CSS3-UI

This page is for collecting resolved issues for the CSS3-UI specification.

CSS3-UI Latest Released draft (REC since 2018)

https://www.w3.org/TR/css-ui-3/

CSS3-UI Editor's draft

https://wiki.csswg.org/spec/css3-ui

Pending More Edits (e.g. post draft publication)

none

Issues for next version

https://wiki.csswg.org/spec/css4-ui

Resolved Issues

Collecting here for producing a disposition of comments upon exiting CR.

Issue 1

Summary

css3-ui should have a test suite

Raised by

Tantek Çelik

URL

eventually http://www.w3.org/Style/CSS/Test/#css3-ui

Proposed Resolution

create at least one test case per feature, update "Current Issues" section above to add the text See the [[2]] versions of the test suite for the "Tests" references.

Status

Resolved. LCWD has link to CSS Tests page. Update this to link to specific test suite page for CSS3-UI for the next CR. Pending next CR and contributions to test suite.

Issue 13

Summary

should CSS3 UI (or some future version) incorporate a 'focusable' property (like in SVG)

Raised by

Tantek Çelik

URL

http://www.w3.org/TR/SVGTiny12/interact.html#focusable-attr

Proposed resolution

postpone til CSS4-UI.

Status

Resolved. Added to http://wiki.csswg.org/spec/css4-ui#focusable-property

Issue 19

Summary

CSS3-UI should define 'overflow-x' and 'overflow-y' properties

Raised by

David Hyatt / Tantek Çelik

URL

https://developer.mozilla.org/En/CSS/Overflow-x

URL

https://developer.mozilla.org/En/CSS/Overflow-y

URL

O'Callahan / David Hyatt emails].

Newer alternative

There is now an overflow spec, this should go there: Issue 30

Summary

Re-add ::selection, define according to existing impl interop of ::selection

Raised by

Tantek Çelik

URL

http://www.w3.org/2013/12/11-css-irc#T18-01-59 (recent IRC discussion)

URL

http://www.w3.org/TR/2005/WD-css3-selectors-20051215/#UIfragments (most recent spec text)

URL

http://www.w3.org/TR/2004/CR-css3-ui-20040511/ (most recent CR reference)

URL

http://www.w3.org/TR/2001/CR-css3-selectors-20011113/#UIfragments (most recent CR spec text)

URL

http://www.w3.org/TR/1999/WD-css3-userint-19990916#pseudo-selection

Status

work on ::selection has been moved to (and is actively being done in) Issue 40

Summary

ime-mode is windows specific and bad design, and should not be standardized. (has replacement suggestion)

Raised by

Masayuki Nakano <[email protected]> and James Craig <[email protected]>

URL

Updated Resolution

Implementations should not support the <css>ime-mode</css> property.

Status

Editor's draft updated with updated Resolution. Awaiting public draft.

Issue 47

Summary

The specced 'resize' property's behavior is *completely* different from the behavior implemented in browsers supporting the property

Raised by

François Remy

URL

Issue 48

Summary

description of cursor: auto insufficient and not interoperable

Raised by

David Baron

URL

Issue 51

Summary

applying transforms to the outline is weird, especially in 3d.

Raised by

Florian Rivoal

URL

http://florian.rivoal.net/csswg/outline.html - test case that demonstrates applying transforms to outlines behaving strangely.

Resolution

Leave explicitly undefined for level 3 (2015-02-10 f2f)

Status

Public draft published.

Issue 52

Summary

behavior of resize on pseudo-elements buggy and ambiguously defined

Raised by

Lea Verou

URL

http://dabblet.com/gist/ab432c3f6a8f672cd077 test case demonstrating non-interop / buggy behavior

Resolution

Explicitly note 'resize' does not apply to pseudo-elements. Note: may apply in the future if there is implementation of Issue 66

Summary

Drop the pseudo-classes, they have a more up to date definition in selectors 4

Raised by

Florian Rivoal

URL

http://dev.w3.org/csswg/selectors4 contains almost exact text copied from CSS3-UI sections #active ":active details", #indeterminate "The indeterminate-value pseudo-class :indeterminate", #pseudo-default ":default", #pseudo-validity ":valid and :invalid" (including XFORMS11 reference heh), #pseudo-range ":in-range and :out-of-range", #pseudo-required-value ":required and :optional". However Selectors 4 :read-only and :read-write definitions were completely different (likely newer). editor's draft updated. awaiting public draft.

