1784Revamp the way that type='' on <link> is defined to actually be accurate. Allow sniffing when the expected type is an image.
1785Allow sniffing when the expected type is an image. (continuation of changes for previous checkin)
1786more notes on URLs
1787Define what happens when you set currentLoop. (bug 5769)
1788yet more mindnumbingly boring preparation work for the upcoming URL reorg
1789final bunch of annotations for url preparation
1790Giant reorg to move things to more appropriate places so that there's a logical place for a URLs section. Please let me know what you think.
1791Define 'URL' and 'valid URL'. Define 'script document context'. Introduce the term 'script browsing context' for the existing concept of 'browsing context associated with a script execution context' and use that term where appropriate. Define origin of onfoo='' scripts. Fix xref errors. Minor editorial fixes.
1792Hm, that didn't work well. Let's try just smashing it together for now.
1793Define 'valid URL' as an IRI reference and 'valid absolute URL' as an absolute IRI.
1794Remove the paragraph that says 'URI' means 'IRI' since we'll just use 'URL' from now on.
1795Define how to resolve a relative URL. Work in progress; not yet integrated with the rest of the spec.
1796Factor out the URL parsing rules and use them to define the URL decomposition attributes.
1797Use 'URL' in various places that used to use URI/IRI. Incomplete; more to come.
1798More work replacing references to 'URI' with references to 'URL
1799Make the 'navigate' step resolve URLs. Make the 'navigate' algorithm more precise and correct and consistent, in particular regarding fragment identifiers. Fix some xrefs in the 'navigate' algorithm. More URLing. Add a placeholder for defining the term 'fetching'.
1800<object> element: bring some old prose in line with future plans regarding 'fetching'.
1801fix markup error, move bookmark down for URL work
1802URLify the 'address of the chosen media resource' concept and related parts.
1803fix blasphemous typo
1804More URLification (<canvas>, <script>, <event-source>).
1805URLification: mention that the term 'URL' isn't the same as 3986 would have us believe; URLify <datagrid> and the main 'command' section.
1806URLification of commands. Made other aspects of commands more consistent while I was at it.
1807URLification of templates,
1808URLification of origin.
1809URLification of javascript: URL section
1810URLification of arguments and attributes in IDLs
1811URLify the registerContentHandler() section. This also modified the rules for handling relative URLs and for escaping content here, it wasn't just s/URI/URL/ as other commits have been.
1812URLification of offline application cache and of session history.
1813URLify Location.
1814URLify navigation and neighboring sections; fix validation errors
1815Urlification of the 'links' section.
1816remaining URLification
1817fix references to RFCs 3986 and 7
1818remove the open issues section for URLs
1819make <base href> not depend on xml:base
1820xref typos
1821Define what should happen when the base URL changes, keeping it as lightweight as possible.
1822Allow reporting bogus URLs as an error.
1823Make backslashes turn into forward slashes when parsing URLs. Sigh.
1824Make the \ -> / conversion not happen in mailto:, data:, etc.
1825If a character in a URL's query component can't be expressed in the page encoding, replace it with a question mark (matches IE and Opera)
1826Bring URL decomposition DOM attributes more into line with IE.
1827Make UTF-16 act like UTF-8 for query encoding of URLs.
1828Add an xref for philip.
1829Ack for the note about the term 'url'
1830'the script that will execute as soon as the parser resumes' is now the 'pending external script' (bug 5787) (credit: hs)
1831Now that <p> is dealt with after looking for an <li> element, make sure we don't stop at <p>. (bug: 5796) (credit: hs)
1832Make the language around parse errors and ignoring tokens more consistent. (bug 5792) (credit: hs)
1833tokens aren't elements...
1834Allow xmlns='...' anywhere in HTML. (bug: 5801) (blame: hs)
1835Don't escape '%' when doing URL resolution. (bug 5802) (credit: hs)
1836Take all the fun out of <embed> and data-*. (bug 5809) (credit: hs)
1837Make <svg:foreignObject> scoping. (bug 5815) (credit: hs)
1838Comment out the parts of the previous checking that don't apply yet since SVG parsing is disabled.
1839Allow the origin to include extra data, such as the host's certificate. (credit: el)
1840Web Socket specification - replaces TCPConnection and related interfaces. Also split 'serialisation of an origin' into two algorithms, one for unicode and one for ascii. And some minor editorial things.
1841xref typos
1842onclose: Event name was inconsistent with event attribute handler name.
1843Handle TLS errors in the Web Socket Protocol. (and minor editorial fixes)
1844Oops, got my maths wrong.
1845Be clearer about what should be echoed as part of the handshake.
1846Make the length encoding big-endian to be compatible with Perl. (credit: pt)
1847Oops, I'm just bad at maths. Also, define how UAs must fail in the face of large data. (credit: pt)
1848Make HTML5 agnostic regarding Cookie and Cookie2. (Re: Web Sockets) (credit: ns)
1849Make the Web Socket Protocol handshake more unique. (Re: websocket HTTP response parsing) (credit: pt, jr)
1850Define handling of poster frame on <video> in more detail.
1851Make <video> and <audio> have a default volume of 1.0 instead of 0.5. (Re: Consider changing the default volume for <audio> and <video> to be 1.0 instead of 0.5) (credit: ap, pj, cd)
1852notes on fetching
1853Try to clean up the 'pixelratio' stuff.
1854Clarify that 'pixelratio' doesn't affect later <source> elements; clarify that load() doesn't wait for earlier invocations to abort. (Re: HTMLMediaElement: more issues and ambiguities) (credit: pj)
1855loadstart is also synchronous (Re: HTMLMediaElement: more issues and ambiguities) (credit: pj)
1857Update WHATWG abstract; update W3C header to match pubrules better; fix links.
1858xref error (duplicate term definition)
1859oops, defined CanvasPattern twice.
1860more typos (credit: gs)
1861Messaging: introduce placeholders for pipes; make all the messaging systems use 'message' and MessageEvent; rename 'server sent dom events' to 'server sent events'.
1862wrong header level
1863Rename <event-source> to <eventsource> for consistency, while we sill can
1864Fix spelling of tokenisation to be american. Sigh. (Re: 8.2.1: tokenisation) (credit: db)
1865Content-sniffing in the unknown type case used the wrong index to check the pattern. (Re: content type sniffing - unknown type) (credit: rk)
1866Content-sniffing for the unknown case had a test backwards. (Re: content type sniffing - stream length) (credit: rk)
1867Define the Pipe constructor; do other maintenance work around constructors.
1868typo fix (credit: pt)
1870Split the pipe section up
1871typo (credit: avk)
1872More work on Pipes.
1873Define Message Channels and Ports. Also, factor out the XXXDISCARD stuff into its own section.
1874Define the three-argument form of postMessage().
1875Define what happens when you call the postMessage() method with a null port.
1876Correct example for postMessage(). (credit: avk)
1878xref typo
1879Split Window into two (editorial-level) definitions. Clean up some timeout stuff.
1880bring more of the interfaces that AbstractView needs to implement into one place
1881Now that we've split Window in two, there are some places where we need to refer to the other interface.
1882Add 'load' and 'error' events to message ports.
1883Fire 'unload' on a message port if it is .close()d or if its other end is garbage collected.
1884Making garbage collection fire an event was a stupid idea. Let's only make an explicit .close() cause unload to be fired.
1885Make the URL easier to remember.
1886xref fixes
1887typo (credit: gs)
1888typos, xref errors (credit: gs)
1889Move some APIs to Window so they are accessible in Workers.
1890add issue markers for sending structured data over postMessage
1892oops, didn't define the menu DOM attributes
1893Make :enabled and :disabled available on <menu><li> elements.
1894First draft (very experimental) of <bb>. Not sure if we'll want to keep this.
1895Stop using <em> for things other than emphasis. Remove title='' from <em> elements.
1896Make the handling of websocket-origin stricter
1897oops, misnamed the interface
1898add a space to make it not empty
1899Make <bb> define a command.
1900minor fixes for <bb>
1901Note how confirmation could take place. (credit: roc)
1902Make addCueRange() have an identifier so that people don't have to use currying. (Re: re-thinking 'cue ranges') (credit: ds)
1903Revamp of the intrinsic dimensions and pixel ratio definitions. (Re: HTMLMediaElement: more issues and ambiguities) (credit: pj)
1904Terminology cleanup near the term 'CDATA'
1905Clarify the bogus comment state. (Bug 5863) (credit: as)
1906Make the tokeniser states into sections for easier navigation. (Bug 5881) (credit: as)
1907Provide a way to mutate the DOM into an infoset. (Bug 5808) (credit: hs)
1908typos (credit: ms)
1909Change how elements are treated in the after after frameset insertion mode. (Bug 5858) (credit: as)
1910Make the coercions section not invent a new syntax. (Bug 5808) (credit: hs)
1911clarify that it's capital letters of hex...
1912tidy up the messes that #whatwg found overnight -- thanks
1913Spec out the behaviour we see in IE for navigating to about:blank or javascript: URLs vs other URLs.'') needs to be like
1915make some notes about window.location near about:blank
1916Fix the spec to use the [Constructor] syntax from WebIDL to define constructors.
1917fix the definition of img.width/height for the case where the image isn't being rendered but is available (bug 5856) (credit: avk)
1918note case-sensitivity issues
1919Add a typographic conventions section; move the terminology section up and split it into subsections.
1920fix markup errors, add images that had been forgotten
1921more typos
1922tweaks to the typographical conventions section
1923oops, wrong heading level
1924Define window.parent. (credit: cm, mjs)
1925Resolve the issue markers in the <embed> section.
1926make <embed> do extension sniffing, for compatibility with servers that send flash as text/plain, sigh. See Mozilla bug 347736. (credit: cb)
1927Make content-sniffing 'better': make the text/binary case actually work out what the binary data might be; make the unknown type case determine the text/plain cases as a first-class citizen instead of falling back on the text/binary algorithm; fix minor grammatical things.
1928make notes about craziness of <embed>
1929Support <embed> with no src=''.
1930copy/paste error (credit: avk)
1931copy/paste error (credit: avk)
1932Define <area alt> authoring conformance criteria.
1934clarify that changes include removal for <object> processing model
1935Make application/octet-stream with <object> with no type='' trigger sniffing.
1936Case Sensitivity Training
1937Define order and case-sensitivity of getElementsByClassName().
1938Since we lowercase attributes anyway, no need to force them further to lowercase in the Node.nodeName etc. (Bug 5898) (credit: sp)
1940A colon for zcorpan.
1941Remove localName from the case quirkyness since IE doesn't support it anyway. Remove some redundant or inappropriate requirements, editorial notes that no longer apply, and editorial fixes. (credit: hs)
1942require title='' for rel='alternate stylesheet'
1943Allow connection to be closed when receiving entity body in ping='' response
1944Make document.write() and document.writeln() support zero and multiple arguments. (credit: avk)
1945Make data-* always lowercase.
1946Make embed attributes always lowercase.
1947change the way ids are given for area shape=''
1948I don't think this pipe dream could be made to work on the average site... :-(
1949Minor fixes to the datagrid section.
1950Extend rel=noreferrer to also blow away the 'opener' when used with target=_blank.
1952Allow frame busting. (also, remove redundant requirement)
1953Make height and width DOM attributes be unsigned.
