[css-pseudo-4] ::first-letter should include space separators

[johannesodland]

Some languages requires a space separator between the first-punctuation and the first letter. Both the punctuation and the space separator should be included in the ::first-letter pseudo element.

Browsers treat this differently at the moment. Safari includes spaces, while Chrome does not.
https://bugs.chromium.org/p/chromium/issues/detail?id=638267.

At the moment the spec does not specify that space should be included: https://www.w3.org/TR/css-pseudo-4/#first-letter-pseudo

Examples:

French
The opening quotation mark in French cannot go without a space:
« Voulez-vous un sandwich, Henri ? »

Source: https://en.wikipedia.org/wiki/Quotation_mark#French

Norwegian
Some quotation marks in Norwegian (replikkstrek) cannot go without a space:
– Pettersen, ropte hun, – kom inn og få kaffe.

Source: https://www.sprakradet.no/sprakhjelp/Skriveregler/tegn/Tankestrek/

[johannesodland]

See #2040 for suggested ::first-punctuation pseudo class.

[johannesodland]

The following example can be used to test how different browsers renders the ::first-letter pseudo element:

https://codepen.io/johannesodland/pen/LYGYWyY

[johannesodland]

This 10 year old bugzilla issue describes the same thing:

https://bugzilla.mozilla.org/show_bug.cgi?id=602459

[fantasai]

Current spec prose for reference:

Punctuation (i.e, characters that belong to the Punctuation (P*) Unicode general category [UAX44]) that precedes or follows the first typographic letter unit must also be included in the ::first-letter pseudo-element.

I believe the pattern we want to collect is: (P+Zs)?[LN]+(ZsP+)? (where letters represent corresponding Unicode categories). So proposed wording:

Punctuation (i.e, characters that belong to the Punctuation (P*) Unicode general category [UAX44]) that precedes or follows the first typographic letter unit, as well as any intervening space separators (characters that belong to the Spaces (Zs) Unicode general category), must also be included in the ::first-letter pseudo-element.

Agenda+ to confirm.

[dbaron]

If that was supposed to be regular-expression-ish, I'd have expected the equivalent of the prose to be something more like (with slightly different notation) ([P]+[Zs])*([L]|[N])([Zs][P]+)*.

[r12a]

(I think your grammar also needs to indicate that this appears at the beginning of the paragraph.)

[fantasai]

@r12a The grammar isn't going into the spec, and the spec already anchors it to the first Letter/Number.

[css-meeting-bot]

The CSS Working Group just discussed [css-pseudo-4] ::first-letter should include space separators.

