Page MenuHomePhabricator

cite_ref shows different letter (თ) than cite_note (ჱ) on Georgian Wikipedia
Open, Needs TriagePublicBUG REPORT

Description

Steps to reproduce:

  1. Go to https://ka.wikipedia.org/wiki/საქართველოს_პარლამენტის_მეათე_მოწვევის_დეპუტატები#cite_ref-29
  2. See 5[თ]
  3. Click the to go to https://ka.wikipedia.org/wiki/საქართველოს_პარლამენტის_მეათე_მოწვევის_დეპუტატები#cite_note-29
  4. See instead of

Original description:

Hello. In Georgian wiki there is a bug in template:Notelist (georgian — თარგი:შენიშვნების სია) when we use georgian-alphabet list.

When several template:note (georgian — თარგი:შენიშვნა) are specified in Wikitext, the letters in the corresponding list (template:Notelist) are incorrect. For example, when the letter "თ" is indicated by a template:note, in the notelist it corresponds not to the same "თ" but to "ჱ". See photos below:

თ.png (472×378 px, 50 KB)
ჱ.png (194×317 px, 10 KB)

See this link for more: https://ka.wikipedia.org/wiki/საქართველოს_პარლამენტის_მეათე_მოწვევის_დეპუტატები#cite_note-29

Sincerely,

Event Timeline

Aklapper renamed this task from Bug at georgian-language notelist in Georgian Wiki (kawiki) to cite_ref shows different letter (თ) than cite_note (ჱ) on Georgian Wikipedia.Dec 10 2021, 9:40 AM
Aklapper updated the task description. (Show Details)

Hmm, this does not seem to be the case for any other Georgian letters on that page?

@Aklapper The fact is that the following letters are not used in the modern Georgian language - ჱ ჲ ჳ ჴ ჵ. Template: Note uses the modern alphabet (i.e. without those five letters), but Template: Notelist also uses those five and is erroneous.

Found out so far:

  • The characters used in the reference section are added by the browser. If you look at the code it's just a HTML list but with a specific style applied to make sure Georgian characters are used. ( seems to be the traditional Georgian alphabet )
  • The characters used in the code are created by the language system in core that is using the configured projects language to figure out what characters to use when counting. ( seems to be the modern Georgian alphabet )
  • So there seems to be a discrepancy on the method use by the browsers and the mediawiki code.

Possible solutions:

There's nothing that we can do here. Moving to done from our perspective.

To add a bit more detail: The effect we see here is because of the unfortunate implementation discussed in T198021#4466089 as well as T239911. The letter "თ" is the 8th letter in the alphabet according to https://ka.wikipedia.org/wiki/MediaWiki:cite_link_label_group-georgian, while "ჱ" is the 8th letter of the alphabet according to the CSS standard. We see "traditional Georgian" being mentioned there, while the community apparently wants to use the modern alphabet. That's a conflict.

The support for cite_link_label_group-… messages is a Cite feature, added via https://phabricator.wikimedia.org/rECITd9b8f329 in 2010, tracked via T24265. However, there is no …-georgian message in the Cite codebase. The message was added by the community not long ago, in 2020.

The list-style-type: georgian we can see in the resulting HTML is not a feature of the Cite extension. This is purely a community decision, encoded via https://ka.wikipedia.org/wiki/Template:Reflist. As such there is not really anything we can do about this from the perspective of the Cite codebase.

Options:

  1. Change https://ka.wikipedia.org/wiki/MediaWiki:Cite_link_label_group-georgian to use the traditional alphabet as well.
  2. Wait for the CSS standard to catch up and add an alternative Georgian list-style-type. However, this is somewhat unlikely given @counter-style exists (see next point).
  3. Use the @counter-style and symbols features from the most recent CSS standard to create a custom style. However, it looks like this was added to Chrome only 2 years ago. Our compatibility guidelines ask for 3 years. Also see T217775.
  4. Add a custom hack that resets the counter via a <li value="9"> attribute to continue with the correct character.
  5. …?

The best suggestion I personally have is to live with #1 as a temporary workaround, and properly implement #3 as a first-class Cite feature the moment the browser support matrix allows us to do this (entirely dropping the problematic cite_link_label_group-… messages in the process).

I have had some discussion with ssastry about #3 in https://en.wikipedia.org/w/index.php?title=User_talk:Izno&oldid=1188378629#Just_a_heads_up_about_Parsoid_Cite_CSS_requests from timestamp 04:42, 13 February 2023, but the whole might be interesting night reading.

Change 982377 had a related patch set uploaded (by Thiemo Kreuz (WMDE); author: Thiemo Kreuz (WMDE)):

[mediawiki/extensions/Cite@master] Don't search for [[MediaWiki:cite_link_label_group-]]

https://gerrit.wikimedia.org/r/982377

Change 982377 merged by jenkins-bot:

[mediawiki/extensions/Cite@master] Don't search for [[MediaWiki:cite_link_label_group-]]

https://gerrit.wikimedia.org/r/982377