1954Allow Lynx-like behaviour for <img>.
1955Move the alt text requirements into their own section so that it is easier to find them.
1956Cleanup around <img> requirements. Disallow scripts to run in <img> images.
1957More cleanup around <img> elements and images: Make the sniffing more predictable (and maybe more secure), make it clear that <img> does sniffing, minor editorial fixes.
1958More notes on what is a valid image.
1959move text around the img section
1960Another (probably vain) attempt at defining what an <img> element represents in all cases.
1961ack for some of the recent img stuff
1962duh, no do
1963experimentally try to allow <a> to contain any flow content in certain cases...
1964Fix the content models to handle <a> containing flow elements without breaking the definition of paragraph, and remove redundant requirements for ins and del along the same lines.
1965Make <a> elements not define paragraphs, allowing them to straddle paragraphs.
1966Clarify what <header> means for document summaries and ToCs, and define what to do in case of missing <hx> elements.
1967Allow implementations to use other IEEE rounding modes if really necessary.
1968Make scrollIntoView() into a no-op if the page is too short.
1969clarify where to bail if you want to ignore the navigation for some reason
1970Rearrange sections around so that related topics are closer together.
1971Define fetching. (bug 5848)
1972Try a new way of handling images that the author doesn't know much about.
1973Change <dl> to sections for convenience.
1974Add a CAPTCHA example to the <img> section.
1975Tweak headers
1976Address the case of a graphic containing only text.
1977ack for last checkin
1978More cases of images for authors to worry about (credit: pt)
1979More slicing examples.
1980subtly mention <meter>...
1981ack for previous checkin
1983highlight images in examples
1984highlight images in examples
1985forgot to escape tags; add a label element in an example
1986typo (credit: pt)
1987typo (credit: sp)
1988Changes to fix the problems Lachy found. (credit: lh)
1989typo (credit: lh)
1990more typos and such errors (credit: lh)
1991example of {...} in non-english (credit: ds)
1992Let's use a language where there is a clear word for 'photo' that doesn't sound like 'photo'... (credit: hs)
1993New feature for noticing updates.
1994Increase update interval to 30 seconds.
1995Fix CSS error.
1996Fix CSS properly; make debugging this script easier.
1998define what happens when you set onclick='' on a document outside a Window
1999define javascript: in Window-less environments
2000Move some paragraphs around so that the basic definitions come before the ones that control script execution. On an unrelated note: Two thousandth revision! Woohoo!
2001Define 'directionality' in terms of the dir='' attribute for cases where the 'direction' property has no computed value. (credit: db)
2002Define processing for the second argument to getDataURL() for image/jpeg.
2003typo (credit: pt)
2004How to handle transparent images for toDataURL(). Just allow the UA to premultiply the alpha and output that whie dropping the alpha channel, instead of worrying about the background color or whatnot. It's up to authors to make the background color non-transparent.
2005Define isPointInPath() uses the non-zero winding number rule. (credit: pt)
2006Fix validation errors.
2007Make the update ui more obvious.
2008Make support for patterns required in the <canvas> API.
2009Simplify the text around getContext()'s arguments.
2010Remove a redundant requirement.
2011typo (credit: pt)
2012typo (credit: pt)
2013Linkify 'void elements' more.
2014Clarify the paragraph that defines how to handle type='' on <script>.
2015Mention text/javascript1.1 et al.
2016When <script type=''> is given, it must match the type of the script, even if the script is JS. (credit: hs)
2017What .detail attribute? Event objects don't have one.
2018Revamp the 'interactive content' definitions so they make sense.
2019Remove autosubmit='' from the <menu> element.
2020Make MessagePort objects not be owned by Windows necessarily, and remove ownerWindow.
2021Merge WindowBrowsingContext and Window together again.
2022Define localStorage in terms of script origins, not script browsing context origins.
2023Simplify message ports: use queueing instead of transient 'active' functionality. Also, make localStorage use the same mechanism for obtaining origin as openDatabase().
2024Simplify garbage collection for ports even further. Define dicarding of Document objects better for ports. Prevent inactive documents from receiving messages.
2025fix typos
2026typos in the WebSocket section
2028Add 'close' button to updates notification.
2029'close' button should work until the next revision.
2030Show diffs link too.
2031MessagePort should implement EventTarget.
2032Remove redundant event listeners on MessagePort.
2033Removing note -- bz is probably right, but we've had even more feedback saying we should not be adding extra differences between serialisations, so punting on this for now.
2034Define outerHTML for HTML elements in HTML documents.
2035Handle outerHTML on elements inside DocumentFragments, because I'm stupid and zcorpan was paying attention.
2036Make it clearer that document.write() and outerHTML don't work in XML documents.
2037prettyifying the section regarding dynamic markup insertion
2038xref error
2039Prepare the ground for insertAdjacentHTML().
2040Define insertAdjacentHTML().
2041Require javascript: execution to be async.
2042No need to call js: out explicitly now that it's async.
2043Add a startConversation() method that wraps new MessageChannel and postMessage() in a less expensive one-step operation. (credit: ab)
2044One area with lack of interop is whether scripts run while alert()s are up. Let's disallow it, to get more predictable behaviour.
2045Actually the previous checkin was redundant with the existing definition of 'pause'.
2046Require that <script src=javascript:...></script> result in no script execution, for any value of '...', for compatibility with most UAs.
2047Clarify what 'using the rules for' means.
2048A sample application cache manifest.
2049Make execCommand('insertHTML') not work in XML documents, since it uses the HTML parser. (Bug 5946) (credit: sp)
2050Allow 'muted' to be stored across sessions; 'require' it to initially start off as not muted.
2051Make the application cache manifest syntax more extensible. (Bug 5937)
2052The 'find a number' algorithm on the string '.' would fail unexpectedly. (Subject: Find a number with . as the input) (credit: hs)
2053Require all <meter> elements to have _some_ value. (Subject: meter and document conformance) (credit: hs)
2054Look for a '%' character after spaces after the second number when looking for an error, for consistency with the earlier check. (Subject: Skipping Zs in finding a ratio) (credit: hs)
2055c/p error (credit: hs)
2056How to determine the language of the node if the language code isn't recognised. (Subject: Effect of bogus lang on determining natural language) (credit: hs)
2057Define the Content-Language pragma, since apparently ~1% of sites use it in some way or another.
2058Make it clear that xml:lang is namespaced.
2059type (credit: avk)
2060more typos
2061Make CanvasGradient use the 'octet' type, and other minor fixes. (credit: #developers)
2062Allow lang='' and xml:lang='' to be specified together in XML, if they are the same.
2063Allow xml:lang= in HTML in certain very specific cases to ease migration.
2064Tighten up's type argument processing.
2065Mention that SVGDocument.title can be readonly (as indeed it is right now).
2066Define order for getElementsByName().
2067Move things around in the IDL so that browsing-context related stuff is closer together.
2068Define window.frameElement.
2069Don't require Document.location to do anything when the Document isn't in a Window.
2070Define a set of properties for browser detection.
2071Refine the rules to allow for more privacy.
2072Change handling of document.write()'s second argument to be safer. (Subject: type parameter of (credit: bz)
2073Setting up the spec for defining the event loop mechanism.
2074Define event loops, task queues, etc; Make 'fetching' use this mechanism (everything will in due course); Fix some cross-references around 'interactive content'.
2075Redefine img.complete in terms of tasks; Make it clearer why 'timeupdate' is fired when seeking; add more XXX notes for queue-related fixes that are coming soon.
2076Tasks for the media element new resource task source.
2077Add a history section.
2078typo (credit: avk)
2079Make it so that multiple pending load()s are coallesced into one. (credit: pj)
2080More poking around getting the media element section to use the new event loop stuff.
2081Further work on the event loop front. Also, make 'will' link to the actual requirements in some more cases.
2082Further work on the event loop front.
2083Further work on the event loop front. (This should complete the bulk of the changes required for the media elements section.)
2084Further work on the event loop front. (<script>, database API, remote events, and other bits and bobs)
2085Further work on the event loop front. (WebSockets, postMessage, MessagePorts, setTimeout)
2086The 'unlink' command should be disabled if there's no link. (Subject: HTML 5 - comments on 5.6 Command APIs)
2087typo (credit: avk)
2088Clarify that we don't currently define the 'user interaction task source'; hopefully DOM3 Events will handle something like this eventually.
2089Note that removing a content attribute doesn't affect a the event handler. (credit: tl)
2090Define onhashchange=''.
2091Allow xml:lang='' and lang='' to differ in case only.
2092Handle empty rows in the 'rules for processing rows'. (credit: w)
2093Fix a logic error in the 'auto' state for scope='' which would have made it never complete in some cases. (credit: w)
2094Turns out that Zs isn't what we want; we want White_Space. (credit: w)
2095Minor fixes for table algorithms. (credit: w)
2096Make it impossible for an SVG file not labelled as SVG to be sniffed as a valid image in <img> (that would allow privilege escalation), but do allow a correctly-labelled SVG to be processed as such.
2097More thoroughly make the <img> processing model work with the new fetching and event loop definitions.
2098ack for part of previous checkin
2099Try to make <style scoped> content model rules clearer. (credit: hs)
2100<meter>0.5</meter> should work. (credit: s)
2101Add an example for image maps.
2102<details> is not for footnotes. (credit: hs)
2103Oops, forgot to escape the example.
2104Explain how UAs are to update the <details> element in response to requests for opening/closing it.
2105Reword the <script> section to allow the element to be used for data blocks that aren't actually used by any scripts. Also include an example.
2106More examples of alternative text. Also, split up the three cases of key content into subsections for easier editing.
2107indenting errors
2108Be more explicit about alt={} cases.
2109Explicitly mention alt='' in the recently added case, so that it isn't ambiguous that alt='' is required.
2110Emphasise that captions are not suitable alternative text.
2111Change from 'anti-patterns' to 'general guidelines' so as to remain positive.
2112Make the broad strokes more obvious in the alt='' section. Also, split the 'sliced image' section in two. (credit: bb)
2113Mention the <button><img></button> case. (credit: w)
2114More alt='' examples.
2115Address all recent alt='' feedback. For further details see:
2116Change 'unload' to 'close' for MessagePorts.
2117Make postMessage() return void. (credit: js)
2118Mention what <area shape=default> covers. (credit: bb)
2119Rename the 'irrelevant' attribute to 'hidden'.
2120type (Credit: avk)
2122Make empty unquoted attribute values not allowed (and some editorial fixes). (credit: sp)
2123Make <foo bar '> a parse error on the ', to match the syntax section. (credit: hs)
2124Define when to coalesce text nodes in a less DOM-prone way. (credit: pt)
2125Recommend against support for EBCDIC-based encodings. (credit: hs)
2126Define what it means for a prefix to be in scope for innerHTML in XML. (credit: sp)
2127be slightly more consistent about well-formedness terminology. (credit: sp)
2128Make the parser handle <option> and <optgroup> more usefully in the 'in body' state.
2129Mention that you can omit the </option> in <option><optgroup> too. Also, make the text support <optgroup> outside of <select>, but comment it out since it's irrelevant for now.
2130Support definitionURL='' attribute in MathML. (credit: hs)
2131Minor fixes: Ignore out-of-band encoding if it's not supported; require 2d context's 'font' attribute to ignore errors on setting.