The full IRC log of that discussion

<dael> Topic: [css-pseudo-4] ::first-letter should include space separators
<dael> github: https://github.com//issues/5154
<dael> fantasai: Issue about how not including spaces but only punct.
<dael> fantasai: NUmber of punct related patterns with punct space letter.
<fantasai> https://github.com//issues/5154#issuecomment-655940861
<dael> fantasai: Prop is update as desc ^ to include intervening whitespace in ::first-letter
<dael> Rossen_: Feedback or objections?
<Rossen_> q?
<Rossen_> ack chris
<dael> faceless2_: If not followed by letter does that go with it?
<dael> fantasai: Currently ::first-letter is a letter or digit. If you have paragraph with only punct there is no ::first-letter
<dael> tantek: I accept the situation exists in prose. What I'm not seeing in issue is documentation or example like print example of ::first-letter with punct and space and letter all as first letter
<dael> tantek: Open to keeping issue open but needs more data to accept b/c otherwise might make worse
<dael> tantek: Inclusion of punct in ::first-letter at all we had numerious examples in multi lang. Since I'm not seeing that I would reject
<dael> fantasai: Do you have those examples?
<dael> tantek: Yeah, need to check my bookshelf
<dael> fantasai: Typical case if if " in french you have to include space. Usually not full sized.
<dael> tantek: My point is a french first letter with a " and a letter. Need to see that example to move forward.
<astearns> there are two references in the initial comment
<fantasai> https://un5h2085w35j89wkxbcf89h0br.julianrbryant.com/p/chromium/issues/detail?id=638267
<dael> myIes: I think I disagree with tantek. If we're doing ::first-letter that punch through pucnt this has to happen if we support french
<dael> florian: Question is valid, does french do that
<dael> fantasai: This raised b/c issue filed against browsers and browsers wanted spec update
<dael> Rossen_: And there are bugs linked in issue
<dael> tantek: At time we found print examples in magizines so it was not hard to find. If that's not true in french I'd push back. Someone more familiar with french I'd ask have they ever seen it in french and share an example rather then rely on theory
<dael> Rossen_: French only?
<dael> tantek: I think Norwegian is also provided in issue
<dael> tantek: Either is sufficent evidence.
<dael> tantek: When talking details I'd rather based on examples and not completionism reasoning.
<dael> dauwhe: We use spacing like this all the time " and than ' would be sep by a space. If that's all ::first-letter we'd want it all selected. We use spaces between punct all the time.
<dael> tantek: That's different than prop.
<dael> fantasai: That's included. All punct before first letter is included as well as any intervening space. So this would solve dauwhe use case
<dael> tantek: Have you seen it in print?
<dael> fantasai: We have it in English
<dael> dauwhe: I can't remember if I've seen that at start of chapter with initial-letter. Somewhere that I'd need a ::first-letter selector to capture that combo of glyphs
<dael> tantek: Anything adding complexity to platform has a cost. It's reasonable to request some examples produces, esp with other lang
<dael> astearns: True there is a cost, but this is a case w/o interop. We can spec what is requested and have engines that don't do it match those that do and we gain interop
<dael> Rossen_: And this is Mozilla right that would need to add to be interop
<dael> tantek: I couldn't determine from issue which engines did what
<dael> astearns: We know some do and some don't. Seems reasonable to spec this knowing there are strings that have this behavior and people like to apply ::first-letter to things. I don't know it's nec to come up with print example to spec this and have engines match
<dael> Rossen_: I think 2nd paragraph in opening issue comment. Safari incl spaces, Chrome doesn't. 10y/o bug from Mozilla desc this.
<dael> Rossen_: Safari is ahead and Moz and Blink need to catch up
<dael> Rossen_: There is precent for interop
<dael> plinss: If there isn't inerop there is fail in testing or unclarity in spec. Need to fix either way
<dael> tantek: I'd worry about compat if there's a 10 y/o bug
<dael> fantasai: Safari shipped with this. Not sure why worried about compat
<dael> tantek: Safari might have compat problem
<dael> fantasai: Doubt it
<dael> florian: Esp since people filing bugs against Chrome and FF
<dael> fantasai: We've spent a long time on this. I'd rather we not spend more time. tantek will you block and want us to go look for more b/c the lack of interop and old bugs is not enough?
<Rossen_> q?
<dael> tantek: Prefer to request examples and leave open
<dael> florian: We're trying to align so I don't object
<dael> Rossen_: Sounds like tantek you object on resolving based on lack of evidence.
<dael> Rossen_: Let's record that objection. It will go into the issue.
<dael> Rossen_: I'm hoping dauwhe or Richard can get an example and we can resolve next week.
<dael> Rossen_: Either way we'll come back to discuss this around interop.
<florian> s/We're trying to align so I don't object/I would agree with tantek if there had been interop and we were proposing to change, but there isn't interop and we're trying to align, so I don't agree/
<tantek> That's fine, didn't intend so much time on this issue
<dael> Rossen_: tantek do you agree with this?
<dael> tantek: Yes, reasonable. Thank you

[fantasai]

@johannesodland So Tantek insists on seeing real-world examples where this behavior is necessary, since he doesn't believe the change is worth making to the spec otherwise. Do you happen to have any examples of existing typography showing the desired behavior?

[plinss]

I think another important aspect is, are there use cases/examples for not including spaces in the first letter? e.g. is there existing content where someone deliberately put a space after punctuation so that the following letter was not considered to be in the ::first-letter. In english at least, I can't come up with a valid use case where this would be desirable, where I can come up with use cases where the space should be included.

If there are valid use cases in both directions, then perhaps we need an additional control to determine what's included in ::first-letter.