Bug 63993 - Wrong cursor position for non-latin text in Web Layout
Summary: Wrong cursor position for non-latin text in Web Layout
Status: RESOLVED WORKSFORME
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: Writer (show other bugs)
Version:
(earliest affected)
4.0.0.3 release
Hardware: Other Linux (All)
: medium normal
Assignee: Not Assigned
URL:
Whiteboard: target:4.1.0
Keywords: bibisectRequest, regression
Depends on:
Blocks: Writer-Web-Layout
  Show dependency treegraph
 
Reported: 2013-04-27 11:52 UTC by Valek Filippov
Modified: 2016-10-17 22:38 UTC (History)
6 users (show)

See Also:
Crash report or crash signature:
Regression By:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Valek Filippov 2013-04-27 11:52:18 UTC
Switch to Web Layout and start typing (or copy/pasting) non-latin characters (Cyrillic in my case). Position of the cursor calculated incorrectly and cursor moves far and far ahead of text as you are typing.

It's only observed in Web Layout.

3.x doesn't have this problem.
Comment 1 Joel Madero 2013-04-30 22:21:31 UTC
Not a problem with Telugu, will attempt to reproduce with Cyrillic
Comment 2 Valek Filippov 2013-05-08 13:47:32 UTC
<vmiklos> yes, i see it on master as well. but only with web layout.
Comment 3 Joel Madero 2013-05-08 16:29:58 UTC
Perfect - in the future please try to get the OS of the person confirming, also the language that he was testing would be nice.

But glad it's confirmed. 

Valek - what do you use for using Cyrillic? Ibus? I want to confirm and then I can bibisect which would be useful in this case most likely. Also was it working in 4.0.0.3 release? If not, please update version to that release version as we try to mark version as the oldest confirmed version that had the issue.
Comment 4 Valek Filippov 2013-05-09 01:38:14 UTC
(In reply to comment #3)
> Perfect - in the future please try to get the OS of the person confirming,
> also the language that he was testing would be nice.
Hmm... I didn;t expect confirmation would take so long, that's why I added Miklos' reply on IRC.
 
> Valek - what do you use for using Cyrillic? Ibus? I want to confirm oand then
> I can bibisect which would be useful in this case most likely. Also was it
> working in 4.0.0.3 release? If not, please update version to that release
> version as we try to mark version as the oldest confirmed version that had
> the issue.

Ibus? I don't know what's that.
I guess yu can simply copy/paste any Cyrillic letters like "АБВГД" to test it.
I only have 4.0.2.2 and 3.6 something installed, so cannot verify with 4.0.0.x.
Comment 5 Khaled Hosny 2013-05-12 22:41:33 UTC
I can’t reproduce this with any of the fonts I tried, which font(s) did you use?
Comment 6 Joel Madero 2013-05-13 00:28:00 UTC
@khaled - please don't set a bug from NEW to NEEDINFO unless you are a developer planning on tackling the bug. If it's new, at least one person has confirmed it independent of the original reporter - in this case it was VMIKLOS who is one of our main developers. Marking as NEW again
Comment 7 Khaled Hosny 2013-05-13 00:48:21 UTC
Whatever.
Comment 8 Joel Madero 2013-05-13 03:16:44 UTC
Just trying to inform you of how QA does things - wasn't trying to be rude. Feel free to read the extensive documentation about how we set status'. Once it's in NEW status it's for a developer to handle, if they are not able to reproduce or have additional info they can ask for it but if a lead developer has reproduced the problem, it seems like no additional info is needed.
Comment 9 Miklos Vajna 2013-05-13 10:29:59 UTC
So, if I try to paste "фывапролдж" with Calibri 12, 5 times, I can reproduce the issue with a 664f14c6cb629404a2fc72caad4ea37734bfd9e build (2013-04-26), but I no longer see this on latest master (57b5ed51d46fd5673dfe35125ceffa71d39f133d).

Joel, Khaled knows much more about lowlevel layout engine than I. :-)

Khaled, do you think this is independent from your harfbuzz rework, or just a fallout from that, so nothing we can backport?
Comment 10 Khaled Hosny 2013-05-13 11:39:13 UTC
(In reply to comment #9)
> Khaled, do you think this is independent from your harfbuzz rework, or just
> a fallout from that, so nothing we can backport?

You can test the old layout engine by setting SAL_USE_ICULE env variable, with Callibri I don’t get any visible rendering at all with the old engine, so probably the issue was accidentally fixed by the switch to HarfBuzz.
Comment 11 Joel Madero 2013-05-14 04:38:44 UTC
I wish he had just said that instead of saying whatever....I can't memorize every developers name. Khaled - if in the future you just say "I'm a developer and I need more info" I'll make sure to leave it alone.

Also please keep in mind that our QA team is growing quite rapidly and most of them will stick to our quasi informal protocol which is that only devs can move bugs from NEW to NEEDINFO in most cases - so a friendly comment helps
Comment 12 Khaled Hosny 2013-05-14 08:43:41 UTC
(In reply to comment #11)
> I wish he had just said that instead of saying whatever....I can't memorize
> every developers name. Khaled - if in the future you just say "I'm a
> developer and I need more info" I'll make sure to leave it alone.

You could have asked instead of just reverting my change.
Comment 13 Michael Stahl (allotropia) 2013-10-04 21:09:32 UTC
can reproduce this on Linux in 4.0.0.3 already
and current libreoffice-4-0 with comment #9.

it appears to be working fine on libreoffice-4-1 and master.

so i guess unless somebody has a really good idea what may cause this,
to resolve this as WFM with 4.1.0 target.
Comment 14 Michael Stahl (allotropia) 2014-04-08 14:31:29 UTC
WFM in 4.1; the 4.0 branch is no longer maintained now => closing
Comment 15 Robinson Tryon (qubit) 2015-12-17 07:14:03 UTC Comment hidden (obsolete)