2132Reorder the SOLIDUS entries in the tokeniser for consistency.
2133Consistent grammar fix.
2134There aren't any other end tag lines here, so I suppose it makes sense to remove the 'other'...
2135Remove redundant issue marker. (credit: as)
2136Make it clear that attributes can't be duplicated on elements.
2137US-ASCII to be treated as win1252.
2138Make U+000B into a parse error and have it convert to U+FFFD in NCRs. (credit: hs)
2139Rearchitect how RCDATA/CDATA blocks work so that they don't involve invoking the tokeniser in a weird way. (credit: w)
2140Allow a DOCTYPE for XSLT compatibility (only).
2141Use XSLT-compat instead of XSLT-generated.
2142WF2: <form> element summary.
2143WF2: <fieldset> element summary.
2144WF2: <input> element summary.
2145WF2: <button> element summary, fixes to <input> to correct copy/paste mistakes.
2146WF2: <label> element summary; make <label> and <fieldset> associate with the form element pointer.
2147Make the term 'associate' in the context of the form element pointer actually link to something.
2148WF2: <select> element summary.
2149Cross-reference interface names.
2150WF2: <datalist> element summary.
2151WF2: <optgroup> element summary.
2152WF2: <option> element summary.
2153WF2: <textarea> element summary.
2154WF2: <output> element summary.
2155WF2: Make sure all elements returned by form.elements have a 'type' DOM attribute.
2156whitespace changes
2157WF2: Defining form=, .form, and the form element pointer. Also, s/initialise/initialize/.
2158xref typo
2159WF2: Define form.elements and related features.
2160WF2: The beginnings of processing models...
2161WF2: Categorise form elements; redefine form.elements; define fieldset.elements; add more steps to form submission (still a work in progress)
2162WF2: Bring together 'disabled' and 'name' attributes that are common to all form controls, so we can have a single definition. Begin to specify 'successful'. Related minor editorial tweaks.
2163WF2: More things that make form controls not be submitted.
2164WF2: Complete the 'construcint the form data set' subalgorithm.
2165WF2: Begin to define how to handle enctype, action, method, and target attributes.
2166WF2: Define 'Submit as entity body', factor out some common prose to make it less redundant.
2167WF2: 'Delete action' and 'Get action'.
2168WF2: Submitting to data: URLs.
2169WF2: mailto submissions
2170WF2: Various tweaks, mostly editorial.
2171WF2: markup error
2172WF2: <form accept-charset> definition (but not the processing model yet).
2173WF2: and form.submit()
2174WF2: Hook up some of the form submission attributes.
2175WF2: form.reset() and a rework of how we categorise form control elements.
2176WF2: Define form.checkValidity().
2177WF2: fix errors and redundant text in today's checkins.
2178WF2: form.dispatchFormInput(), form.dispatchFormChange()
2179WF2: Simplify the APIs by removing the helper functions dispatchChange and dispatchFormChange -- they don't really give us much over people just defining their own functions.
2180WF2: Prepare the <menu> section for upcoming edits to the <input> element's section.
2181Clean up all the notes at the bottom of the document; merge in WF2 notes; mark up each note separately as an XXX item.
2182WF2: Merge the XHTML2 and XForms sections in the intro, and remove the WF2 bit, since we're merging it in now.
2183WF2: fieldset.disabled
2184WF2: Fix <fieldset> content model; define the <fieldset> element; mention the form='' attribute in the element description section.
2185WF2: typo...
2186WF2: Mention the name='' attribute in the element description section for various form elements that have one.
2187WF2: fieldset.type
2188WF2: Cross-refence the Constraint Validation API attributes and methods into their own dedicated section.
2189WF2: Move <label> up a bit so I have another excuse to not do <input> yet...
2190WF2: xref error
2191WF2: <label>.
2192WF2: <input type=''>. Also, some minor fixes relating to <label>.
2193WF2: make the input type table neater
2194WF2: Prepare <input> section for subsections.
2195WF2: <input> type sections (placeholders)
2196Change occurances of <em> that to <i> where that would be more applicable per HTML5.
2197WF2: Fill in the various <input> element's attribute definitions (but not with actual definitions yet).
2198WF2: link to disabled='' definition.
2199WF2: Add placeholders for stepUp() and stepDown()
2200WF2: mentioned attributes in the <input type=text> section; plus some editorial fixes.
2201WF2: Prevent the attributes that _aren't_ allowed from appearing in the backrefs, to make the backrefs more useful.
2202WF2: Restructure the <input> section to allow for attributes to be defined once instead of multiple times. Define autocomplete=''.
2203WF2: Try moving some more stuff around in the <input> section, maybe this will work better.
2204WF2: Begin defining value='' for <input type=text>
2205WF2: Fill in some detail for <input type=text>
2206WF2: Fix minor issues with the last checkin
2207WF2: Define resetting for >input> elements.
2208WF2: change the names of the algorithms for style reasons
2209WF2: Add a table for Anne's benefit.
2210WF2: Reorder the types to make more sense
2211WF2: input.value is going to be generic and use a type-specific algorithm instead of being type-specific like the others. Also, fiddle with the way we define and use the dirty flag.
2212WF2: Define input.valueAsDate and input.valueAsNumber for the cases where they don't apply.
2213WF2: Define input.value, especially for type=text; and related tweaks to make things read better.
2214WF2: <input type=password>; mention the DOM methods and attributes explicitly for type=text also.
2215WF2: Prepare the document for autofocus=''.
2216Try to restructure the section on focus to be more understandable.
2217Move the Scripting section up, so that the form controls and interactive elements are next to each other.
2218WF2: Define the autofocus='' attribute. Define generic task sources.
2219WF2: Define input.readonly.
2220WF2: Remove inputmode='' -- turns out most handsets support -waf-input-format already, so it's better to just use that. Reference:
2221WF2: Minor tweaks for constraint validation features.
2222WF2: Define 'candidate for constraint validation'.
2223WF2: Define .willValidate
2224WF2: fix xref typo
2225Oops, redundant queueing.
2226WF2: define the syntax of <input type=file accept='...'>
2227WF2: <input type=email>.
2228WF2: <input type=url>. Define constraints for how/when autocomplete='' happens. Say that input controls not in a document aren't mutable.
2229WF2: <input type=datetime> initial cut. Still needs filling in.
2230WF2: Collapse down some of the <input> algorithms for simplicity. Fill in a little more type=datetime detail (not much). Minor editorial fixes.
2231WF2: Fill in the attributes table for URL and date/time types.
2233WF2: maxlength=''.
2234WF2: Oops, broke the reset algorithm for <input> elements.
2235WF2: allow type=url to be set to a non-valid URL (setting it to not a URL is meaningless since everything is a URL in HTML5, technically...)
2236WF2: minor tweaks and some new issue markers.
2237WF2: Put password type after email/url.
2238WF2: r comes after m...
2239WF2: <input type=datetime min> and <input type=datetime max>.
2240WF2: Rewrite the way we're doing these type='' sections so they're less babbly.
2241WF2: typo
2242Change the 'floating point number' syntax to allow an exponent.
2243WF2: Revamp min/max/step, valueAsDate, valueAsNumber, stepUp, stepDown, and type=datetime.
2244WF2: xref typos (used dfn instead of span)
2245more typos...
2246Editorial tweaks in the type=datetime section.
2247WF2: stepUp/stepDown shouldn't go out of range. Allow rounding of datetime controls.
2248simplify the prose a bit, to not have unnecessary lists (type=datetime)
2249WF2: type=datetime-local
2250WF2: move datetime-local down a bit
2251WF2: Restructure the way dates and times are specified.
2252WF2: type=date, type=month, type=week
2253WF2: <input type=time>
2254WF2: <input type=number>
2255WF2: <input type=range>
2256Switching to the new spec generator. Disruptions in service are expected. (credit: gs)
2257WF2: <input type=checkbox> and significant fallout therefrom.
2258WF2: input.value default for checkboxes; readonly for type=range; somes notes; update the informative table of attribute applyingness.
2259WF2: <input type=radio>
2260WF2: these xrefs to <input name> were wrong.
2261WF2: Placeholders and minor rearrangements.
2262WF2: Complete the informative table for <input>; redefine input.value so it handles the various behaviors that the states have.
2263WF2: move type=hidden up.
2264WF2: merge common columns in the summary table.
2265remove remaining non-big issue markers
2266oops, had issues marked as titles instead of classes.
2267move from the 'big-issue' class to 'XXX'.
2268WF2: <input type=hidden> and some preparations for type=image.
2269WF2: <input type=submit>
2270WF2: <input type=reset>
2271WF2: <input type=button>
2272WF2: Revamp of activation behavior definitions and related details.
2273WF2: required constraint.
2274WF2: <input type=file> and some editorial fallout.
2275lining things up in the source
2276<input type=image>
2277WF2: <input size>, content side. The rendering side will be done later and is much harder...
2278WF2: <input pattern>
2279WF2: Rejig the references for pattern=''.
2280WF2: <button>.
2281fiddle with some wording for consistency.
2282WF2: input and change events; activation behaviors for <input> element.
2283minor editorial tweaks for change event
2284WF2: add links for autofocus and labels
2285WF2: <select> -- the beginnings. Also: typo fixes, fix <input size> definition to make better use of rfc2119 terms, and move .labels in the IDL down to after the CVA stuff, so that it matches the order in the prose.
2286.labels is an attribute, not an element, oops.
2287WF2: select.type
2288WF2: select.options
2289wait... there's a better way to say this... -_-
2290WF2: select.add, select.remove, select.options.add, select.options.remove
2291testing new issue-delta tracking script
2292WF2: <select> interaction.
2293WF2: final bits of the <select> API
2294The user is not a conformance class... (credit: sp)
2295WF2: input.list; input.selectedOption, <datalist>; minor fixes around previous checkins
2296WF2: <input list=''>
2297WF2: <optgroup>.
2298WF2: <option>
2299WF2: new Option() constructors.
2300WF2: option.form
2301Set up the infrastructure for <textarea>.
2302WF2: Oops, <textarea> parsing wasn't fixed when it should have been. Also, WF2: textarea.defaultValue should just be a synonym for textarea.textContent, if I'm not mistaken.
2303The term 'insertion mode' lost its definition at some point...
2304WF2: how to reset a <textarea>.
2305WF2: editing a <textarea>.
2306WF2: Change styling for book-keeping details of the <input> element states for easier reading.
2307update some broken links (credit: jcr)
2308Ack for last checkin.
2309WF2: textarea - newlines, rows='' and cols=''.
2310WF2: <textarea wrap=''> and related fallout.
2311WF2: <textarea maxlength>; also, factoring out common stuff for this from <input> to reduce duplication.
2312<textarea required>
2313WF2: Remove <textarea accept=''>. I couldn't come up with enough compelling reason to keep it.
2314WF2: required='' only applies to mutable text controls; <output> .value, .defaultValue
2315WF2: Finish <output>.
2316WF2: name='' for form controls.
2317WF2: disabled='', value, checkedness
2318WF2: Attributes for form submission
2319Remove data templates and repetition templates.
2320WF2: Define the ailments for a form control.
2321WF2: constraint validation API
2322WF2: Form submission encoding algorithms and related fallout. Also, change 'int' to 'long' in IDL blocks.
2323Clean up some stuff around the forms.
2324Editorial fixes for bug 6103.
