Bug 152108 - spell-checking of mixed-language phrases
Summary: spell-checking of mixed-language phrases
Status: RESOLVED DUPLICATE of bug 131487
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: Writer (show other bugs)
Version:
(earliest affected)
7.3.6.2 release
Hardware: All All
: medium minor
Assignee: Not Assigned
URL:
Whiteboard:
Keywords:
Depends on:
Blocks: Spell-Checking a11y, Accessibility
  Show dependency treegraph
 
Reported: 2022-11-18 13:18 UTC by Ulrich Windl
Modified: 2022-11-22 08:23 UTC (History)
4 users (show)

See Also:
Crash report or crash signature:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Ulrich Windl 2022-11-18 13:18:39 UTC
I'm unsure whether this is a bug or not:
With German language being used I can enter "Firewall-Regeln", and it's accepted, even though "Firewall" is not a native German word (German adopted many IT words).
However when I want to be very correct and mark "Firewall" and set the language to English, then the _whole_ phrase is marked as being not correct.
That's not correct IMHO.

Why would I do that? Assuming that one day the language tags will be exported to PDF so that screen readers can pronounce the words correctly without having a huge table of exceptions (When spelling "Firewall" according to German rules, it would sound quite different). Also hyphenation rules may be different for the languages involved.
Comment 1 Stéphane Guillou (stragu) 2022-11-18 14:49:34 UTC
It is true it is surprising that the spellchecker takes the hyphenated string as a whole, using the English dictionary for the whole thing, when in fact it there are two different languages used for the different elements.

Reproduced with:

Version: 7.4.2.3 / LibreOffice Community
Build ID: 382eef1f22670f7f4118c8c2dd222ec7ad009daf
CPU threads: 8; OS: Linux 5.15; UI render: default; VCL: gtk3
Locale: de-DE (en_AU.UTF-8); UI: en-US
Calc: threaded

A similar report is Bug 131487. Could you please see if that covers your usecase and if your voice should rather be added there?

Thank you!
Comment 2 V Stuart Foote 2022-11-18 15:16:37 UTC
To my mind => NAB.

Perhaps László should comment on how the spell-checking is structured. And what if anything additional could be achieved for crafting the dictionaries.
Comment 3 Ulrich Windl 2022-11-21 06:58:46 UTC
(In reply to Stéphane Guillou (stragu) from comment #1)
> A similar report is Bug 131487.

Indeed; to me it seems that any marking for  specific language should override the built-in word-boundaries.

So for the example (using pseudo-language)"<english>Firewall</english>-Regeln"
there should be two components: "Firewall" Using english language rules) and "-Regeln" (using default (German) language rules); maybe the leading minus in "-Regeln" actually causes the problem.
Comment 4 QA Administrators 2022-11-22 03:35:59 UTC Comment hidden (obsolete)
Comment 5 Stéphane Guillou (stragu) 2022-11-22 08:23:42 UTC
Thank you, Ulrich and Stuart.
I am marking this as a duplicate, we can continue the conversation there (I've already added a comment).

*** This bug has been marked as a duplicate of bug 131487 ***