From kde-devel  Fri Dec 29 16:19:24 2023
From: "Friedrich W. H. Kossebau" <kossebau () kde ! org>
Date: Fri, 29 Dec 2023 16:19:24 +0000
To: kde-devel
Subject: Re: fix for KConfigPrivate::groupList (kconfig.cpp)
Message-Id: <3523755.V25eIC5XRa () klux>
X-MARC-Message: https://marc.info/?l=kde-devel&m=170386674214216

Am Freitag, 29. Dezember 2023, 16:36:41 CET schrieb Tommaso Massimi:
> On Fri, Dec 29, 2023 at 3:35=E2=80=AFPM Friedrich W. H. Kossebau
> <kossebau@kde.org> wrote:
[snip}
> > After a more close look I would partially disagree and point to another
>=20
> > possible cause:
> correct, my analisys was wrong, I apologize

All fine, things are so complicated we all get some things wrong in the fir=
st=20
round of working theory, actually my long reply was also edited & corrected=
 3=20
times (and perhaps still got some thing wrong) :)

> > Now, the backtrace you shared shows that there are more threads active =
at
> > the same time.
>=20
> I'm not sure:
>=20
> valgrind can shown more backtraces, in this case:
>=20
> the backtrace where the error was detected (read unallocated memory)
>=20
> > > =3D=3D70026=3D=3D Invalid read of size 16
> > > =3D=3D70026=3D=3D    at 0x668FAF7: ???
>=20
> the backtrace where that memory was allocated
>=20
> > > =3D=3D70026=3D=3D  Address 0xcd3c40a is 26 bytes inside a block of si=
ze 38
> > > alloc'd
> > > =3D=3D70026=3D=3D    at 0x4848899: malloc (in
>=20
> 70026 is the PID, reported on each line. so I see only one thread.

Ah, I might have misread the log as being a execution stack backtrace. So m=
y=20
conclusion of this being a threading issue might also be wrong?`

I guess you best upload the full log of valgrind (including how to reproduc=
e=20
what you see, i.e. which tool you used with which flag) to a bug report on=
=20
bugs.kde.org.

Cheers
=46riedrich