2325more things waiting on WebIDL
2326Reorganise the video section (should be no normative changes here)
2327Change 'actively playing' to 'potentially playing' to be more accurate.
2328More slight reorganisation of the video section.
2329How to handle a moving window of backbuffer; how to handle audio playback in non-1.0 playback rates.
2330Seriously revamp the <video> state processing models and events.
2331Allow UAs to not autoplay.
2332Add navigator.canPlayType() to introspect video and audio types.
2333Turns out that embed, object, and iframe reflect strings for .height and .width, not numbers. Go figure.
2334wrong argument name for new Audio(). (bug 6057)
2335Clean up event dispatch issues in the media element section. (there were some asychronous event dispatches without event loops)
2336ack for last checkin and lots of other video feedback
2337Make the online whitelist a prefix match feature, not an exact match feature.
2338Remove opportunistic caching; leave only the fallback feature.
2339handle captive portals better in the offline appcache feature
2341unnecessary, the line never has leading whitespace
2342there's a security risk if we allow pages in one domain to fallback to pages in another domain.
2343Make fallback namespaces use the whole URL instead of just the path.
2344Make application caches work at the iframe level as well, not just whole-window level. This allows <iframe>s to contain gadgets from other hosts that are themselves cached.
2345Appcache: Rename 'implicit entries' to 'master entries'.
2346Appcache: Fix a number of race conditions; introduce the concept of the cache lifecycle, allowing caches to be obsoleted; fire checking and downloading events on ApplicationCache objects that join an update process already in progress.
2347Allow an obsolete appcache to be renewed. Move the manifest storing to later in the process to reduce more race conditions. Use 'obsolete' as the event name instead of 'error' when the cache is made obsolete. Support obsolete caches in the API.
2348Appcache: Handle errors for dynamic and master entries differently, since they aren't representative of manifest errors.
2349Appcache: try to clarify some requirements.
2350Appcache: Check the manifest again after updating the cache to make sure it hasn't changed.
2351Appcache: Allow expiring non-manifest files.
2352Appcache: Change the approach used for the dynamic entries API so that it makes more sense in a threaded environment.
2353Appcache: Allow unprompted background updates for app caches.
2354Change handling of postMessage() to fire an exception when the targetOrigin has a path/query/fragment.
2355Change handling of postMessage() again to do the previous checkin in a better way.
2356Serialise a unique (non-host) origin to the string 'null' instead of the empty stirng.
2357Define MessagePort such that they won't be garbage collected while a message is outstanding. (credit: ap)
2358MessagePorts shouldn't be GCed even when their queue is closed if they have events targetted at them. (credit: ap)
2359<eventsource>: clarify error handling rules.
2360the word 'browsing' doesn't belong here. (credit: cp)
2361disabled form controls don't see 'click' events. (credit: op)
2362WF2: Define title= on <input> with pattern=; clarify how validity states work with disabled controls.
2363Don't escape '<' and '>' in attribute values, to handle (credit: sp)
2364Make 'History' section non-normative.
2365'the it' is not english
2366Remove obsolete attributes that were removed for datatemplates that I forgot to remove before.
2367Make the styling in the elements section work better with the annotation script.
2368WF2: Define conformance criteria for <input value>. (credit: hs) Also, remove a redundant conformance requirement for contentEditable. (credit: gl)
2369WF2: Allow UAs to auto-escapify data entered for type=url.
2370ack for last checkin
2371Define four-argument form of new Option(), and clarify that selectedness is not set to true even if the default state is true.
2372WF2: typo fix; textarea.value shouldn't expose the CRLF transform.
2373xref typos
2374Note that radio buttons don't autocheck when none are checked.
2375Mention that client-side validation is not secure.
2376Make list='' only work with <datalist>, not <select>. (credit: sp, avk)
2377Clarify that processing of action='' (empty) is wrong per RFC3986. (credit: tb)
2378WF2: Clarify that hitting enter when type=image is focused is equivalent to 0,0. (credit: bz)
2379Define 'button' and 'default button' and define implicit form submission.
2380Move the 'vaguer moments' section down so the precise moments are all together.
2381Redefine datetimes to be stricter. Merge the two definitions of timezoned datetimes.
2382Define 'local date and time', 'date', 'time', and redefine 'global date and time' in a more factored-out way.
2383Define 'month' syntax, and tighten up the rest of the datetime definitions.
2384Define 'week' syntax. Details coming later.
2385WF2: Define weeks.
2386s/token/production/ fix
2387minor typos
2388note a request for forms
2389xref c/p errors
2390WF2: forminput/formchange, rework how <option> defines a command, and editorial fixes.
2391WF2: Try to clear up when you fire input/change events.
2392Fix 'interactive content' definitions for forms stuff
2393fix IDL for new Option()
2394Include an example of changing the value of a text field. Make terminology consistent.
2395Clarify that resetting a form doesn't fire 'input' events.
2396oops, that should be max
2397Make the text on proprietary languages different.
2398SSE: Add a commented out feature for triggering an immediate reconnection.
2399Split Navigator so that Workers can refer to it.
2400Remove startConversation() for now, as it is distracting in the worker discussions. May return in some form later.
2401Replace all the looping mechanisms with a single loop attribute. Allow the media URI to include a fragment identifier to seek to at the start.
2402c/p error (credit: mg)
2403Revamp the load() algorithm so that it tries each possible file in turn.
2404Make <source type> optional. Move canPlayType from navigator to HTMLMediaElement. Make it return a string instead of an integer. Fix up the candidate URL list generation. Make loadstart/play event sequencing deterministic.
2405Remove bufferedBytes and totalBytes.
2406Define pixelratio='' for values <= 0.
2407Specify how frequently the timeupdate event should fire.
2408Clarify 'static'.
2409WF2: <input placeholder=''>
2410WebSocket is cross-origin. (credit: ak)
2411Mention the lack of limits in WebSocket.
2412Correct rules for when you can omit </body> and </html>. (credit: dw)
2413location.resolveURL() for resolving relative URLs (for sending to workers, in particular) (credit: js)
2414Make <iframe>s not inherit the cache of their parent.
2415Define better the restrictions on attributes of <command> (credit: hs)
2416Change the content model of <menu> to be simpler.
2418xref error
2419copy/paste error in example (credit: jj)
2420Clarify that there can be more than one footer and that the footer is not the same as the header. (credit: jj)
2421Make <object> a form-associated element.
2422Also make <output> have the same form-association magic, for consistency. (Also, simplify how this magic is specified.) (And note the previous checkin, which means that <object> is now also affected by this magic.)
2423Define how <object> fits into form submission.
2424Add the Audience section to the intro.
2425<embed>: it turns out that unknown types cause the UAs to defer to the extension. what a mess.
2426For nested objects, only the first successful one works, the others all show fallback (and don't load plugins, documents, or whatnot).
2427NPAPI implementations apparently include the attributes of the <object> in what they send to the plugin.
2428xref error
2429WF2: <input type=search>
2430Make the request sent by ping='' include an entity body, to enable even easier recognition. (credit: jr)
2431Add an example of UI for ping=''. (credit: jr)
2432Define a way to expose HTTP login forms to spiders in 401 (or even 200) responses.
2433Ban year zero. (credit: hs)
2434<time>: Revamp the 'vaguer moments in time' concept.
2435Update examples for prior checkin; remove note on feedback since we didn't get any negative feedback and it's been a while now.
2436remove orphan section (became orphaned after we removed start/end/etc from <video>)
2437Define meaning of dates before 1582.
2438markup issue
2439WF2: Further simplify the allowed syntax for <input type=email>
2440WF2: <input type=email multiple>, <input type=file multiple>
2441Update summary tables for previous checkin.
2442remove resolve request
2443fix the definition of the categories
2444clarify use cases for <time>
2445WF2: Add novalidate='' for submission; define what happens for re-entrant .submit().
2446Add DataTransfer.clearData() with no arguments. (credit: nd)
2447Clarify what the empty string means when passed to DataTransfer.setData(). (credit: nd)
2448Make DragEvent inherit from MouseEvent so we get mouse and key info into the events. Add a note about the difference between addElement() and setDragImage(). (credit: nd)
2449fix references to <input type=text>
2450WF2: <input type=color>. Also, some editorial fixes.
2451WF2: Sigh. Browsers do <input type-image> coordinates relative to the border, not the image. (credit: sp)
2452base 16, not base 36, oops. (credit: lh)
2453fix confusion in activation behavior text (credit: lh)
2454coordinates for ismap='' clicks near borders
2455<img usemap> is interactive. (credit: sp)
2456<map name=''> must not have space characters, since they'd be illegal if you added an id='' attribute and browsers are kinda weird about them anyway.
2457SQL: Clarify how ?-substitution happens, clarify that an ObjectArray is really an 'any' array.
2458Storage should fire a quote-specific exception when out of quota. Also, replace 'security exception' with SECURITY_ERR, and cross-ref all exceptions.
2459markup fix
2460Fix a note about <object> now that it really is a form element.
2461Fix compliance issue with the CSS (removes a hack for IE6).
2462Clarify that if a DB has version='', then openDatabase() with a version won't work. (credit: ab)
2463Comment on the current violation of RFC2617. We may need to fix this at some point.
2464Mention that <abbr> won't necessarily be linked to other <abbr>s.
2465ack for previous checkin
2466Add some examples for <dfn> and <abbr>.
2467Add explanatory text around <abbr> uses.
2468Define 'the document's address'.
2469Clear up document.cookie; make update the doc address.
2470Remove the WWW-Authenticate 'HTML' scheme and replace it with requirements on browsers to ignore unknown schemes.
2471xref errors
2472make note of <select value> proposal
2473make note of <select placeholder='...' required value=''> proposal
2474Change the rules for the handling of dimension attributes so that the requirement that they keep their ratio takes into account the integer nature of the attributes.
2475Split the rendering and obsolete features sections from each other.
2476Experimental attempt at listing a set of errors that validators can comfortably classify as lower priority to aid transition.
2477typo (credit: pt)
2478Guidance on writing alt='' text for markup generators.
2479Update guidance on writing alt='' text for markup generators to handle image links.
2480grammar edit (credit: sp)
2482<img> in a section has to be the only paragraph in that section.
2483The title of the section was over-specific given the section's contents
2484Define authoring requirements for <meter> and <progress> in more detail; add v2 note for <credit>.
2485Oops, boundary checking on <meter> was off. (credit: ab)
2486<q>: now with UA-inserted quotes. Again.
2487Remove the association of <cite> and <blockquote>/<q>. (credit: sk)
2488update the note about longdesc='' to be more accurate, since longdesc='' was considered during the past year.
2489Drop HAVE_SOME_DATA. (credit: ec)
2490Make play() not reset playbackRate, since media engines need to be able to preload at the right rate, which means that we need to set the playback rate before playing generally. (credit: ec)
2491Clarify the requirement that UAs must make the 'duration' attribute accurate before playing any media..
2492Add another example for 'durationchange'.
2493Put pixelratio='' to rest.
2494Clarify that playbackRate, and allow it to be 0.0.
2495Clarify the requirements around HAVE_METADATA.
2496Fire 'waiting' after play() if playback is not immediately started.
2497Prevent firing double 'waiting' events in play(). Make sure we fire a 'timeupdate' when we begin playing, even if we only paused very briefly.
