Re: SIGSEGV when doing undo after putting bib items in a note

2024-07-03 Thread Scott Kostyshak
On Thu, Jul 04, 2024 at 12:39:22AM GMT, Jean-Marc Lasgouttes wrote:
> Le 02/07/2024 à 18:01, Scott Kostyshak a écrit :
> > To reproduce:
> > 
> > 1. Fresh build on current master.
> > 2. Open User Guide.
> > 3. Make sure that editing is enabled, and turn off change tracking.
> > 4. Scroll to the end.
> > 5. Select multiple bib items.
> > 6. Click on the yellow note.
> > 7. Do Undo with ctrl + z.
> > 
> > I get the attached backtrace. Can anyone else reproduce?
> > 
> > If I don't turn off change tracking, different strange behavior occurs.
> 
> I cannot reproduce, and I am surprised because I obtained that with 2.3 once
> when trying to fix the previous bug.
> 
> Do you have a more precise recipe? I tried running under valgrind (because
> your backtrace is about deleting paragraph objects), without much success.

This one I can reproduce pretty reliably. I think the only thing
imprecise about the recipe is (5), right? I attach a screenshot of a
specific selection of multiple bib items.

Scott


signature.asc
Description: PGP signature
-- 
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: Assertion when selecting large table in User Guide on master

2024-07-03 Thread Scott Kostyshak
On Wed, Jul 03, 2024 at 06:35:03PM GMT, Scott Kostyshak wrote:
> On Wed, Jul 03, 2024 at 11:59:14PM GMT, Jean-Marc Lasgouttes wrote:
> > Le 03/07/2024 à 19:34, Scott Kostyshak a écrit :
> > > On Wed, Jul 03, 2024 at 05:48:23PM GMT, Jean-Marc Lasgouttes wrote:
> > > > Le 03/07/2024 à 06:17, Scott Kostyshak a écrit :
> > > > > I still get an assertion while selecting the phone table.
> > > > 
> > > > Hmm, I was two quick in my testing.
> > > > 
> > > > I just pushed two patches which:
> > > > 1/ avoids setting Update::SinglePar in a place where this does not make
> > > > sense
> > > > 2/ make BufferView::singleParUpdate robust in such a case (but still 
> > > > emit a
> > > > warning if the position was not set, which should not happen).
> > > > 
> > > > Please test.
> > > 
> > > Works well! Much harder to break, but 1 out of 30 times I still get an
> > > issue. I don't have a recipe to reproduce. It has to do with scrolling
> > > up fast, and then pausing (by moving the cursor down) then scrolling up
> > > again. I know that's not a useful description. Does the attached
> > > backtrace give a clue?
> > 
> > Well, it looks like the second one we had, fixed by 6e0ea4269a, but it seems
> > that you found a new path to trigger it. The best would be to be able to
> > reproduce it.
> > 
> > Can you open a ticket, please, so that it is not forgotten? I can always fix
> > it with a new test, but I'd rather understand why it happens.
> 
> Sounds good, I'll try to figure out a way to reproduce it somewhat
> reliably, at least on my system. I might experiment with ydotool (like
> xdotool but works additionally with Wayland) to see if that helps to
> reproduce.
> 
> Scott

I made a ticket here and added some more details on a recipe, although
still I can only reproduce it very infrequently:

  https://www.lyx.org/trac/ticket/13081

Scott


signature.asc
Description: PGP signature
-- 
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: 2.4.1

2024-07-03 Thread Richard Kimberly Heck

On 7/3/24 6:01 PM, Jean-Marc Lasgouttes wrote:

Le 24/06/2024 à 22:26, Richard Kimberly Heck a écrit :

On 6/24/24 1:20 PM, Jean-Marc Lasgouttes wrote:

Le 09/06/2024 à 19:18, Richard Kimberly Heck a écrit :
We do not seem to have had any panicked reports of crashes. Yea! So 
I propose to merge 2.4.1-devel into 2.4.x and plan to release 2.4.1 
at the end of the month. We have quite a few fixes already in the 
pipeline. OK?


Otherwise, I'm going to try working to some kind of release 
schedule, of the sort that José suggested. So 2.4.2 would be 
expected two months later (beginning of September) and then 2.4.3 
three months later (start of December), unless we have enough fixes 
to do an earlier release.


There is a potentially interesting bug here:
https://bugs.launchpad.net/ubuntu/+source/lyx/+bug/2069792


Keep me posted. I also cannot reproduce but lack the same files.


