1308113 - 'tab-size' should account for letter and word spacing

Status: RESOLVED FIXED

(Core :: CSS Parsing and Computation, defect, P3)

Description

[Sebastian Zartner [:sebo]]

According to the CSSWG resolution from 2016-10-05[1] 'tab-size' should account for letter spacing and word spacing.

'-moz-tab-size' should be changed accordingly.

Sebastian

[1] https://lists.w3.org/Archives/Public/www-style/2016Oct/0068.html

Comment 1

[Thomas Wisniewski]

The proposal[1] says this:
>tab stops are calculated as
>   n*( width of U+0020 plus letter-spacing plus word-spacing)

But this seems strange to me, as tab characters are already affected by letter-spacing, aren't they? (That's what I'm seeing while playing around in devtools in Firefox and Chrome). Why count letter-spacing twice for tabs?


In addition, tabs aren't word-separators in the spec[2], which also states:
>If there are no word-separator characters, or if a word-separating character has a zero advance width
>(such as the zero width space U+200B) then the user agent must not create an additional spacing between words.

How is this meant to be reconciled? Should the word-spacing be applied regardless of whether there are word-separator characters in the run?


[1] https://lists.w3.org/Archives/Public/www-style/2015Jan/0548.html
[2] https://drafts.csswg.org/css-text-3/#letter-spacing-property

Comment 2

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

That's a fair point.  Do you mind raising an issue in https://github.com/w3c/csswg-drafts/issues ?

Comment 3

[Thomas Wisniewski]

Not at all. I just raised it as issue 643: https://github.com/w3c/csswg-drafts/issues/643

Comment 4

[Thomas Wisniewski]

dbaron, do you think it's worth waiting for this spec-issue to be resolved before unprefixing -moz-tab-size? (Chrome is shipping tab-size without it, so I suspect not?)

Comment 5

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Not necessarily -- but given that this bug should be easy enough to fix, I think the best approach would be:
 * fix this bug in the way that seems to be correct, and describe what we did in the spec issue
 * then unprefix.

Comment 6

[Thomas Wisniewski]

Attachment: 1308113-ensure-tabs-in-pre-elements-are-affected-by-word-spacing.diff

Sure, but based on the discussion for the resolution, my understanding is that tab characters in <pre> should now just be affected by word-spacing (as well as letter-spacing, which they are already affected by).

I'm not 100% sure if that might affect existing content in some negative way, given that word-spacing is an inherited property. I'm game for trying it out, though.

Here's a patch to do that; I can add a reftest to it if you feel that's the way to go.

Comment 7

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

So I think your assertion that tabs are already affected by letter-spacing isn't quite right.  In particular, I think integer values N of tab-size should add in (N-1) times the letter-spacing so that a tab remains equal to N spaces.  The -1 is because the tab already gets *one* letter-spacing for being there.

I guess I don't understand what the intent of the word-spacing change is, though.  So it's not at all clear to me that the change you've posted for review actually makes sense.

Comment 8

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Comment on attachment 8820482 [details] [diff] [review]
1308113-ensure-tabs-in-pre-elements-are-affected-by-word-spacing.diff

Marking review- for now because I don't understand why this is a sensible thing to do.

I'd be happy to review a patch to fix letter-spacing, though.

(And if you can explain what adjusting for word-spacing accounts for, happy to re-review a word-spacing patch too.)

Comment 9

[Thomas Wisniewski]

dbaron, it looks as though fantasai has replied in the thread at https://github.com/w3c/csswg-drafts/issues/643

Do you think he's answered our questions to your satisfaction so we can continue here?

Comment 10

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

OK, I guess her response says we can continue here.

The above patch is wrong, since we actually *should* be considering letter-spacing.  And https://dbaron.org/css/test/2017/tabs seems to show we currently don't.  And I guess the word-spacing thing does make more sense to me now, assuming that it results in the x's always lining up with the 9's in that test.

(It's probably worth a little comparison to other browsers, though, and consideration of whether you're getting styles from the block rather than the inline.)

Comment 11

[Thomas Wisniewski]

Attachment: screenshot.png

>And I guess the word-spacing thing does make more sense to me now, assuming that it results in the x's always lining up with the 9's in that test.

Nope. At least, I just adjusted the patch to do this instead:
>  return spaces * (spaceWidthAppUnits + WordSpacing(aFrame, aTextRun, textStyle) +
>                                        LetterSpacing(aFrame, textStyle));

And I get the result in the attached screenshot. I haven't checked whether I'm getting styles from the block, though.

>It's probably worth a little comparison to other browsers, though

I don't think any browser does this yet, as it was a recent resolution. Chrome and WebKit render similarly to Firefox nightly for me on your test, though WebKit seems to nudge the X's a bit more to the right in the letter/word spacing cases than Fx/Chrome).

Comment 12

[Thomas Wisniewski]

Attachment: 1308113-ensure-tabs-in-pre-elements-are-affected-by-word-and-letter-spacing.diff

Hmm. Even when I adjust the code to ensure it uses the nearest block frame (hopefully I'm doing so correctly in this new version of the patch), the testcase still renders similarly to my previous screenshot.

Comment 13

[felice]

Yeah, but the blocking bug is prioritized as an enhancement, not a bug, causing this actual bug in standards support to stall. Should the depended bug be upgraded in priority due to this?

Comment 14

[felice]

(In reply to felice from comment #13)
> Yeah, but the blocking bug is prioritized as an enhancement, not a bug,
> causing this actual bug in standards support to stall. Should the depended
> bug be upgraded in priority due to this?

Pardon, didn't click back twice and posted to the wrong bug. Please disregard (or delete) these two comments.

Comment 15

[felice]

This bug is blocking support for the "text-size" property. It's currently prioritized as an enhancement, and as such I suspect it has been back-burnered, so text-size support is now stalled.

Should it be re-prioritized to account for that blockage?

Comment 16

[felice]

Note: Just glanced at this to see if anything had changed and noticed I had done a Freudian slip above. It's not "text-size" that's blocked; it's "tab-size".

See: bug 737785

Comment 17

[felice]

Okay, it's been another two years, and bug 737785 (lack of support for 'tab-size' from CSS spec) is STILL blocked on this bug, which is now EIGHT YEARS OLD.

Can we please get someone assigned to this?

Comment 18

[Emilio Cobos Álvarez (:emilio)]

You are also welcome to submit a patch, fwiw :)

Comment 19

[felice]

Ten years ago, I would already have tried, but I have a degenerative neurological disease that, along with the meds to treat it, really affects my ability to focus on and contribute to large projects like a browser with confidence. Basically, I don't have the capacity to fix this problem myself. I feel like a sculptor who's been reduced to whittling twigs on the porch. You don't want me trying to fix it.

I'm asking if someone who actually works on mozilla code and is familiar with it and knows various implications could take it on. It's languished for a very long time for a bug that's actually causing the browser to fail spec testing.

Comment 20

[samuel]

What can we do to get some more eyes and hands on this issue?

Comment 21

[felice]

This bug is still the only thing blocking a nine-year-old standards failure, yet it isn't assigned to anyone.

Comment 22

[Jonathan Kew [:jfkthame]]

Attachment: Bug 1308113 - Account for letter- and word-spacing when resolving tab-size <number>. r=#layout-reviewers

Comment 23

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/26fa8cc71f85
Account for letter- and word-spacing when resolving tab-size <number>. r=emilio

Comment 24

[Iulian Moraru]

https://hg.mozilla.org/mozilla-central/rev/26fa8cc71f85