Issue 69

Summary

box-sizing insufficiently specified for replaced elements

Raised by

Boris Zbarsky

URL

Issue 71

Summary

WG resolved to drop all XForms related pseudo elements from CSS-UI, but ::value and ::choices were retained, as they apply to HTML in addition to XForms as they were initially specified for. However, they are not implemented anywhere, and it is not clear that they effectively solve the problem they were introduced for.

Raised by

Florian Rivoal

URL

Issue 72

Summary

All browsers should be treated equally with regards to obsoleting the 'ime-mode' property, instead of saying "must not add support" for those that don't have it but "should drop support" for those that do.

Raised by

David Baron (in telcon)

When

2015-01-28

URL

http://www.w3.org/mid/CADhPm3uJkugzOcz-vaKXkFMGik=R2AXwxLmXaFtUj3E_HRnsyA@mail.gmail.com

URL

Status

Public draft published.

Issue 75

Summary

As we mark ime-mode as obsolete, we should point readers of the spec to the alternative ways to address the issue that ime-mode was attempting to solve.

Raised by

Florian Rivoal

When

2015-01-28

URL

http://logs.csswg.org/irc.w3.org/css/2015-01-28/#e513809

Proposed Resolution

Add a note to ime-mode pointing readers to relevant html attributes, and explaining that dropping ime-mode was because it was a bad solution, not because the problem isn't worth addressing. Draft phrasing: "

The Working Group recognizes the importance of presenting the most relevant text input UI to users depending on what type of input is expected from them. However, instead of trying to control directly the UI as ime-mode attempts to do, a better approach is to give as much context as possible to the User Agent, and let it present the best UI possible to the user out of the various possibilities available on the platform. In HTML, the following attributes are relevant:

• the lang attribute https://html.spec.whatwg.org/multipage/forms.html#the-pattern-attribute" - Shorter/briefer informative note wording that covers these features has been added to draft.

Status

Pending publication of existing editor's draft, then updated editor's draft, then another publication.

Issue 76

Summary

the change introduced to address #Issue 54 causes problems with bidi, scrolling, and is not necessarily an improvement anyway

Raised by

Florian Rivoal

URL

Issue 77

Summary

text-overflow: clip should not clip at the end of the line box

Raised by

Florian Rivoal

URL

Status

Public draft published.

Issue 78

Summary

Clarify that the :focus pseudo class matches when the focus is move to an element using the directional navigation properties, even if that element cannot otherwise be focused.

Raised by

Florian Rivoal

URL

Status

Resolved at telcon 2015-02-25. Public draft published.

Issue 79

Summary

CSS3-UI "does not define" stacking behavior of outlines, but CSS2.1 and css-position do

Raised by

Florian Rivoal

URL

Issue 82

Summary

Drop the note and the at-risk mention about the "block container" aspect of the clip property, since it is the historic behavior implemented by everybody

Raised by

Florian Rivoal

URL

Status

Public draft published.

Issue 83

Summary

A couple of (mostly editorial) tweaks to the definition of the resize property

Raised by

Florian Rivoal

URL

Issue 84

Summary

no file format required by the spec for cursor

Raised by

Florian Rivoal

URL

Status

Editor's draft updated.

Issue 85

Summary

Tiny leftover of nav-index removal should be erased

Raised by

Florian Rivoal

URL

http://www.w3.org/mid/[email protected]

Proposed Resolution

http://www.w3.org/mid/[email protected]

Status

Public draft published.

Issue 86

Summary

caret-color should be inherited, and auto should compute to auto

Raised by

Florian Rivoal

URL

Issue 87

Summary

define what happens if a target-name starting with an "_" is provided

Raised by

Florian Rivoal

URL

Issue 88

Summary

incorrect prose (rtl issue) in in text-overflow

Raised by

Florian Rivoal

URL

Issue 89

Summary

The precise direction of resizing is *not* left to the UA

Raised by

Florian Rivoal

URL

Issue 90

Summary

"the cursor’s coordinate system" is insufficiently defined for various <image> values

Raised by

Florian Rivoal

URL

Issue 91

Summary

A few minor tweaks (essentially editorial) to css-ui

Raised by

Florian Rivoal

URL

Issue 92

Summary

In vertical text with text-orientation upright, U+2026 does not seem to be the best character for the ellipsis

Raised by

Xidorn Xuan

URL

http://www.w3.org/mid/CAMdq69_N1aGTswQM7807xoZSK4berTsqq5H-cg07AmbgBPjUFA@mail.gmail.com

Resolution

Allow UAs to take writing mode into account when choosing the ellipsis character.