2498New plan. Introduce a 'playing' event for when the playback actually begins.
2499Always put javascript: into the online whitelist. Make some comments about HTML and HTTPS security. Vaguely define 'or equivalent' for HTTP concepts.
2500Support <script> in <select>/<optgroup>/<option>.
2501<h1><h2> should close <h1> (but <h1><span><h2> should not)
2502Lower-case the DOCTYPE name so that UAs can do case-sensitive matches on interned strings instead of case-insensitive matches. DOCTYPE name is now lowercase in the DOM (to match XHTML case -- might as well keep the differences minimal).
2503People are always complaining about this being too verbose, so let's tone it down.
2504Fix text that makes no sense for handling </body>. (credit: tt)
2505Fix </form> handling for the case where the tag isn't matched. (credit: tt)
2506Update comment for previous checkin (note -- previous checkin was substantial, not editorial as annotated. sorry about that.)
2507Disambiguate </li> and </dd>/</dt> processing. (credit: tt)
2508The 'after head' insertion mode wasn't updated to handle the new CDATA rules. (credit: tt)
2509Make <code> into a formatting element affected by the AAA. (credit: pt)
2510Remove the note about nesting <optgroup>s since this is no longer allowed anyway. (credit: hs)
2511Explain(?) that entities inside escaping text spans don't have any effect. (credit: hs)
2512Attempt to define that <script>document.write('<script src=a.js><'+'/script>1');document.write('2<script src=a.js><'+'/script>');</script> must output 12 not 21.
2513make this parenthetical more pedantically correct
2514document.close() should insert the EOF at the end of the file, not in the middle...
2515Make not fire mutation events.
2516Make Content-Language non-conforming.
2517Mention HTTP-only cookies. (credit: ak)
2518Clarify that anything that mutates the document via the parser doesn't involve mutation events.
2519SVG </script> in text/html will be an issue...
2520WF2: Fire 'input' for <textarea> too.
2521typos and c/p... (credit: pt)
2522I should stop using numbers. Yet I continue, because it just reads better... (credit: pt)
2523WF2: placeholder='' should strip newlines. (credit: hs)
2524WF2: placeholder='' must not contain newlines. (credit: hs)
2525CSRF mitigation -- add Origin header to all non-GET requests.
2526WF2: Sigh. <legend> can now be associated with forms.
2527WF2: Allow placeholder='' to be empty. No need to be overzealous in our requirements. (credit: sp)
2528WF2: Change <legend> to be treated more like <option> than like a fully-fledged form-associatd element. Simpler to implement.
2529MAJOR CHANGES: Revamp the way scripts are specified, along with their interaction with resolving relative URLs, etc. Give enough detail to justify objects in the DOM not being garbage collected randomly when still in use. Define script groups, to handle scripts going away during and session history navigation. Define why and how setTimeout(), database transactions, etc, handle page transitions. Drop the terms 'with' and 'without' script, use script is 'enabled'/'disabled' instead. Define 'unload' and 'beforeunload'. Rework how onfoo='' and .onfoo event handler attributes are defined. Rework how the content model of <noscript> is defined. Reword the way javascript: is defined to use the new terminology. Add a few notes of things that came up while I was doing all that.
2530Update terminology to help Workers specification.
2531Simplify the way messages in ports are handled when the destination's document is not available.
2532Drop MessagePort.onclose. Make MessagePort's message queue actually plug straight into the event loop so you can't lost events when sending ports across to other event loops.
2533Add structured data support to postMessage().
2534Oops, we were leaking information about the entanglement status.
2535Clean things up with the 'structured cloning' stuff.
2536WebIDL: Put in the anonymous [IndexGetter], [NameCreator], etc, operations.
2537Make non-reentrant.
2538Make focused elements fire onclick='' when the user activates them with the keyboard.
2539Add a note pointing out that buttons aren't submitted unless they submitted the form.
2540Add another example for <var>.
2541Mention popup blockers near showModalDialog(). (credit: je)
2542Make clearer what is meant by 'system focus'.
2543typo (credit: cf)
2544Make James happier.
2545Make Philip happier.
2546fix ack to use full name
2547* Header cells can now themselves have headers. * Reversed the way the header association algorithm is presented, such that it starts from a cell and reports the headers rather than generating the list of headers for each cell on a header-by-header basis. * Made it so that if headers="" points to a <td> element, the association is now set up, despite being non-conforming. * Header cells that are automatically associating now no longer stop associating when they hit equivalent cells until they have also hit a data cell. * The "col" and "row" scope values now act like the implied auto value except that they force the direction. * Empty header cells no longer get automatically associated. * Removed the wide header cell heuristic. * Made headers="" use the same ID discovery mechanism as getElementById() (namely, first match), to avoid implementations having to support multiple such mechanisms. * Made the spec define if a header is a column header or a row header in the case where scope="" is omitted. * Made scope=rowgroup and scope=colgroup non-conforming when not in an appropriate group.
2548Define how to register new http-equiv values. (credit: ma)
2549Minor tweaks to the script group stuff to fix cases I'd overlooked.
2550Redefine onerror='' handling to be reusable from other specs.
2551Define that onerror is undefined in a more convenient location.
2552Example of using RDF in XHTML.
2553Make sure to propagate add() and remove() changes when updating the cache.
2554xref the word 'master'.
2555Clarify that HTML doesn't override CSS. (credit: db)
2556Require <meta charset> within the first 512 bytes. (credit: hs)
2557Downplay <a name=''> error. (credit: hs)
2559Downplay <table summary=''> error, to aid transition. (credit: hs)
2560Downplay <noframes> error. (credit: hs)
2561Allow height/width attributes on <input type=image>. (credit: hs)
2562Downplay <td/th axis=''> error, to aid transition. (credit: hs)
2563Downplay <a longdesc=''> error, to aid transition. (credit: hs)
2564s/axis/abbr/ in recent checkin. Also, add data in comments. (credit: hs)
2565WF2: give up on <form autocomplete>.
2566Make <img src=''> a special case. (credit: bz)
2567usemap='#foo' no longer does case-insensitive matching against id=''. (credit: js)
2568Add a 'storageArea' attribute on the storage event object. (credit: be)
2570Note that the sync model for Storage doesn't imply disk-level consistency. (credit: zx)
2571typo (credit: ak)
2572typo (credit: ak)
2573Explain parseInt() usage. (credit: pt)
2574remove impossible case (thanks to webidl for making this unambiguous) (credit: pt)
2575Make it clear that sessionStorage and localStorage return the same object each time, but a unique object per Document. (credit: pt)
2576Don't fire a 'storage' event if the methods did nothing. (credit: pt)
2577xref to the document's address that i missed when doing those earlier (credit: pt)
2578Pedants unite! (credit: pt)
2579Same as last checkin, but for 'source'. (credit: pt)
2580threads aren't really predictable, but what we want to guarantee can still be required... (credit: pt)
2581obsolete note. (credit: ap)
2582Be more explicit about which document we're talking about in cache selection. (credit: ap)
2583Change the allowed content model for <svg:foreignObject> when it contains HTML. (credit: hs)
2584Vaguely try to handwave a better definition of 'subtree' into the spec.
2585Tone down the rhetoric around the <i> element being a last resort, as it doesn't really help authors to say this. (credit: hs)
2586Oops, missed a case of </br> magic in the parser. (credit: tt)
2587Remove bogus term 'transformation sheet'. (credit: eh)
2588Clarify meaning. (credit: kg)
2589Oops, <textarea> was being inserted twice. (credit: kg)
2590Change how <font> is handled in foreign content.
2591copy/paste error (credit: pt)
2592editorial fix (credit: pt)
2593Make 'phrasing content' elements explicitly 'flow content', and 'embedding content' both of those, so that readers of the spec don't have to know about what type of content is a subset of what other kind of content.
2594Replace the paragraphs about relationships between kinds of content with a diagram.
2595Define how to mark up tag clouds.
2596window.close(), window.focus(), window.blur().
2597Remove the showNotification() API due to lack of interest from implementors.
2598Remove the traces of the <span><ul>...</ul></span> idea.
2599First attempt at making document.readyState start off as 'complete' for anything but a parser-associated document.
2600make a note about mutation events wrt document.title, so that once the dom3 events spec is done, i don't forget to check if this ended up resolved.
2601Add a section explicitly talking about killing scripts that run for too long.
2602ack for previous checkin
2603remove obsolete issue note
2604Make the presence of profile='' a toned-down error since that attribute is used in WordPress templates. (Also, rearrange that section to make more sense.)
2605Update a diagram to be an SVG diagram. Thanks to Karl and Simon for the help.
2606Say which character '?' is.
2607SQL: Split the transaction() method into two, one for read/write access and one for shared read access.
2608terminology consistency fixes
2609mime type
2610Define what numbers actually mean. Fix ranges in the date microsyntaxes.
2611make negative zero not negative to please philip
2612Be more precise about what makes an invalid DOCTYPE when serialising XML.
2613That last checkin was wrong. Let's try for an even more accurate spec by noting the difference between public and system IDs in XML. (bug 6032)
2614Allow <label for=x><input id=x></label>. (bug 5997)
2615WF2: make <input type=url> require absolute URLs. (bug 6300)
2616When changing encoding on the fly after a POST, don't hit the network again whatever happens. (bug 6258)
2617Make insertCell() and insertRow() treat missing arguments like -1. (bug 6214)
2618Make HTMLCollection.namedItem() apply to non-HTML nodes. (bug 6036)
2619<noscript> in <head> needs to not cause parse errors. (bug 6213)
2620Mark .opener as [Replaceable]. Over time, more and more of these will be marked that way. (bug 5909)
2621Define callbacks using WebIDL. (bug 5814)
2622Use 'any' instead of DOMObject in IDL blocks. (I may have gone too far; let me know if you want some DOMObjects back.) (bug 6182)
2623Note that the syntax is not for XML. (bug 5994)
2624xref typos (bug 5963)
2625Picked one of the 'pending external script' cases and made it the <dfn> case. (bug 6143)
2626run-on sentence (bug 6134)
2627New examples of <ruby>. Thanks to MikeSmith and heycam for the help.
2628Further update ruby examples; add ruby introductory text.
2629ack for previous checkin
2630note for ruby rendering
2632alt text error
2633Support .indeterminate from IE/Safari. Hopefully this is an accurate specification for it. Let me know if it isn't.
2634And we're expecting regular authors to use this stuff? Sheesh. (credit: hs)
2635add missing feature note
2636add missing feature note
2637Minor editorial tweaks for WebSocket.
2638Reorganise the WebSocket section in preparation for splitting out the protocol section.
2639Extract WebSockets API (for W3C) and protocol (for IETF).
2640Add placeholder IANA considerations section.
2641Collapsable TOC.
2642Define Document's [[Get]]ter. (credit: sp)
2643Oops, missed a case -- <object> containing stuff musn't count for access. (credit: sp)
2644Clean up some sections and cross-refs.
2645<img name id> should be accessible via (credit: ms)
2646Add <embed> to the HTMLCollection nmedItem() rules. Made the HTMLDocument NameGetter override builtins. Made it return Window objects instead of <iframe>s. Added a similar NameGetter for Window. Fixed typos, added clarifications, fixed cross-reference issues.
