1057329 - update values for the text-orientation property to match css3-writing-modes

Status: RESOLVED FIXED

(Core :: CSS Parsing and Computation, defect)

Description

[Jonathan Kew [:jfkthame]]

See http://www.w3.org/TR/css3-writing-modes/#propdef-text-orientation.

Comment 1

[John Daggett (:jtd)]

If this is about renaming 'auto' to 'mixed', meh, whatever but I don't think we should add other values unless other browsers appear to be implementing them. In particular, I don't think we should add 'sideways-right', 'sideways-left' or 'use-glyph-orientation' unless other browsers appear to be implementing them and authors using them.

Comment 2

[Jonathan Kew [:jfkthame]]

(In reply to John Daggett (:jtd) from comment #1)
> If this is about renaming 'auto' to 'mixed', meh, whatever but I don't think
> we should add other values unless other browsers appear to be implementing
> them. In particular, I don't think we should add 'sideways-right',
> 'sideways-left' or 'use-glyph-orientation' unless other browsers appear to
> be implementing them and authors using them.

At this point, I wasn't intending to do anything about use-glyph-orientation, which AFAIK is a relatively obscure value only relevant to SVG, and is marked at-risk in the spec anyway. However, I assumed we'd want to support 'sideways-left' and 'sideways-right', in addition to changing 'auto' to 'mixed'.

I'm aware you questioned these values in http://lists.w3.org/Archives/Public/www-style/2013Aug/0256.html, but fantasai responded with http://lists.w3.org/Archives/Public/www-style/2013Aug/0279.html which makes the case for keeping them. If you still don't want us to have them, that's something to take up again with the CSS WG, I think.

Comment 3

[John Daggett (:jtd)]

(In reply to Jonathan Kew (:jfkthame) from comment #2)
> I'm aware you questioned these values in
> http://lists.w3.org/Archives/Public/www-style/2013Aug/0256.html, but
> fantasai responded with
> http://lists.w3.org/Archives/Public/www-style/2013Aug/0279.html which makes
> the case for keeping them. If you still don't want us to have them, that's
> something to take up again with the CSS WG, I think.

There's basically no common use case for 'sideways-left'. Implementing it adds complexity to vertical text line breaking and I think we should be aiming for the common case at this point and not wasting time implementing (and testing!) features that aren't needed or used.

I doubt me raising the incongruity of these property values again with the CSSWG will change anything since no one is interested unfortunately.

Comment 4

[Jonathan Kew [:jfkthame]]

Attachment: update the values of text-orientation to match current css-writing-modes draft (omitting use-glyph-orientation and [for now] sideways-left)

In principle, this supports all the values from the current draft except 'glyph-orientation', but I've commented out 'sideways-left' for now on the grounds that we may not implement it right away, though I expect we'll want it eventually.

Comment 5

[John Daggett (:jtd)]

Comment on attachment 8486615 [details] [diff] [review]
update the values of text-orientation to match current css-writing-modes draft (omitting use-glyph-orientation and [for now] sideways-left)


Changing 'auto' to 'mixed' for the automatic default case is silly, but, whatever.

There are only three values needed for actual vertical text usage, the default, automatic behavior, and explicit overrides for the vertical and sideways cases. I'd prefer these be 'auto, vertical, sideways' but the spec wants these to be 'mixed, vertical, sideways-right'. As defined by the spec, 'sideways-left' and 'sideways' are unnecessary.

Given that this is still behind a pref, I'd prefer to implement only what is needed. If you'd prefer these match the spec for now, fine, we can use 'mixed', 'vertical' and 'sideways-right'. I think we should not parse or otherwise refer to 'glyph-orientation' or other sideways values in our code.

I'll post yet again to www-style to try and get this fixed.

Comment 6

[Jonathan Kew [:jfkthame]]

Attachment: update the values of text-orientation to match current css-writing-modes draft (omitting use-glyph-orientation and [for now] sideways and sideways-left)

This patch parses only 'mixed', 'upright' and 'sideways-right', as the three values we need to support first. It leaves 'sideways' and 'sideways-left' commented out for now, with the expectation that these will be wanted in due course; and 'use-glyph-orientation' is omitted completely on the assumption that we aren't interested in it and the spec may drop it anyway.

Comment 7

[John Daggett (:jtd)]

Comment on attachment 8489234 [details] [diff] [review]
update the values of text-orientation to match current css-writing-modes draft (omitting use-glyph-orientation and [for now] sideways and sideways-left)

Based on the www-style discussion, the use of sideways-right/sideways-left is needed for table headers, so we should go ahead with this. The 'sideways' behavior of changing based on the writing mode is unnecessary but that's a minor point.

Comment 8

[Jonathan Kew [:jfkthame]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/6eb716762751

(Note that 'sideways-left' and 'sideways' are not yet parsed, even when CSS vertical writing mode support is enabled; the values are present in this patch but commented out, until we're ready to begin implementation.)

Comment 9

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/6eb716762751

Comment 10

[Jean-Yves Perrier [:teoli]]

Quick questions so that we can triage our giant doc heap: this specific bug has an effect only on the parsing (I see the property parsed with the right value when I activate the pref), but that has no visual effect (layout) for the user yet. Is it right?

Which bug should I follow to know when these new CSS properties (writing-mode, text-orientation and text-combine-upright) will have an effect on the layout (even still behind a pref)?

Comment 11

[John Daggett (:jtd)]

(In reply to Jean-Yves Perrier [:teoli] from comment #10)
> Quick questions so that we can triage our giant doc heap: this specific bug
> has an effect only on the parsing (I see the property parsed with the right
> value when I activate the pref), but that has no visual effect (layout) for
> the user yet. Is it right?

Right, it's under both a compile flag and a pref currently.

> Which bug should I follow to know when these new CSS properties
> (writing-mode, text-orientation and text-combine-upright) will have an
> effect on the layout (even still behind a pref)?

I think bug 145503 is the one you want to follow for this.

Comment 12

[Jean-Yves Perrier [:teoli]]

Thx!