Could you take a look at it now? I do not understand at all this 
bibio.cpp stuff to fetch author names. I guess we should just limit 
the number of authors that we parse.


Will do. We had another bug like this when some physicist was 
referencing a paper with 3000 authors or something.


Riki


--
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: SIGSEGV when doing undo after putting bib items in a note

2024-07-03 Thread Jean-Marc Lasgouttes

Le 02/07/2024 à 18:01, Scott Kostyshak a écrit :

To reproduce:

1. Fresh build on current master.
2. Open User Guide.
3. Make sure that editing is enabled, and turn off change tracking.
4. Scroll to the end.
5. Select multiple bib items.
6. Click on the yellow note.
7. Do Undo with ctrl + z.

I get the attached backtrace. Can anyone else reproduce?

If I don't turn off change tracking, different strange behavior occurs.


I cannot reproduce, and I am surprised because I obtained that with 2.3 
once when trying to fix the previous bug.


Do you have a more precise recipe? I tried running under valgrind 
(because your backtrace is about deleting paragraph objects), without 
much success.


JMarc

--
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: Assertion when selecting large table in User Guide on master

2024-07-03 Thread Scott Kostyshak
On Wed, Jul 03, 2024 at 11:59:14PM GMT, Jean-Marc Lasgouttes wrote:
> Le 03/07/2024 à 19:34, Scott Kostyshak a écrit :
> > On Wed, Jul 03, 2024 at 05:48:23PM GMT, Jean-Marc Lasgouttes wrote:
> > > Le 03/07/2024 à 06:17, Scott Kostyshak a écrit :
> > > > I still get an assertion while selecting the phone table.
> > > 
> > > Hmm, I was two quick in my testing.
> > > 
> > > I just pushed two patches which:
> > > 1/ avoids setting Update::SinglePar in a place where this does not make
> > > sense
> > > 2/ make BufferView::singleParUpdate robust in such a case (but still emit 
> > > a
> > > warning if the position was not set, which should not happen).
> > > 
> > > Please test.
> > 
> > Works well! Much harder to break, but 1 out of 30 times I still get an
> > issue. I don't have a recipe to reproduce. It has to do with scrolling
> > up fast, and then pausing (by moving the cursor down) then scrolling up
> > again. I know that's not a useful description. Does the attached
> > backtrace give a clue?
> 
> Well, it looks like the second one we had, fixed by 6e0ea4269a, but it seems
> that you found a new path to trigger it. The best would be to be able to
> reproduce it.
> 
> Can you open a ticket, please, so that it is not forgotten? I can always fix
> it with a new test, but I'd rather understand why it happens.

Sounds good, I'll try to figure out a way to reproduce it somewhat
reliably, at least on my system. I might experiment with ydotool (like
xdotool but works additionally with Wayland) to see if that helps to
reproduce.

Scott


signature.asc
Description: PGP signature
-- 
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: lyx 2.4.0 crashes on the command line on Silicon Mac

2024-07-03 Thread Eberhard W Lisse
On 02/07/2024 23:18, Stephan Witt wrote:
[...]
> Yes, the crash was to be expected. This was to let the dynamic linker
> talk about its doing.
> 
> First info: LyX is using the bundled Qt5 frameworks. Second info: Qt
> loads libqmacstyle.dylib just before the crash.
> 
> So, probably it is a problem in Qt indeed. At the moment I have to
> look into the Qt source to get an idea how to improve the situation.
> 
> Here it is working on macOS 11.7 and on Johns machine it runs fine
> too - so there is some hope :)
> 
> Do you have any Qt related environment variables set?
> 
> BR, Stephan
[...]

I'll check on my other Silicone (Mac Book Air) but no QT related
environment variables are set.


-- 
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: lyx 2.4.0 crashes on the command line on Silicon Mac

2024-07-03 Thread Eberhard W Lisse
On 02/07/2024 23:18, Stephan Witt wrote:
[...]
> Yes, the crash was to be expected. This was to let the dynamic linker
> talk about its doing.
> 
> First info: LyX is using the bundled Qt5 frameworks. Second info: Qt
> loads libqmacstyle.dylib just before the crash.
> 
> So, probably it is a problem in Qt indeed. At the moment I have to
> look into the Qt source to get an idea how to improve the situation.
> 
> Here it is working on macOS 11.7 and on Johns machine it runs fine
> too - so there is some hope :)
> 
> Do you have any Qt related environment variables set?
> 
> BR, Stephan
[...]