Status

spec edited

Issue 93

Summary

"cursor: default" collides with the css-wide "default" value defined in css-cascade-4

Raised by

Florian Rivoal

URL

Status

No change needed.

Issue 94

Summary

The default stylesheet for HTML should have "resize: both" for <textarea>

Raised by

Timeless

URL

http://www.w3.org/mid/CACsW8eGS3kZQ8a3jPAgrDaXW1WVmUQLo5vYzYty991JZzxMVdw@mail.gmail.com

Proposed Resolution

accept proposal, since all browsers that implement resize do this.

Status

2015-06-10 WG telcon resolved to accept. Editor Draft updated to match.

Issue 96

Summary

Clarify the at-risk section

Raised by

timeless

URL

http://www.w3.org/mid/CACsW8eEQVD9K48FMt0UR7=c7S1=3LmX=itcnfc8UNE6WPRsBGg@mail.gmail.com

Proposed Resolution

Rephrased, and included more specific links, to make it more clear

Status

2015-06-10 WG telcon deferred to Editors. Fixed in Editor draft

Issue 99

Summary

spec unclear as to what the limits of the area where the cursor should apply are when using border-radius

Raised by

Florian Rivoal

URL

http://www.w3.org/mid/[email protected]

Proposed Resolution

clarify that the cursor applies within the border edge, to match implementations

Status

editor draft updated

Issue 100

Summary

Non standard error handling for unsupported cursor value considered harmful

Raised by

Florian Rivoal

URL

https://wiki.csswg.org/spec/css3-ui )

Rejected Issues

Issue 49

Summary

there is no interop between UAs about showing tooltips of the ellipsed text

Raised by

Philip Walton

URL

Issue 50

Summary

add user-select back to css3-ui, it has 3 implementations

Raised by

Tab Atkins and Edward O'Connor

URL

Issue 55

Summary

Outline has a very loose definition, and significant interop problems. Should we tighten it up?

Raised by

Florian Rivoal

URL

Issue 68

Summary

apply text-overflow even when overflow is visible?

Raised by

Mats Palmgren

URL

Issue 54

Summary

Can text-overflow ellipsis not break in the middle of a word?

Raised by

Stuart Langridge

URL

https://twitter.com/sil/status/537627213789949952

Proposed Resolution

Consider allowing implementation flexibility to ellipse at a text wrap opportunity instead of at a character boundary, and ask for further clarification (should this be automatic per implementation? or under author control). Add to spec "Implementations may ellipse at a text wrap opportunity instead of a character boundary."

Status

Published 2015-02-24. Subsequently re-opened (as issue 76) and removed in editor's draft per issue 76.

Issue 80

Summary

"Should drop ime-mode ASAP" is too strong as long as a better alternative has not been standardised

Raised by

Masayuki Nakano

URL

Status

Duplicate. Group already resolved to drop "should drop" as part of issue 72 (and issue 40 updated) resolution.

Issue 95

Summary

"Ellipsed" isn't in the dictionary, use "Ellipsized" instead

Raised by

Timeless

URL

http://www.w3.org/mid/CACsW8eGS3kZQ8a3jPAgrDaXW1WVmUQLo5vYzYty991JZzxMVdw@mail.gmail.com

Proposed Resolution

reject, "Ellipsed" is in some dictionaries, "Ellipsized" is not

Status

2015-06-10 WG telcon resolved to reject. Such editorial wording is up to editors.

Issue 97

Summary

Make stronger statement regarding authors disabling the outline on :focus

Raised by

timeless

URL

http://www.w3.org/mid/CACsW8eEQVD9K48FMt0UR7=c7S1=3LmX=itcnfc8UNE6WPRsBGg@mail.gmail.com

Proposed Resolution

Clarify why it is important, do not make threats of what third parties might do if this requirement is violated.

