Bug 458442 - Arabic keyboard layout sends ligatures as one character (Laa Problem)
Summary: Arabic keyboard layout sends ligatures as one character (Laa Problem)
Status: REPORTED
Alias: None
Product: kwin
Classification: Plasma
Component: input (show other bugs)
Version: unspecified
Platform: Other Linux
: NOR normal
Target Milestone: ---
Assignee: KWin default assignee
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2022-08-28 20:28 UTC by AvidSeeker
Modified: 2024-10-05 14:22 UTC (History)
2 users (show)

See Also:
Latest Commit:
Version Fixed In:
Sentry Crash Report:

Attachments
Add an attachment

Note You need to log in before you can comment on or make changes to this bug.
Description AvidSeeker 2022-08-28 20:28:59 UTC 
SUMMARY
KDE sends [Arabic ligature glyphs](https://en.wikipedia.org/wiki/Arabic_alphabet#Ligatures) as a single glyph. For example, Laa+Alif ligature "لا" (U+0644, U+0627) is sent as "ﻻ" (U+FEFB), and similarly for (ﻷ، ﻵ، ﻹ).


STEPS TO REPRODUCE
1. Settings > Keyboard > Add default Arabic layout
2. Type "ﻻ" (i.e: "b" in QWERTY keyboards)

OBSERVED RESULT
Output is ﻻ (U+FEFB)

EXPECTED RESULT
Output is لا (U+0644, U+0627). 

SOFTWARE/OS VERSIONS
All latest update

ADDITIONAL INFORMATION
To clarify for English, this is like having a key to type a ligature. Say that you want to press "b" to type two characters: "fi" but instead you get "fi" as a single character. This is exactly what's happening with Arabic.
Comment 1 Nate Graham 2022-08-30 20:28:01 UTC 
I don't know what part of the stack is responsible for this, but I very much doubt it's anywhere in KDE code.
Comment 2 AvidSeeker 2022-08-30 22:34:22 UTC 
(In reply to Nate Graham from comment #1)
> I don't know what part of the stack is responsible for this, but I very much
> doubt it's anywhere in KDE code.

At first I expected it to be from ibus, but then I tried a fresh installation of KDE without ibus and the problem persisted.
Comment 3 Wismill 2024-10-05 14:22:14 UTC 
I cannot reproduce this Plasma 6.1.5.

@AvidSeeker The sequence you expect is defined at: https://gitlab.freedesktop.org/xorg/lib/libx11/-/blob/f2ebbce6d0978d2d84f840196b03ee35a3ca1736/nls/en_US.UTF-8/Compose.pre#L4523, while the relevant key in this Arabic layout is: https://gitlab.freedesktop.org/xkeyboard-config/xkeyboard-config/-/blob/0f42a9a5b06262e35392680a3ad3d9cad3175086/symbols/ara#L53.

So you may have an issue with the Compose file parsing, which would lead to not replacing UFEFB for "لا". Do other Compose sequence work? Is there other that do not work?