First list of vendor-prefixed tests ----------------------------------- http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/s77-caption-side-vrl-002.xht http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/s77-caption-side-vlr-003.xht http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/s77-caption-side-vrl-004.xht http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/s77-caption-side-vlr-005.xht Expected results ---------------- http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/ref-filled-green-100px-square.xht Notes ----- - More caption-side tests (for the values of: left, right, block-start, block-end, inline-start, inline-end, top-outside, bottom-outside, block-start-outside, block-end-outside) will be developed and added later in this bug report - Firefox 42.0a1 buildID=20150810092500 fails these 4 tests - Chrome 46.0.2471.2 passes these 4 tests; IE11 passes the 2 caption-side:top tests and fails the caption-side: bottom tests - I use Linux 3.13.0-61-generic x86_64, Qt: 4.8.6, KDE 4.13.3; Kubuntu (trusty) 14.04.3 LTS - I've searched for duplicates and did not find any.

This test http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/CaptionSide-writing-modes-dhtml.html suggests that block-start, block-end, inline-start and inline-end caption-side values (along with 'vertical-align: top', 'vertical-align: middle', 'vertical-align: bottom' for inline-start and inline-end values) are correctly implemented in all writing-modes in Firefox 42.0a1 buildID=20150810092500

The spec says: For implementations that do support side captions ... the top and bottom values must place the caption on the top and bottom sides of the table box, respectively. Since our impl do supports side captions, it seems to me that our current behavior matches what the spec says. I do see your email to CSSWG about this. But before the spec gets changed with your proposal, I don't think it is a bug of Gecko.

> The spec says: For implementations that do support side captions ... the top > and bottom values must place the caption on the top and bottom sides of the > table box, respectively. Xidorn, Where exactly do you read this? I do not see where it mentions to place the caption-side at the bottom of the table box. It says "must treat *both* top and captionbottom as block-start" https://drafts.csswg.org/css-writing-modes-3/#caption-side > Since our impl do supports side captions, it seems to me that our current > behavior matches what the spec says. This is what the spec says: " For such implementations [referring to implementations that do support side captions], the top and bottom values must place the caption on the top and bottom sides of the table box, respectively. " but here, there is a possible confusion. Eg. where is the top side of a table box which has its writing-mode set to vertical-rl? It would think it is on the righthand side. It would be a lot more clear if the spec mentioned "logical top" or block-start side or something like that... or even "physical top" if I'm wrong. In the realm of vertical writing-mode, the word "top" does not mean much.

> Xidorn, > > Where exactly do you read this? > > I do not see where it mentions to place the caption-side at the bottom of > the table box. It says "must treat *both* top and captionbottom as > block-start" > https://drafts.csswg.org/css-writing-modes-3/#caption-side Okay, I see what you mean now. I'm sorry about my mistake. I read your comment too quickly. I maintain Firefox 42 fails those 4 tests. In my opinion, the caption box of a vertical-rl table with caption-side set to top should be rendered on the right-hand side of the table box. Just like Chrome 46+ does.

(In reply to Gérard Talbot from comment #4) > > Xidorn, > > > > Where exactly do you read this? > > > > I do not see where it mentions to place the caption-side at the bottom of > > the table box. It says "must treat *both* top and captionbottom as > > block-start" > > https://drafts.csswg.org/css-writing-modes-3/#caption-side > > > Okay, I see what you mean now. I'm sorry about my mistake. I read your > comment too quickly. > > I maintain Firefox 42 fails those 4 tests. In my opinion, the caption box of > a vertical-rl table with caption-side set to top should be rendered on the > right-hand side of the table box. Just like Chrome 46+ does. No, that's what the block-start value is for. If 'top' and 'bottom' were supposed to be reinterpreted as logical-top and logical-bottom, there'd be no reason to introduce the block-start/end values. It seems clear to me that the spec intends the physical values 'top' and 'bottom' (and 'left' and 'right', if supported) to remain physical, and to introduce a parallel set of logical values.

> If 'top' and 'bottom' were supposed to be reinterpreted as logical-top and logical-bottom, > there'd be no reason to introduce the block-start/end values. Correct. Not having to create new values is not necessarly a bad thing... > the spec intends the physical values 'top' and 'bottom' (and 'left' and 'right', if supported) > to remain physical I did not understand the intent of the spec in such way. When reading the spec, this is not how I initially understood this. Jonathan, Thank you for your accurate explanation. I will make appropriate changes to the tests I did and continue testing. To make things more clear and easier for everyone, I will resolve this bug report for now. I may reopen it later if I find something new .. or create another bug report. Resolving as WORKSFORME

" The prop is to remove block start and end and to make top, bottom, left, and right logical (...) RESOLVED: make the block caption side property logical " http://log.csswg.org/irc.w3.org/css/2015-08-27/#e586467 F2F Paris 2015-08-27 I will create another bug report (probably with identical summary) with new tests. Gérard