Status

Partly rejected. Clarifications to the existing warning made, and class=advisement applied, but 2015-06-10 WG telcon resolved to reject threat request, as out of scope for a W3C spec.

Issue 98

Summary

Add note saying the outline-color may be ignored if outline-style is auto

Raised by

Timeless

URL

http://www.w3.org/mid/CACsW8eEQVD9K48FMt0UR7=c7S1=3LmX=itcnfc8UNE6WPRsBGg@mail.gmail.com

Proposed Resolution

Reject, this is already stated 2 lines above.

Status

Editors consensus to reject.

Closed Issues

Resolved issues that have been reflected in a public CSS3-UI draft. E.g.

• http://www.w3.org/TR/2015/WD-css3-ui-20150224/

Issue 2

Summary

Change name/title of spec to be consistent with other CSS3 modules

Raised by

Tantek Çelik

URL

n/a

Proposed Resolution

Change name/title of CSS3-ui from "CSS3 Basic User Interface Module" to "CSS Basic User Interface Module Level 3"

Status

Closed.

Issue 3

Summary

::value needs to specify which properties are allowed on that pseudo-element

Raised by

Tab Atkins

URL

Issue 18

Summary

How does text-overflow:ellipsis work with overflow-style: marquee-line?

Dependency

Since CSS3 Marquee is already in CR and text-overflow is the "new" feature, it is up to text-overflow to define the interaction.

Raised by

Andrew Fedoniouk

URL

http://www.w3.org/TR/2008/CR-css3-marquee-20081205/#overflow-style

Proposed Resolution

Render same as user controlled scrolling. Add mention of overflow-style to parenthetical examples listed after "When an element is scrolled".

Status

Resolved in http://www.w3.org/TR/2012/WD-css3-ui-20120117/

Update

CSS3 Marquee has been retired, as nobody implemented it and the WG does not encourage anyone to. This clarification is therefore not longer needed, and has been removed from the draft, along with all references to CSS3 Marquee

Issue 20

Summary

text-overflow definition is ambiguous as to whether it applies to vertical overflow of text

Raised by

Alan Hogan in private email to Tantek Çelik

URL

http://dl.dropbox.com/u/105727/web/text-wrap-ellipsis.html

Proposed Resolution

No one has implemented text-overflow for anything other than inline progress. Therefore we should explicitly clarify this deatil. Insert "in its inline progression direction and " into "when text overflows its block container element that has ‘overflow’ other than ‘visible’" just after "element".

Status

Closed.

Issue 21

Summary

text-overflow definition must include <string> value and 2 values option, but with both explicitly at-risk.

Raised by

CSS WG

URL

text-overflow string

text-overflow: <string>. consider incorporating dropped <string> value.

Originally text-overflow: <string> was [in CSS3 Text CR 2003-05-14] - but no one implemented it. Thus any request for including this MUST include some justification as to why/how implementations would consider it differently than they did (and reject) for CSS3 Text CR 2003.

Real-world use-cases:

• none.

Theoretical use-cases:

• text-overflow:"" can be used to clip inline content on whole characters (grapheme clusters) / atomic elements rather than clipping them at a pixel boundary. (is there any evidence of designers wanting such an effect?)
• multilingual (need specific examples here. which languages and which characters?)
• author might wish to use "(...)" for instance in citations (need a reference to this style of ellipsing in citations)
• vertical ellipsis in maths (need a reference to this style of ellipsing in math)
• a line that has nothing but images on it. Having two markers also motivates having a <string> value to be able to specify different symbols on each side.

CSSWG resolution:

semantic: "two strings (left and right) in addition to one string (both)."

unprefixed: "[FF is] implementing it unprefixed, because everyone else already has."

Possible spec markup:

( clip | ellipsis |

  <a class="noxref" href="http://www.w3.org/TR/CSS21/syndata.html#value-def-string">
    <string></a> ){1,2}

...

<string>

Render the given string to represent clipped text. The string is treated as an independent paragraph for bidi purposes.

...

[At risk]. If there is one value, use it for both the left and right line edges. If there are two values, use the first value for the right edge, and the second value for the left edge.

Issue 22

Summary

How should text selection behave in the presence of a text-overflow ellipsis?