I'll check on my other Silicone (Mac Book Air) but no QT related
environment variables are set.

el

-- 
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: 2.4.1

2024-07-03 Thread Jean-Marc Lasgouttes

Le 24/06/2024 à 22:26, Richard Kimberly Heck a écrit :

On 6/24/24 1:20 PM, Jean-Marc Lasgouttes wrote:

Le 09/06/2024 à 19:18, Richard Kimberly Heck a écrit :
We do not seem to have had any panicked reports of crashes. Yea! So I 
propose to merge 2.4.1-devel into 2.4.x and plan to release 2.4.1 at 
the end of the month. We have quite a few fixes already in the 
pipeline. OK?


Otherwise, I'm going to try working to some kind of release schedule, 
of the sort that José suggested. So 2.4.2 would be expected two 
months later (beginning of September) and then 2.4.3 three months 
later (start of December), unless we have enough fixes to do an 
earlier release.


There is a potentially interesting bug here:
https://bugs.launchpad.net/ubuntu/+source/lyx/+bug/2069792


Keep me posted. I also cannot reproduce but lack the same files.


Could you take a look at it now? I do not understand at all this 
bibio.cpp stuff to fetch author names. I guess we should just limit the 
number of authors that we parse.


JMarc

--
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: Assertion when selecting large table in User Guide on master

2024-07-03 Thread Jean-Marc Lasgouttes

Le 03/07/2024 à 19:34, Scott Kostyshak a écrit :

On Wed, Jul 03, 2024 at 05:48:23PM GMT, Jean-Marc Lasgouttes wrote:

Le 03/07/2024 à 06:17, Scott Kostyshak a écrit :

I still get an assertion while selecting the phone table.


Hmm, I was two quick in my testing.

I just pushed two patches which:
1/ avoids setting Update::SinglePar in a place where this does not make
sense
2/ make BufferView::singleParUpdate robust in such a case (but still emit a
warning if the position was not set, which should not happen).

Please test.


Works well! Much harder to break, but 1 out of 30 times I still get an
issue. I don't have a recipe to reproduce. It has to do with scrolling
up fast, and then pausing (by moving the cursor down) then scrolling up
again. I know that's not a useful description. Does the attached
backtrace give a clue?


Well, it looks like the second one we had, fixed by 6e0ea4269a, but it 
seems that you found a new path to trigger it. The best would be to be 
able to reproduce it.


Can you open a ticket, please, so that it is not forgotten? I can always 
fix it with a new test, but I'd rather understand why it happens.


JMarc

--
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: Assertion when selecting large table in User Guide on master

2024-07-03 Thread Scott Kostyshak
On Wed, Jul 03, 2024 at 05:48:23PM GMT, Jean-Marc Lasgouttes wrote:
> Le 03/07/2024 à 06:17, Scott Kostyshak a écrit :
> > I still get an assertion while selecting the phone table.
> 
> Hmm, I was two quick in my testing.
> 
> I just pushed two patches which:
> 1/ avoids setting Update::SinglePar in a place where this does not make
> sense
> 2/ make BufferView::singleParUpdate robust in such a case (but still emit a
> warning if the position was not set, which should not happen).
> 
> Please test.

Works well! Much harder to break, but 1 out of 30 times I still get an
issue. I don't have a recipe to reproduce. It has to do with scrolling
up fast, and then pausing (by moving the cursor down) then scrolling up
again. I know that's not a useful description. Does the attached
backtrace give a clue?

