92176 - ::first-letter pseudo-element should consider both characters within letter combinations like the Dutch "ij"

Status: RESOLVED FIXED

(Core :: Layout: Text and Fonts, defect, P5)

Description

[Shanjian Li]

This is a problem split from 54467. 
As CSS specs 5.12.2 metioned :
   "Some languages may have specific rules about how to treat certain letter 
combinations. In Dutch, for example, if the letter combination "ij" apprears at 
the beginning of a word, both letters should be considered within the 
:first-letter pseudo-element."
   Should we include those situations? I doubt if anyone will list a lot of such 
examples. Even to include above mentioned example is rather difficult, because 
we have to know what language it is. (Encoding normally is easier to get. )

Comment 1

[Shanjian Li]

It is not clear either we want to implement this or not. But anyway, I don't see 
any solution in near future. 

Comment 2

[Pierre Saslawsky]

This bug belongs to Internationalization, like bug 54467.

i18n folks need to identify the combinations of letters that should be considered 
as a single letter, and the language in which they occur, then extend the patch 
crafted by Shanjian in bug 54467 keeping in mind that the :first-letter pseudo 
can in fact cover up to 3 characters (1 punctuation sign + 2 letters).

Comment 3

[anthonyd]

reassining to the correct core and core owner.

Comment 4

[Roy Yokoyama]

-> shianjian

Comment 5

[Shanjian Li]

accepting

Comment 6

[Anne (:annevk)]

Apparently, \u0132 en \u0133 are the Unicode characters for the Dutch "ij".
Could this be resolved as WONTFIX now there are characters for this? (I will
bring this up at www-style.)

Comment 7

[Albert de Klein]

In most cases ij or IJ isn't seen as one character in Dutch and in most
publications it isn't treated as such. I doubt whether the single character
unicode values for ij/IJ are ever used in Dutch publications at all.

Some Dutch references on the usage of ij can be found on:

http://taalunieversum.org/taal/advies/vraag/1194/
http://taalunieversum.org/taal/advies/vraag/1188/

Comment 8

[Ilja Kamps]

Seeing popular use and the very limited support of the unicode signs it doesn't
make sense to not support it. Although it has been tried to get computers to see
it as one character before (Dutch DOS character set and keyboards had it as a
single sign), no one actually uses it. It is not even auto converted for in most
popular wordprocessors.

Comment 9

[Sebastian Zartner [:sebo]]

Corrected the summary.
Also, that's already documented https://developer.mozilla.org/en-US/docs/Web/CSS/::first-letter.

Sebastian

Comment 10

[Jean-Yves Perrier [:teoli]]

Sebastian: doesn't this mean that this is a duplicate of another bug? (meaning it should be RESOLVED DUPLICATE or RESOLVED WORKFORME)

Comment 11

[Sebastian Zartner [:sebo]]

I couldn't find any other bug related to this. And it is also not implemented yet.
What is done is the documentation of the feature according to the specification (though no browser seems to support this yet). FWIW I added a test case in the URL field.

Maybe I was too hasty, as the browser compatibility table still needs to be updated once this bug is fixed. Should the keyword be changed back to dev-doc-needed?

Sebastian

Comment 12

[Jean-Yves Perrier [:teoli]]

Yep, that we update Fx XY for devs and the compat table when this lands :-)

Comment 13

[Zibi Braniecki [:zbraniecki][:gandalf]]

Is this bug still relevant?

Comment 14

[Anne (:annevk)]

Yeah, the specification still calls this out at https://drafts.csswg.org/css-pseudo/#first-letter-pseudo and we still ignore it. (As do other browsers, but I don't think that necessarily calls for changing the specification as in many ways this is a quality-of-implementation issue.)

Comment 15

[Zibi Braniecki [:zbraniecki][:gandalf]]

Hi Jonathan, can you triage that?

Comment 16

[Jonathan Kew [:jfkthame]]

This still seems like a valid issue, but given that afaik no other browser implements it either, it's relatively low priority. (Authors can't currently have any expectation of this working.) Though we do special-case the Dutch "ij" for casing (i.e. with text-transform: capitalize), so it would make sense to respect it for ::first-letter as well.

Comment 17

[Jonathan Kew [:jfkthame]]

Attachment: Bug 92176 - Make the ::first-letter pseudo respect the Dutch 'ij' digraph and treat it as a unit. r=dholbert

Comment 18

[Jonathan Kew [:jfkthame]]

Attachment: Bug 92176 - Add wpt reftest for ::first-letter and Dutch ij digraph. r=dholbert

Depends on D103223

Comment 19

[Jonathan Kew [:jfkthame]]

More of a Layout:Text issue than i18n, I think; moving to that component.

Comment 20

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/02b46d68f0ec
Make the ::first-letter pseudo respect the Dutch 'ij' digraph and treat it as a unit. r=dholbert
https://hg.mozilla.org/integration/autoland/rev/acf19242273c
Add wpt reftest for ::first-letter and Dutch ij digraph. r=dholbert

Comment 21

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Created web-platform-tests PR https://github.com/web-platform-tests/wpt/pull/27365 for changes under testing/web-platform/tests

Comment 22

[Alexandru Michis [:malexandru]]

https://hg.mozilla.org/mozilla-central/rev/02b46d68f0ec
https://hg.mozilla.org/mozilla-central/rev/acf19242273c

Comment 23

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Upstream PR merged by moz-wptsync-bot

Comment 24

[Rachel Andrew]

I have updated the ::first-letter page, BCD, and release notes.