Raised by

Tantek Çelik

URL

http://www.w3.org/Style/CSS/Tracker/issues/279

URL

Issue 23

Summary

Update CSS IS AWESOME example to show non-ellipsed last line.

Raised by

Rossen

URL

Issue 24

Summary

Provide horizontal scrolling text-overflow example revealing more text that is elided

Raised by

Tantek Çelik

URL

http://www.xanthir.com/etc/text-overflow.html

URL

Issue 25

Summary

nav-index property is not well implemented, nor expected to be. also, is problematic (per similar issues on tabindex in HTML).

Raised by

Tantek Çelik

URL

here's one: http://wiki.csswg.org/spec/css4-ui#nav-properties

URL

https://wiki.csswg.org/ideas/nav-index

URL

http://www.w3.org/2013/11/27-css-irc ).

Status

Closed.

Issue 26

Summary

add 'cursor' values 'grab' and 'grabbing' per existing interop support by Gecko and -webkit- prefixed in WebKit

Raised by

Frank Yan / Tantek Çelik

URL

http://frankyan.com/labs/cursor/

URL

https://bugzilla.mozilla.org/show_bug.cgi?id=880672

Proposed Resolution

edit the spec, add the values, definitions, inline tests

Status

Closed.

Issue 27

Summary

specify that 'cursor' propagates from the root to the viewport

Raised by

Tab Atkins

URL

Issue 28

Summary

link to nav-* directional navigation tests from nav-* section, and in tests section in header

Raised by

Tantek Çelik

URL

Issue 29

Summary

nav-* directional navigation properties definition prose fixes

Raised by

Simon Sapin

URL

Issue 31

Summary

add 'caret-color' property at risk to css3-ui

Raised by

Andrey Rybka (Bloomberg) <[email protected]>

URL

Resolved https://bugzilla.mozilla.org/show_bug.cgi?id=1063162

URL

caret-* proposals Issue 32

Summary

use line box edge instead of block container edge for the purpose of determining text-overflow

Raised by

David Baron

URL

https://bugzilla.mozilla.org/show_bug.cgi?id=944200#c3

URL

proposal https://bugzilla.mozilla.org/show_bug.cgi?id=944200

URL

https://bugs.webkit.org/show_bug.cgi?id=115746

Proposed Resolution

change the opening sentence of text-overflow section from "overflows its block container element" to "overflows its line box edge", and any other similar references to block container edge to line box edge.

Status

Effective consensus from Webkit (per their fix of 115746), Chrome (Tab's email), Gecko (per 944200 discussion). Closed.

Issue 33

Summary

In the description of the <target-name> parameter of the nav-* properties, the word string should be replaced by <string>

Raised by

Kang-Hao (Kenny) Lu <[email protected]>

URL

Status

Closed.

Issue 34

Summary

The text-overflow property definition uses () for grouping, but we should use []

Raised by

Kang-Hao (Kenny) Lu <[email protected]>

URL

Issue 35

Summary

Editorial: distinguishing public web sites from web applications is non obvious

Raised by

Kang-Hao (Kenny) Lu <[email protected]>

URL

Status

Closed.

Issue 36

Summary

"Content" properties for several input elements are locale specific

Raised by

Kang-Hao (Kenny) Lu <[email protected]>

URL

Status

Closed.

Issue 37

Summary

spec differs from implementations regarding whether resize only applies in 1 dimension if overflow-x or -y is visible in one of the dimensions.

Raised by

Kang-Hao (Kenny) Lu <[email protected]>

URL

Issue 38

Summary

what to do if outline-offset is negative, especially to the point that the offset is more than half the element size or similar values that make it impossible to draw?

Raised by

Kang-Hao (Kenny) Lu <[email protected]>

URL

Status

Closed.

Issue 39

Summary

Various issues with ime-mode

Raised by

Kang-Hao (Kenny) Lu <[email protected]>

URL

Issue 41

Summary

The 'inherit' value is globally defined for all properties, and the

preferred style appears to be to omit it in property definitions.

Raised by

Tab Atkins

URL

Issue 42

Summary

drop “inverted” from outline-color, webkit and gecko don’t do it

Raised by

Tab Atkins

URL

Issue 43

Summary

add note or author conformance requirement to not make outlines invisible on focusable elements, for accessibility reasons

Raised by

Michael Cooper <[email protected]>

URL

Status

Closed.