2647Oops, <object><param></object> is still fallback-free. (credit: sp)
2648Make collections callable for compat with IE. (credit: mr)
2649Making the Window NamedGetter be OverrideBuiltins, apparently people rely on it.
2650Add <frame> and <frameset> to Window [[Get]] and getElementsByName().
2651Plug in WebIDL boilerplate.
2652Add NameDeleter to Storage (controversial) and remove it from UndoManager (to avoid future controversy).
2653Revamp the DOMStringMap text. (credit: cm)
2654Revamp the CanvasPixelArray and Window[n] texts, for better compatibility with WebIDL.
2655Revamp how <form> handles enumeration, for better compatibility with WebIDL and IE.
2656Revamp how <select> handles enumeration, for better compatibility with WebIDL and, to some very limited extent, IE. Full compatibility with IE here is impossible without me losing my sanity.
2658Define what is actually enumerated in collections, as per WebIDL.
2659WebIDLize DOMTokenList.
2660WebIDLize remaining enumerated interfaces.
2661Update IDL blocks to use NoInterfaceObject and ImplementedOn=... appropriately.
2662Simplify IDL by using [Optional].
2664Define how to handle out-of-range content attribute values in the DOM.
2665getContext() always return an object of some kind.
2666Rename confusing parameter 'arguments' in showModalDialog() - the plural was a legacy from MSDN.
2667fix mistake in one of the recent checkins - is also allowed. (credit: sp)
2669Rename the 'Origin' header to 'XXX-Origin' for now, pending a conclusion from public-webapps.
2670Redefine how the enumeration works on <form> to be more compatible with content.
2671Mute hashchanged until after the load event has fired, to avoid race conditions. Make pushState() objects survive a reload.
2672Define that relative URLs in CSS in HTML documents are not reresolved when the base URL changes.
2673clarifications for <script language=''> (credit: jj)
2674Allow authors to include documentation inside <script> blocks that have src='' attributes. (credit: sp)
2675Replace fake BNF with real ABNF.
2676markup consistency
2677Correct mistakes pointed out overnight.
2678Use the %x... syntax in ABNF. (credit: jr)
2679window.toolbar and company.
2680Note that .onLine might not be accurate
2682mention boolean attributes can't be false
2683Prevent cross-origin javascript: navigation of browsing contexts. Define the base URL and document's address of pages generated by javascript: URLs. Minor editorial tweaks.
2684Bring the event handler attribute stuff in line with reality. Make onerror='' and .onerror make sense and compatible with legacy content.
2685Define how the Window-magic event handler attributes like window.onload work. Tighten up the text around event handler attributes.
2686Make sure to fire the relevant events on the Window object now instead of the body element. Define onpopstate. Other minor fixes for the previous checkin.
2687consistency fix
2688xref fixes; mention that audit pings don't happen for broken urls.
2689Fix minor mistakes in recent checkins.
2690Oops, missed Window.applicationCache in the IDL.
2691Remove dynamic entries from the application cache for now.
2692Big revamp of how Application Caches work. The core of the algorithm is the same, but it is described quite differently and the edge cases should be smoother now.
2693Make it possible for appcaches to become obsolete.
2694Add minor clarifications about interactions of appcache and HTTP. (credit: ap)
2695Add more text about rationale for data-*='' attributes.
2696When getting the manifest fails, don't associate the resource with the cache, since the resource is probably screwed up / too new / a captive portal anyway.
2697Define handling of MIME type charsets in <script src> external scripts
2698Mention that rel='' is case-insensitive.
2699Clarify that there's no encoding necessary for innerHTML and such like.
2700Handle <!DOCTYPE> in parser.
2701xmlns:link='bla' is a parse error for all but one value of bla.
2702script node was never pushed onto the stack, oops
2703Allow <frameset> to override the 'in body' mode in certain cases for compatibility with IE and pages that rely on IE's behaviour.
2704Oops, Unicode characters can be up to six digits hex, not five.
2706accept-charset isn't a DOM attribute, duh
2707Fix collected editorial mistakes. (credit: w)
2708I missed a few non-characters. Not sure how I missed these. (credit: w)
2709Abstract out the determination of the base URL from the resolving-a-relative-URL section; remove the resolving step from the navigating and fetching algorithms; make the base URI used for relative URL resolution by scripts be set by the first script that is running, not the last.
2710Change the document's address when is invoked to match the _first script_'s document's address, not the address of the document of the script that invoked the method.
2711Define base URIs and document addresses for various freaky cases with about:blank.
2712Allow Emacs-style mode lines on the first line of a cache manifest file. (credit: ap)
2713Oops, recent changes made it impossible for an image to be served from a cache while top-level. Make it possible for anything that isn't HTML or XML to be served that way. (HTML and XML need the manifest='' attribute.) (credit: ap)
2714Avoid sending 'error' events twice to some documents. (credit: ap)
2715Fire 'noupdate' for all browsing contexts, not just the new ones. (credit: ap)
2716Simplify the cache failure steps.
2717Yay pedants!
2718Example for <section>.
2719prevent document.domain from being set to 'com' or ''.
2720Make outerHTML and insertAdjacentHTML work on XML too. Reorganise things and define the interaction of XML and the DOM for the purposes of HTML APIs.
2721Missed that onload was a global that was also on <body>.
2722Mention the global attributes everywhere.
2723Fix weirdness from last checkin. (credit: sp)
2724There is such a thing as being too subtle and cutting hairs too finely. This is one of these cases. (credit: sp)
2725Make the legacy DOCTYPE string be compatible with more tools. Clarify some XML DOCTYPE comments.
2727Change from the 'prevent scripts from running when documents are active' model to the 'split window' model, as it has better performance characteristics.
2728Tweak wording in table header algorithm.
2729Minor editorial fixes to the parser section (credit: ey)
2730Fix handing of end tags allowed before, in, and after <head> elements.
2731Floats: Allow '1e+0' syntax. Require it for serialising floating point numbers. Other editorial fixes.
2732Simplify the foster parenting definition a tiny bit.
2733Mark <datagrid> as broken for now.
2734Start filling in the rendering section.
2735Filling in the rendering section: editorial fixes; title=''.
2736Filling in the rendering section: editorial fixes; sectioning content <h1>.
2737Filling in the rendering section: editorial fixes; default styles for inline elements.
2738Filling in the rendering section: editorial fixes.
2739Filling in the rendering section: adding placeholders for future text.
2740Filling in the rendering section: moving stuff around.
2741Filling in the rendering section: moving notes around.
2742Filling in the rendering section: moving notes around.
2743Filling in the rendering section: more placeholder sections.
2744Filling in the rendering section: moving things around.
2745Filling in the rendering section: table placeholders; contenteditable placeholder
2746Filling in the rendering section: research for placeholders
2747Filling in the rendering section: more research for placeholders
2748Filling in the rendering section: more research for placeholders
2749Filling in the rendering section: <body> margins; administrivia
2750Filling in the rendering section: more margin and alignment presentational hints; fixing the infrastructure around this stuff.
2751Filling in the rendering section: color-related presentational hints; fixes for previous checkin.
2752Filling in the rendering section: <font> and related changes.
2753Filling in the rendering section: list styles; table borders; align=justify.
2754Filling in the rendering section: move the 'maps' stuff back to the intro now that we're using it in more sections.
2755Filling in the rendering section: change to a style for @namespace that i'm less likely to screw up all the time.
2756Filling in the rendering section: hook up the semantics to the rendering.
2757Filling in the rendering section: embedded content section.
2758Filling in the rendering section: regexp editing error...
2759Filling in the rendering section: Frames.
2760Filling in the rendering section: Fill in more sections -- br, tool bars, ruby...
2761Filling in the rendering section: <figure>, minor fixes
2762Filling in the rendering section: <bb>, <button>, <details>, and a rearranging of the sections with minor tweaks.
2763Filling in the rendering section: <hr>, and a rearranging of the sections with minor tweaks. Again.
2764Filling in the rendering section: <fieldset>.
2765Filling in the rendering section: <input type=text>; also, some tweaks to <img>'s requirements.
2766Filling in the rendering section: more <input> types -- date, number, range, color.
2767Filling in the rendering section: the remaining <input> types -- all the buttons.
2768Filling in the rendering section: <select>, <progress>, <meter>; minor editorial fixes here and there.
2769Filling in the rendering section: <textarea>; minor editorial fixes here and there.
2770Filling in the rendering section: <mark>, links.
2771Filling in the rendering section: remove hit testing (see www-style); require more UI for hyperlink auditing.
2772Filling in the rendering section: printing.
2773Filling in the rendering section: <link> interactivity and the browser UI
2774Recommend using dir='' rather than CSS.
2775Improve alt text. (credit: ns)
2776How to parse a color in an IE-compatible way. (credit: sp)
2777A variety of rendering section fixes: hide noscript; indent <dd>; redefine (poorly) how we do alignment; reset text-indent on form controls.
2778Defined how pseudo-classes match HTML concepts. Redid how rtl/ltr changes are marked. Stopped <legend> centering for <figure>. Removed <multicol> support. Removed <basefont> support. Removed obsolete notes. Fixed minor typos.
2779Fix the bug whereby color rules on 'html' elements in author CSS wouldn't work. (credit: sp)
2780fix typos, copyright dates
2781Remove <layer>, make <td>/<th> margins collapse too.
2782Align more closely with CSS3 UI on selector definitions.
2783:enabled doesn't match <input type=hidden>
2784rendering rules for <area>
2785Oops, <input type=hidden> can still be disabled.
2786Remove redundancy in definition of :read-write. (credit: rk)
2787note that we might want to make coords='' take %s
2788Making everything white wasn't quite what I intended. Oops.
2789Make coords='' processing (specifically error handling) more like IE for the ASCII range. (what IE does is nuts outside the ASCII range -- or rather, I can't work out what the pattern is -- so for now, we ignore that)
2790Try to clarify why this note says 'some case'.
2791Make <code> into a formatting element actually affected by the AAA. See r2509. (credit: kg)
2792Forgot to include height/width DOM attributes on <input>. (credit: ak)
2793Comment out broken link for now.
2794Attempt to simplify the AAA a bit and make it less dependent on the DOM state. Please send feedback if this breaks anything. (It is expected that this will increase the number of DOM nodes in certain cases.)
2795Make it possible to change a <script>'s text nodes or set its src='' attribute if it hasn't yet executed.
2796<small> examples (credit: bl)
2797Vaguely reduce the burden on data mining tools.
2798Support <form autocomplete>
2799Include more examples of small print in the prose.
2800ack for prev checkin
2801spellcheck='' attribute
2802Support BOMs in <script src=''> JS files. (credit: mp)
2803Pimp My Spec changes... (should be no actual difference here)
2804Handle elements losing focus when they are hidden, etc.
2805Try once more to handle elements losing focus when they are hidden, and other focusing fixes.
2806Allow </rt>, </rp>, </option>, </optgroup> to be omitted in fragment case.
2807Simplify window.onerror.
2808Allow HTML and XML files to be fetched from an application cache despite not having a manifest (same-origin caches only, of course).
2809If an appcache is obsolete, swapCache() will now unassociate altogether.
2810Stop making <noframes> a downplayed error.
2811'this' in the global scope in JS shouldn't return the window object, it should return the proxy. Hopefully one day this'll move to the ES specs.
2812Trade willful violation of ECMAScript for unintentional violation of Unicode - update the character encoding alias rules.