Scott
Thread 1 "lyx" received signal SIGABRT, Aborted.
Download failed: Invalid argument.  Continuing without source file 
./nptl/./nptl/pthread_kill.c.
__pthread_kill_implementation (no_tid=0, signo=6, threadid=) at 
./nptl/pthread_kill.c:44
warning: 44 ./nptl/pthread_kill.c: No such file or directory
(gdb) set pagination off
(gdb) bt
#0  __pthread_kill_implementation (no_tid=0, signo=6, threadid=) 
at ./nptl/pthread_kill.c:44
#1  __pthread_kill_internal (signo=6, threadid=) at 
./nptl/pthread_kill.c:78
#2  __GI___pthread_kill (threadid=, signo=signo@entry=6) at 
./nptl/pthread_kill.c:89
#3  0x7604526e in __GI_raise (sig=sig@entry=6) at 
../sysdeps/posix/raise.c:26
#4  0x760288ff in __GI_abort () at ./stdlib/abort.c:79
#5  0x5591312f in lyx::lyx_exit (exit_code=1) at 
/home/scott/lyxbuilds/master-master/repo/src/LyX.cpp:263
#6  0x55a56720 in boost::assertion_failed (expr=0x561ba997 "false", 
function=0x561d75b4 "void lyx::doAssertWithCallstack(bool)", 
file=0x561d75da 
"/home/scott/lyxbuilds/master-master/repo/src/support/lassert.cpp", line=45) at 
/home/scott/lyxbuilds/master-master/repo/src/boost.cpp:47
#7  0x56172abb in lyx::doAssertWithCallstack (value=false) at 
/home/scott/lyxbuilds/master-master/repo/src/support/lassert.cpp:45
#8  0x5617325e in lyx::doAssert (expr=0x561a1f39 "hasPosition()", 
file=0x561a1f47 
"/home/scott/lyxbuilds/master-master/repo/src/ParagraphMetrics.cpp", line=71)
at /home/scott/lyxbuilds/master-master/repo/src/support/lassert.cpp:54
#9  0x55983043 in lyx::ParagraphMetrics::position (this=0x59930d98) 
at /home/scott/lyxbuilds/master-master/repo/src/ParagraphMetrics.cpp:71
#10 0x557b7a7a in lyx::BufferView::getPos (this=0x58306120, 
dit=...) at /home/scott/lyxbuilds/master-master/repo/src/BufferView.cpp:3430
#11 0x557ceab5 in lyx::BufferView::checkCursorScrollOffset 
(this=0x58306120) at 
/home/scott/lyxbuilds/master-master/repo/src/BufferView.cpp:3627
#12 0x557cc103 in lyx::BufferView::draw (this=0x58306120, pain=..., 
paint_caret=false) at 
/home/scott/lyxbuilds/master-master/repo/src/BufferView.cpp:3684
#13 0x55e4fb66 in lyx::frontend::GuiWorkArea::paintEvent 
(this=0x58454fa0, ev=0x7fffba20) at 
/home/scott/lyxbuilds/master-master/repo/src/frontends/qt/GuiWorkArea.cpp:1402
#14 0x777b0df8 in QWidget::event (this=this@entry=0x58454fa0, 
event=event@entry=0x7fffba20) at kernel/qwidget.cpp:9045
#15 0x77862647 in QFrame::event (this=0x58454fa0, e=0x7fffba20) 
at widgets/qframe.cpp:550
#16 0x76ad7e82 in 
QCoreApplicationPrivate::sendThroughObjectEventFilters 
(receiver=receiver@entry=0x582fee10, event=event@entry=0x7fffba20)
at kernel/qcoreapplication.cpp:1190
#17 0x7776bd35 in QApplicationPrivate::notify_helper (this=, receiver=0x582fee10, e=0x7fffba20) at kernel/qapplication.cpp:3634
#18 0x55dd4241 in lyx::frontend::GuiApplication::notify 
(this=0x56586310, receiver=0x582fee10, event=0x7fffba20)
at 
/home/scott/lyxbuilds/master-master/repo/src/frontends/qt/GuiApplication.cpp:3022
#19 0x76ad8118 in QCoreApplication::notifyInternal2 
(receiver=0x582fee10, event=0x7fffba20) at 
kernel/qcoreapplication.cpp:1064
#20 0x76ad8352 in QCoreApplication::sendSpontaneousEvent 
(receiver=, event=) at 
kernel/qcoreapplication.cpp:1474
#21 0x777a8aee in QWidgetPrivate::sendPaintEvent 
(this=this@entry=0x580b1020, toBePainted=...) at kernel/qwidget.cpp:5479
#22 0x777a94a9 in QWidgetPrivate::drawWidget 
(this=this@entry=0x580b1020, pdev=, rgn=..., offset=..., 
flags=..., sharedPainter=sharedPainter@entry=0x0, 
repaintManager=) at kernel/qwidget.cpp:5429
#23 0x7777dd21 in QWidgetRepaintManager::paintAndFlush 
(this=0x570c49c0) at kernel/qwidgetrepaintmanager.cpp:1016
#24 0x777b146c in QWidget::event (this=0x56a42a00, 
event=0x59bea580) at kernel/qwidget.cpp:8990
#25 0x55e22e93 in lyx::frontend::GuiView::event (this=0x56a42a00, 
e=0x59bea580) at 

LyX 2.4.2

2024-07-03 Thread Richard Kimberly Heck

Binaries are being uploaded, so we'll do the release of 2.4.1 tomorrow.