Issue 44

Summary

use <image> instead <uri> in the cursor property, like Issue 45

Summary

include an example highlighting the <target> part of the nav-* syntax. Suggestion: nav-up: #foo "bar"

Raised by

Yves Lafon

URL

Issue 46

Summary

what happens if <x> and <y> are out of bounds for the cursor property

Raised by

Kang-Hao (Kenny) Lu and Tab Atkins

URL

Status

Closed.

Issue 53

Summary

resize factor behaviour is weird when width / height are later dynamically modified

Raised by

Lea Verou

URL

Status

Closed.

Issue 56

Summary

WG resolved to make outline corners rounded to match borders, but it was only lightly discussed, and not obvious how that should work

Raised by

Florian Rivoal

URL

Status

Closed.

Issue 57

Summary

the default object size for cursors is defined in CSS2.1, but not in CSS3-UI, which replaces it

Raised by

Tab Atkins

URL

Status

Closed.

Issue 58

Summary

border box is a well defined term. Let's use it.

Raised by

Fantasai

URL

http://www.w3.org/mid/[email protected]

Proposed Resolution

replace "when over the element’s border, padding, and content." with "when within the element's border-box". Looks good. Changed as suggested.

Status

Closed.

Issue 59

Summary

Spec is ambiguous as to what happens when the start end of the line overflows with a single value used on text-overflow

Raised by

Florian Rivoal

URL

Note

blocks Issue 24

Status

Closed.

Issue 60

Summary

Information about when the resize factor gets reset is not sufficiently specific

Raised by

Boris Zbarsky

URL

Issue 61

Summary

Cross reference should be updated to css3-background instead of css2.1

Raised by

Florian Rivoal

URL

Issue 62

Summary

Overview is redundant with Introduction, and introduction contains out of date sentence.

Raised by

Florian

URL

Issue 63

Summary

Section "Dependencies on other modules" is redundant with bibliography, but normative "Module interactions" section defining what the spec replaces is missing

Raised by

Florian

URL

Issue 64

Summary

Drop the icon property and the icon value of the content property, they are not implemented.

Raised by

Florian Rivoal

URL

Issue 65

Summary

Drop the XForms related pseudo-elements

Raised by

Florian Rivoal

URL

Mozilla has an open bug to implement them, however dropped XForms related ::value example. Separate edit will add new examples for HTML5 inputs and ::value and ::choices.

Issue 67

Summary

Editorial improvement to clarify where the outline is

Raised by

Axel Dahmen

Raised on

2014-12-08

URL

Status

Closed.

Issue 70

Summary

CSS3-UI defines the 'default object size' for cursors, but not the 'concrete object size'

Raised by

Tab Atkins

URL

incapable of rendering a cursor above a given size, cursors larger than that size must be shrunk to within the OS-supported size bounds, while maintaining the cursor image's intrinsic ratio, if any."

Status

Closed.

Issue 73

Summary

Support for negative outline-offset is marked as optional and at-risk. However, all browsers that support outline-offset support negative values. The only issue was the lack of interop for large negative values. The new spec prose introduced to define interoperable behavior is fine.

Raised by

Florian Rivoal

When

2015-01-28

URL

http://www.w3.org/mid/CADhPm3uJkugzOcz-vaKXkFMGik=R2AXwxLmXaFtUj3E_HRnsyA@mail.gmail.com

Proposed Resolution

Remove "UAs may ignore negative values. UAs that support " and mentions of "at risk". To deal with the risk of non implementation of the interoperable behavior in the CR time frame, replace "Neither [...] may become" with "Both [...] must not become".

Status

Closed.

Issue 74

Summary

The phrasing of what happens with large negative values in the outline-offset property is unclear, and makes it sound like we should keep a blank space of outline-width * 2 inside the outline at minimum.

Raised by

Fantasai

When

2015-01-28

URL

Issue 81

Summary

A few (mostly editorial) tweaks to the definition of the resize property

Raised by

Florian Rivoal

URL

Status

closed. Note: After publication, consider either drop the "should not apply" sentence, or rephrase to "does not apply" for CSS3-UI if we get interop (multiple do nothings).

See Also

• CSSWG#css3-ui