2813Stop referring to 'semi-transparent' elements.
2814provide a .value API for <input type=file>
2815use 'C:' instead of 'c:'; xref and typo fixes; changes to the generator again.
2816Switch to ASCII.
2817legacy colors include 'transparent' and CSS2 system colors, for some reason
2818Actually designMode does not make the document stop running scripts. That was a lie, just like the cake. Also, designMode, despite being an on/off attribute, is a string, not a boolean...
2819DOMTimeStamp betrayed us. Switch it to Date. Hopefully WebIDL or Web DOM Core will define what that means for non-JS languages...
2820Change how we define the handling of 'cursor' for <area>, to handle area { cursor: inhert }. (credit: db)
2822Some of us are just in this for the fame. :-P
2823Scripting being disabled didn't always block all scripts from being compiled.
2824Clarify that sandbox's origin features only take effect at navigation.
2825Clarify that iframe seamless can be dynamically set.
2826Make 'transparent' as a color in legacy attributes be thrown out (no real change compared to what the spec said before).
2827Clarify <small> even further, by popular request. Clarify the rules involved in the handshake of the Web Socket protocol. Some editorial tweaks.
2828Minor WindowProxy fixes.
2829Make clear that <iframe seamless> doesn't change height based on scroll position. (credit: ro)
2830Change the way we find fragment identifiers to be slightly more compatible with legacy content.
2831Fix a couple of typos; remove a feature that has now been added from the todo list.
2832Defer ideas for modal dialogs to the next version; remove some 'XXX' markers for v2 features.
2833Remove mention of the idea of an explicit calendar widget showing dates with events. This area is not yet mature (blog calendars, sites like Google Calendar, hCalendar, <time>, RDFa all have various ideas on how to approach the problem).
2834Strip a variety of XXXs that are obsolete or inappropriate: CSS stuff belongs in CSS; keygen will be dealt with as part of e-mail feedback; the other things are already done.
2835Collapse the XXXs for things for other specs into one XXX, and move it to the 'other specs' section.
2836Move some v2 ideas at the bottom of the spec up into their relevant sections; move some more non-HTML ideas into the non-HTML section.
2837Merge the text selection APIs into the relevant IDL blocks.
2838move some XXX and v2 issues around
2839watch() is a gecko-specific feature so I'm not going to address it in the spec
2840Fill in media element security/privacy section.
2841remove XXX markers for canvas v4 notes, since they probably won't make this version of the spec
2842Abstract out the encoding stuff from the parser to the infrastructure section so that it also affects form submission
2843Flesh out minor point about text nodes. Remove redundant internal note.
2844Give up on 205. Define how session history updates work without breaking scripts.
2845Define that <canvas>'s drawImage() can take a <video>.
2846note feature requests for audio
2847Define that <video> and <audio> delay the 'load' event.
2848Avoid firing 'canplay' twice if media buffers slowly. Also, what's this about a 'prevous' ready state...?
2849Change the way resources are loaded for media elements to make it actually work.
2850Tweak document.domain to handle IPv6 addresses.
2851Mark the semantics of cross-origin event sources as work-in-progress.
2852Stop duplicate 'playing' events.
2853Ugh. Fix the order of events in the case of autoplay.
2854typos, mark bad examples, include feature requests for video
2855Add the autobuffer='' attribute as a hint to browsers.
2856Fix a mistake in the table algorithm where a couple of conditions were exactly backwards for some reason. (bug 6438)
2857Ban cyclic references in headers=''.
2858Remove mention of proprietary competing platforms.
2859Allow <meta charset='utf-8'> in XML. (bug 6613)
2860Clarify the <meta charset> rules for XML documents.
2861Reword how we require that XML documents that use <meta charset> must use UTF-8. Also require it in the first 512 bytes.
2862Reverse r2732 since it broke pages.
2863Give more details of how <caption> elements should be used.
2864Remove some notes that were intended for the rendering section but which are now addressed; address others.
2865Remove redundant prose about content models for <table>.
2866Make it clear that layout tables are wrong. Suggest that tools use heuristics to detect them anyway. Allows browsers to use the summary='' attribute.
2867WebSocket-Location should use ws: or wss: (credit: wm)
2868Accessibility of <canvas> is not optional. (credit: sf)
2869Stylistic tweaks; mention the rendering rules for dim attributes.
2870Drop <eventsource>. Replace the API with an EventSource object. Rename onclosed to onclose on WebSocket. Remove the mostly worthless event definitions. I'll add more useful intro text to replace them later. Defined garbage collection specialness for WebSocket and EventSource.
2871There were a couple of places where I forgot that 'load' was also special.
2872Drop bufferingRate and bufferingThrottled for now.
2873video: Use better terminology; make all invokations of the resource selection algorithm async; fix the problem whereby you could get that algorithm running twice on one element; add a missing loop step. (credit: cp)
2874Clarify where the lack of 'autobuffer' should cause a UA to stop buffering. (credit: roc)
2875Make clear that EventSource objects don't reconnect on failure.
2876Make it possible to hide implementation requirements with an alternative style sheet. Add author documentation for most of the APIs. Fix minor typos and oversights here and there.
2877Extract Web Sockets, Event Source, and Web Storage out of HTML5.
2878Clarify <time>'s use cases and provide examples. (bug 6536)
2879Oops, missed the allowed range for 'minute'. (bug 6356)
2880Clarify that the date in <time> elements with no datetime='' attribute musn't be Julian dates that happen to look like Gregorian dates.
2882Change the way the Web Workers spec is generated.
2883Integrate the specs together a bit more, fix some legacy issues.
2884Change the term 'script's character encoding' to be less confusing.
2885postMessage() methods that take MessagePort objects now take MessagePortArray objects.
2886Add a conformance section to the WebSocket Protocol spec (e.g. reference RFC2119).
2887Make the structured data cloning algorithm support RegExp, ImageData, and be more explicit and prototypes and Error objects.
2888Define how to determine the character encoding of worker scripts.
2889Support select.add(e) and select.options.add(e) with no second argument. (credit: cm)
2890Rename attributes for form submission to avoid clashes with existing usage.
2891update styles for w3c; fix xref errors
2893Fix some headers and text/plain cross-reference issues.
2894Require CRLF, not just CR, as the blank line after the headers.
2895Remove localStorage from Workers for now.
2896Remove the commented-out notification API.
2897note for later about websocket security issue
2898Define the storage mutex. Will be used in the next few checkins.
2899Block concurrent access to .cookie and .localStorage using the storage mutex.
2900xref fixes for the merged spec
2902fix errors related to use of class=impl (credit: sp)
2903make expressions not be considered broken references
2904SVG in text/html: Second try, based on more recent feedback.
2905Fire load/error events on <link> (for external resources) and <style> (credit: gh)
2906v2 note
2907Fix race conditions in the <video> load mechanism.
2908Make <video> delay the load event until the first frame is available.
2909mark issue regarding event inconsistency
2910Allow mixed SVG and MathML and HTML in <foreignObject> so long as it is all flow content. (credit: ed)
2911Poster frames delay the 'load' event; stalling the video altogether doesn't.
2912Implicitly open the port message queue of dedicated workers. (Before, they would never actually receive messages, oops...)
2913update refs for web workers, given the recent spec splits
2914Rename navigator.releaseLocks() to navigator.getStorageUpdates(); make the network layer's seting of cookies grab the lock.
2915ack for last checkin
2916Release the storage lock whenever something modal happens, like window.alert(), or printing.
2917tweak header levels; remove obsolete issue marker
2918this was a test and it should have not been submitted.
2919Rephrase the appcache section to talk about 'cache hosts' instead of documents and browsing contexts.
2920Define the interaction of the appcache feature and the workers feature.
2921Add placeholder= for <textarea>.
2922Defer to mutation events spec.
2923Add example of form association in weird case. (credit: js)
2924Make pushState() and <a href='#foo'> actually update the window.location object's value.
2925remove some XXX markers since i decided to use internal links after all
2926Down-played errors: XHTML DOCTYPEs without SYSTEM URI are not valid. (credit: sp) (bug 6508)
2927They collapse even with a border there. (credit: sp) (bug 6586)
2928Rendering: <textarea> also resets text-indent. (credit: avk) (bug 6659)
2929Define that <embed> and <applet> elements in <object> fallback do not instantiate plugins. (bug 6434)
2930Discuss overlapping paragraphs. (bug 6507)
2931Remove mention of renameNode() since it is going to be dropped from Web DOM Core. (credit: hs) (bug 6690)
2932Not sure how I missed these 'prose content' bits, but that term was renamed to 'flow content' long ago.
2933Expose event handler DOM attributes on Document objects. (credit: sp) (bug 6474)
2934Remove line breaks from title='' examples. (credit: sp) (bug 6497)
2935Propagate the quirks mode to the fragment parser in preparation to speccing the parsing quirks. (credit: hs) (bug 6563)
2936Allow spellcheck='' to be set with no value as a shorthand for 'true'. This isn't compatible with what existing implementations do, so please let me know if this is a problem... (credit: avk) (bug 6652)
2937Make HTMLAreaElement stringify; make <textarea wrap=''> support the 'off' value.
2938Rendering: Hide noscript when script is _enabled_, duh. (credit: sp) (bug 6669)
2939Tighten up the definitions of event handler attributes. (bug 6490)
2940Improve consistency in syntax definitions. (credit: hs) (bug 6370)
2941Clarify how to handle legacy color attributes when the parser for them returns an error. (credit: sp) (bug 6556)
2942Move the enumerated attributes section up. (credit: sp) (bug 6637)
2943Allow the dimension attributes to be set to zero. (credit: sp) (bug 6544)
2944Clarify when start tags can be omitted, in the face of empty elements. (credit: sp) (bug 6630)
2945Clarify when start tags can be omitted, in the face of empty elements. (credit: sp) (bug 6630)
2946eventsource.URL should return an absolute URL. (bug 6628)
2947Instead of cloning elements in the DOM, use the original tokens to reconstruct elements. (credit: hs) (bug 6743)
2948Merge in the WF2 credits. (bug 6352)
2949Support quotes around URLs in the Refresh pragma's value. (credit: sp) (bug 6464)
2950Add a note about the head element pointer not being null in the 'after head' mode. (credit: hs) (bug 6390)
2951alt='' is required, more or less... conformance checkers can say so, anyway (credit: sp) (bug: 6494)
2952Clarify that you can't have two encoding decls. (credit: sp) (bug 6614)
2953Visit the confessional. (credit: sb) (bug 6684)
2954Refer to ECMAScript as JavaScript more consistently; add the various MIME types that JavaScript is sometimes known as.
2955Correct a note about JavaScript MIME types.
2956xref error w/ MessagePortArray
2957Set the stage for a synchronous Database API.
2958Introduce a Synchronous Database API for Workers.
2959Avoid switching to foreign content if you're already in it. Oops. (credit: jg)
2960First cut at defining <keygen>. Affects parser, rendering, and form submission.
2961Add keygen to the void element list; make those element lists alphabetical; correct an acknowledgement spelling.
2962<datagrid> mark II, first draft
2963Typos: round 1.
2964<fieldset> and <br> rendering fixes (credit: ds); also, additional ack for previous checkin
2965Add a note that omitting a start tag doesn't omit the element. (credit: fs)
2966Typos: round 2. (credit: gb)
2967Typos: round 3.