In accord José's suggestion, my intention is to release 2.4.2 at the 
beginning of September (2 months out), unless there are sufficiently 
important fixes to justify an earlier release.


Riki


--
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: [LyX/master] Update format in lyxrc.dist

2024-07-03 Thread Richard Kimberly Heck

On 7/3/24 6:56 AM, Enrico Forestieri wrote:

On Wed, Jul 03, 2024 at 10:32:56AM +, Enrico Forestieri wrote:


commit 1dd0bf3253a7d6abd5f4b13fe46e9a4d83887b5e
Author: Enrico Forestieri 
Date:   Wed Jul 3 12:32:18 2024 +0200

   Update format in lyxrc.dist


Riki, Ok for stable?


Yes, again.

Riki


--
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: [LyX/master] Avoid bogus warnings when configuring for Qt6

2024-07-03 Thread Richard Kimberly Heck

On 7/3/24 6:55 AM, Enrico Forestieri wrote:

On Tue, Jun 18, 2024 at 09:29:58PM +0100, José Matos wrote:


On Tue, 2024-06-18 at 20:04 +, Enrico Forestieri wrote:

    Avoid bogus warnings when configuring for Qt6


Thank you. It works. :-)


Riki, Ok for stable?


Yes, thanks.

Riki


--
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: 2.4.1 Rebuilt

2024-07-03 Thread Jean-Pierre Chrétien

Le 02/07/2024 à 21:36, Richard Kimberly Heck a écrit :

Rebuilt tarballs, with a few more fixes (one crash), are here:

http://ftp.lyx.org/pub/lyx/devel/lyx-2.4/

We have had no other reports of problems, so please build binaries.


Works fine on Debian stable. I forgot a shortcut conflict,
it will have to wait for 2.4.2 :-(

--
Jean-Pierre


--
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: Assertion when selecting large table in User Guide on master

2024-07-03 Thread Jean-Marc Lasgouttes

Le 03/07/2024 à 06:17, Scott Kostyshak a écrit :

I still get an assertion while selecting the phone table.


Hmm, I was two quick in my testing.

I just pushed two patches which:
1/ avoids setting Update::SinglePar in a place where this does not make 
sense
2/ make BufferView::singleParUpdate robust in such a case (but still 
emit a warning if the position was not set, which should not happen).


Please test.

JMarc
--
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: [LyX/master] Update format in lyxrc.dist

2024-07-03 Thread Enrico Forestieri

On Wed, Jul 03, 2024 at 10:32:56AM +, Enrico Forestieri wrote:


commit 1dd0bf3253a7d6abd5f4b13fe46e9a4d83887b5e
Author: Enrico Forestieri 
Date:   Wed Jul 3 12:32:18 2024 +0200

   Update format in lyxrc.dist


Riki, Ok for stable?

--
Enrico
--
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: [LyX/master] Avoid bogus warnings when configuring for Qt6

2024-07-03 Thread Enrico Forestieri

On Tue, Jun 18, 2024 at 09:29:58PM +0100, José Matos wrote:


On Tue, 2024-06-18 at 20:04 +, Enrico Forestieri wrote:

    Avoid bogus warnings when configuring for Qt6


Thank you. It works. :-)


Riki, Ok for stable?

--
Enrico
--
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel


Re: Assertion when selecting large table in User Guide on master

2024-07-03 Thread Kornel Benko
Am Wed, 3 Jul 2024 00:17:29 -0400
schrieb Scott Kostyshak :

> On Wed, Jul 03, 2024 at 12:35:38AM GMT, Jean-Marc Lasgouttes wrote:
> > Le 02/07/2024 à 05:27, Scott Kostyshak a écrit :  
> > > I no longer get the assertion in that place. But if I work a bit harder
> > > and drag the selection up and down, eventually I get an assertion.
> > > Backtrace attached (this was while testing the patch).  
> > 
> > OK, I just pushed a different solution: instead of being robust to badly
> > updated metrics, let's do the one-line change that makes sure that
> > everything is always correctly updated.
> > 
> > Tell me how it fares.  
> 
> I still get an assertion while selecting the phone table.
> 
> I checked with current master (6e0ea426), and I don't have any patches
> applied and I did a fresh build.
> 
> I do not get an assertion while doing up-down-up-down.
> 
> Scott

Same here.

Kornel


pgpiiRjy20yyj.pgp
Description: Digitale Signatur von OpenPGP
-- 
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel