I could not reproduce on Ubuntu 20.04 with GNOME 3.36.8 and:

Version: 7.5.1.2 (X86_64) / LibreOffice Community
Build ID: fcbaee479e84c6cd81291587d2ee68cba099e129
CPU threads: 8; OS: Linux 5.15; UI render: default; VCL: gtk3
Locale: en-AU (en_AU.UTF-8); UI: en-US
Calc: threaded

With LO 7.4, I saw a different issue. It's Orca that crashed, but not LibreOffice. Ubuntu then opened an internal error report "orca crashed with SIGABRT". However, I wasn't able to reproduce it.

Version: 7.4.6.2 / LibreOffice Community
Build ID: 5b1f5509c2decdade7fda905e3e1429a67acd63d
CPU threads: 8; OS: Linux 5.15; UI render: default; VCL: gtk3
Locale: en-AU (en_AU.UTF-8); UI: en-US
Calc: threaded

Can you please share the information copied from Help > About LibreOffice?
Does this happen with all LO components? Instantly as soon as Orca is turned on?

From About I get:

Version: 7.4.6.2
Build ID: 40(Build:2)
CPU threads: 16; OS: Linux 6.2; UI render: default; VCL: gtk3
Locale: de-DE (de_DE.utf8); UI: de-DE
Calc: threaded

The problem happen as well for Calc and Writer. I had no other LO applications open.

A new test with all kind of documents opened, was not successful, no crash!

After thinking why, I think that my tests occurred some tines while stopping and starting Orca. At some tine I wanted to check the exchanges between orca and writer and the crash happened,

Cannot reproduce with Orca 43.1 and

Version: 7.6.0.0.alpha0+ (X86_64) / LibreOffice Community
Build ID: 457af772f52e447ef4d2dd22c1eb4624e50e6df7
CPU threads: 4; OS: Linux 6.1; UI render: default; VCL: gtk3
Locale: en-GB (en_GB.UTF-8); UI: en-US
Calc: threaded

on Debian testing with just starting Orca while Writer is running.

(In reply to jj.sarton from comment #2)
> A new test with all kind of documents opened, was not successful, no crash!
> 
> After thinking why, I think that my tests occurred some tines while stopping
> and starting Orca. At some tine I wanted to check the exchanges between orca
> and writer and the crash happened,

Can you try to find a way/scenario to reliably make it crash? Otherwise it will be hard to analyze and fix any potential issue here.

I will try, but I can't promise e result. If this has to do with timing or suchthinks, it will not be easy to find e method which reproduce the crash.

Thanks. Otherwise, if it's no more reproducible, would it be OK to close this bug as WORKSFORME for now? It can still be reopened once the crash reappears, ideally with more knowledge on how to provoke it.

I had already such a crash for a few days.
Of course, m< use of orca is not normal, since I have stopped and started a few time.
for the last start at which the crash appeared, I have configured orca for debugging in
order to see what for information are delivered from Writer.

At writing this, Calc was opened, I started Orca and the error appeared again.

Here it is late evening, so I need some hours before checking this again.

(In reply to jj.sarton from comment #6)
> Here it is late evening, so I need some hours before checking this again.

No need to hurry, thanks for looking further into this whenever you find time!

If you don't see any clear pattern/steps of how to reproduce it but it still happens every once in a while, what can be very helpful is to reproduce the crash and create a backtrace with a LibreOffice daily (development) build that has debug symbols, since that will show where in LO the crash happens.

Some more details here:

https://wiki.documentfoundation.org/QA/Testing_Daily_Builds
https://wiki.documentfoundation.org/QA/BugReport/Debug_Information

To get a daily build that contains debug symbols, you can use the link for "Linux-rpm_deb-x86_64@tb87-TDF-dbg" at https://dev-builds.libreoffice.org/daily/master/current.html .

I had tried to reproduce the crash, i what not successful. My idea for this what now to
try to get a crash dump, as you have suggested. Thanks for the information how I may do this.
The crashes seem to only appear occasionally, so I may not be able to send a crash dump soon.

(In reply to jj.sarton from comment #9)
> I had tried to reproduce the crash, i what not successful. My idea for this
> what now to
> try to get a crash dump, as you have suggested. Thanks for the information
> how I may do this.
> The crashes seem to only appear occasionally, so I may not be able to send a
> crash dump soon.

Hello,
Any update on this ?

I'm closing this as a duplicate of tdf#155449 since it sounds like the same issue.

*** This bug has been marked as a duplicate of bug 155449 ***