2968Typos: round 4.
2969Typos: round 5.
2970Typos: round 6.
2971Fix links to references section.
2972Typos: round 7.
2973W3C policy: fear of change, caution to the point of paralysis. Let's change the world, but without taking any risks!
2974pubrules fixes
2975there is debate over whether SQL is a good idea
2976rename the database section while we debate whether we should have databases.
2977Add a protocol flag to WebSocket(), which, if set, must be echoed by the server.
2978Limit number of non-established WebSocket connections per client.
2979Make it possible to do document.forms.f.r.value = 'b' to check the second radio button in <form name=f><input type=radio name=r value=a><input type=radio name=r value=b></form>
2980ack for prev checkin
2981Make an empty <time> render the attribute value in a locale-specific manner.
2982Update definition of :enabled to match updated Selectors spec better.
2983Make it clear that images on web sites don't fall into the 'private communication' bucket.
2984Add another example of badness.
2985Remove confusing terminology.
2986insertion mode isn't a flag it's a state variable
2987IE8 says the <body> token should kill frameset-ok flag. Also, framset-ok wasn't supported correctly in foreign-content mode yet.
2988lastModified returns current time according to IE, not the epoch. Also, never foster parent <input type=hidden>, <style>, or <script>.
2989Simplify event terminology. Make 'load' events fired on the Window have the Document as the target.
2990EOF in tag tokens throws away the token now (for security reasons) (credit: sp)
2991loadeddata fires even when jumping throught to HAVE_FUTURE_DATA. (credit: ec)
2992Clean up event dispatch, with a particular emphasis on whether events get canceled or not.
2993Make EventSource available to workers, and add some notes to EventSource's definition that clarifies how to reduce the number of connections per server. (credit: jf)
2994Fix a variety of mistakes in the rendering section. (credit: jb)
2995reference SVG for the behaviour of document.title on SVGDocument
2996Allow other specs to define default namespaces for elements they don't define. (credit: avk)
2997Make the parser pause flag become true earlier than it has done before. (credit: kg)
2998typos (credit: avk)
2999Warn about not trusting trusted sources in postMessage(). (credit: jw)
3000Fix typos, fix the silly 'escaping' of out-of-range characters in submission to actually match what browsers do.
3001ack for last checkin
3002Make pushState() affect Referer:, since otherwise analytics results would depend on when the user reloads the source page.
3003Move onundo and onredo from Document to Window.
3004Add onbeforeprint and onafterprint.
3005Update the list of event handlers to handle all the event names used in HTML5 that we want supported at the DOM level. If I missed any let me know.
3006<dl> can be ordered... it's not appropriate for dialogs because dialogs aren't name/value pairs. (credit: rk)
3007Disambiguate what document's load event is delayed. (credit: roc)
3008Clarify <section> example's use of <h1>.
3009Define <map> as transparent.
3010Rendering: video poster frames preserve aspect ration and are centered. (credit: roc)
3011typo fix
3012Fix a variety of issues with how media playback works when playbackRate is backwards.
3013markup change
3014Fire 'error' events at the <source> element when it fails to load.
3015Add the URL decomposition attributes to <a> and <area>.
3016Clarify that cloning a port doesn't cause messages to go spewing onto the floor temporarily.
3017Design notes.
3018Mention that uncommitted transactions get thrown away without blocking the entire world.
3019plugins release the storage mutex
3022Forgot [NameCreator] on Storage.
3023Disable shadows when they're at their default values.
3024note the idea of layers for canvas
3025Define the task sources used for WebStorage; define that 'storage' events are queued not synchronous. (Also: Add some v4 notes for <canvas>.)
3026Mention that tasks from discarded documents are also discarded. Change some prose to mention origins and sites instead of domains. Fire 'storage' on inactive documents as well so they all get notified.
3027Make how <video> loads resources a bit more stable -- setting src='' once fixes the src to load; <source>s are examined when the DOM is stable; error events are only fired on <source> (or once on <video> if the UA gave up) when using <source> elements.
3028Add some examples of processing of text nodes in HTML5.
3029Prefixes can be adjustd so a prefix with a ':' in it is not really a problem. (credit: sp)
3030Allow serialisers to not exactly match the DOM when it comes to CDATA nodes. (credit: sp)
3031typo (credit: ab)
3032Define the margins of h1 elements in section elements to match h2-h6, like with the font-size. (credit: sp)
3033Hide mention of datagrid selections that I missed.
3034Mention that the earliest possible position might be non-zero for non-streaming resources too.
3035Introduce video.startTime to return the earliest possible position.
3036oops, shouldn't have removed the local name part (credit: sp)
3037Twiddle some text in case we ever use the dom manipulation task source here.
3038Define the ordering of 'load' events in more pedantic detail.
3039Rename <header> to <hgroup> and restrict it just to supporting subheadings.
3040Introduce a new <header> element.
3041Clarify <header>/<hgroup> etc.
3043Attempt to resolve the terminology confusion of heading vs header.
3045note more requests for canvas
3046createImageData() now also takes a single ImageData argument, so that you can get new ImageData objects of the same dimensions, for filters that need that kind of thing.
3047Simplify Philip's crazy arc ideas using more crazy arc ideas. I hope this makes sense to someone. If it doesn't, let me know, please... (credit: pt)
3049Allow createPattern() based on <video> also, for completeness.
3050Disallow <footer> in <header> since that's probably indicative of an error, so validators should probably report it. (credit: ms)
3051Don't allow <header> in <address> or <footer>.
3052Clarify meter length definition.
3054Tentatively add a control for entering phone numbers.
3055Allow browsers to tweak the punctuation of <input type=tel>.
3056various minor webidl fixes (credit: so)
3057I keep forgetting the text/css quirk is already defined, so let's link to it explicitly.
3058Removed per hsivonen's request.
3059Redefine getElementsByTagName() for performance.
3060Correct syntax errors in the SQLException definition.
3061Update text of getElementsByTagName() for clarity.
3062Update references to Origin header to match the Origin I-D again.
3063Change the way the Command API is defined so that it applies to all elements, so that we can make any element a command without jumping through hoops.
3064fix copypasta, add more xrefs
3065Define accesskey='' and related magic to try to Solve the Problem.
3066No triggering hidden or disabled commands, oops.
3067Comment out the <command default=''> feature because 'default' is a reserved word in JS and I can't come up with a better name. Also, 'default' isn't a critical feature, so it can wait for a future version.
3068Bring setTimeout() and setInterval() up to date with recent changes and bring them back into the fold.
3069Add cite='' attribute to <section> and <article>.
3070note for future work (not written up yet as it depends on whether we pick sql or not)
3071whitespace issues
3072Clarify what 'return' means. (credit: ec)
3073Update the definition of rel=license to match actual usage (in particular Flickr's usage).
3074Add the ability to annotate custom microdata in HTML.
3075Fix some typos and add the first draft of an intro section for microdata.
3076Minor clarifications based on overnight comments on IRC.
3077Minor editorial fixes for Philip and jgraham.
3078Clarify (well, fix, really) handling of rel='' tokens in RDF conversion.
3079Minor fixes to intro.
3080More microdata intro text.
3081more intro text for microdata
3082Make the 'alternate stylesheet' keyword uppercase in RDF to prevent people from being able to set it explicitly.
3083change the order to reduce confusion (microdata intro example)
3084Finish off the microdata intro, and define the order and duplicate handling for property names.
3085markup errors
3086Oops, forgot <video>'s poorer sibling.
3087Rename property='' to itemprop=''.
3088Add to the microdata intro paragraphs introducing item='' and itemprop=''.
3089Generate a triple for each top-level item, so that they can be enumerated easily.
3090yeah i can't spell to save my life
3091Note about performance characteristics of SQLResultSetRowList. (credit: ap)
3092Strip 'about' properties when converting to RDF.
3093Add language information to the RDF output. (credit: id)
3094Introduce a vCard vocabulary.
3095Introduce a vEvent vocabulary.
3096Introduce a BibTeX vocabulary.
3097Include a summary table for bibtex (and fix some minor issues from the previous checkin).
3098Clarify a requirement regarding dtend's syntax which I had missed when going through vevent before.
3099Define 'about'; allow top-level vcards that apply to documents to trigger more RDF triples.
3100note future extension ideas for microdata
3101Add examples for vEvent.
3102oops, forgot to rename attributes
3103Add vCard example; remove CR from vEvent example; mention 'article' includes blog entries; remove bogus 'conference' entry type in table; add bibtex examples; add support for citation object in JSON output.
3104What if there's more than one? Let's continue our campaign of precision...
3105Changed HTMLPropertyCollection to always return PropertyNodeLists when accessed by name. Make itemprop=about only work for RDF on the linking elements. vCard output support. Minor tweaks to microdata. Editorial fixes.
3106Fix some algorithms aronud 'about'.
3107Describe the model expected of processors that would allow vCard, iCalendar, BibTeX, etc, extraction.
3108note language which me might add at some point
3109Allow validators to catch the case where someone round-trips microdata through RDF and forgets to convert custom properties back to their short form (since that would result in different output in non-RDF processors that support forms other than URIs, unlike RDF)
3110iCalendar output support.
3111add reference to definition of the term used
3112Not particularly great bibtex serialisation. If anyone knows of improvements that could be made here, please let me know.
3114Include microdata when dragging-and-dropping or copying-and-pasting.
3115Fix how drag and drop exports data.
3116How to extract an Atom feed from an HTML5 document
3117Fix up some Workers examples; remove extraneous <link rel=alternate> step in Atom export. (credit: sp)
3118body.background is not a url dom attribute apparently. (bug 6097)
3119Update spec to reference RFC5322 instead of RFC2822. (bug 6300)
3120Drop in some 'Extensibility' text from WF2.
3121Reduce the number of parse errors for EOF in DOCTYPE. (bug 6351) (credit: hs)
3122Fix lack of consistency of naming of confidence flags. (bug 6423) (credit: hs)
3123Make the formatBlock execCommand more discerning. (bug 6440)
3124registerContentHandler() clarifications.
3125Clarify how registered handlers work a little. (bug 6457)
3126Try to make the definitions of 'available', 'supported', image.width, image.height, and image.complete more correct.
3127ack for last checkin
3128clean up some stuff around event handler attributes
3129Mention what happens to event handlers on <body> elements obtained by XHR or createDocument(). (bug 6515)
3130Clarify what document.fgcolor does when there's no <body> element. (bug 6516)
3131Mention what to do when reflecting to a missing attribute. (bug 6516)
3132Parsing-mode quirks. (bug 6562)
3133<marquee>, DOM side
3134Rename the spellcheck 'inherit' state to 'default'.
3135Don't spellcheck passwords.
3136fix some references
3137Make 'default' spellcheck behaviour definition clearer.
3138Elaborate on willful violations.
3139note autofocus='' for contentEditable idea
3140Try to clarify what it means to change the encoding from UTF-16 to UTF-8.
3141<marquee>, CSS side
3142various typos
3143Big changes to Workers and SharedWorkers to make their lifetime model easier. Move 'close' events to v2. Also, fix some omissions and xref problems that I ran across.
3144Only directly created nested workers inherit the lifetime requirements; ports on their own don't do it.
3145Allow workers to be GC'ed when they become useless.
3146Allow h