Bug 257944

Created attachment 53774 [details]
New crash information added by DrKonqi

dolphin (1.5.9) on KDE Platform 4.5.80 (4.6 Beta1) using Qt 4.7.1

- What I was doing when the application crashed: I close dolphin, after some clicking on icons

-- Backtrace (Reduced):
#18 0x0804e529 in _start ()

[Comment from a bug triager]
From bug 257951:
-- Information about the crash:
Start Dolphin, select some Folder, unselect the information window (F11),
select another folder, close dolphin -> crash

From bug 258093:
-- Information about the crash:
Dolphin will randomly crash on exit - does not always happen. Does not matter
what is being displayed.

*** Bug 257951 has been marked as a duplicate of this bug. ***

*** Bug 258093 has been marked as a duplicate of this bug. ***

[Comment from a bug triager]
Note that the reporter of bug 254249 (a similar crash) said that downgrading the version of the "strigi" library to 0.7.2 fixed this crash.
- What is your "strigi" library version ?
Thanks

My strigi is version 0.7.3:

libstrigi0-32bit-0.7.3-41.1.x86_64
strigi-0.7.3-41.1.x86_64
strigi-devel-0.7.3-41.1.x86_64
libstrigi0-0.7.3-41.1.x86_64

*** Bug 257158 has been marked as a duplicate of this bug. ***

*** Bug 258259 has been marked as a duplicate of this bug. ***

Tried to downgrade libstrigi to 0.7.2 to check whether the crashing stops with the lower version. Unfortunally, downgrading to libstrigi0-0.7.2 has to many dependencies and would downgrade my whole desktop.

Created attachment 53905 [details]
New crash information added by DrKonqi

dolphin (1.5.9) on KDE Platform 4.5.80 (4.6 Beta1) using Qt 4.7.1

Whenever I close Dolphin in KDE SC 4.6 Beta 1, it crashes.

-- Backtrace (Reduced):
#18 0x0804e529 in _start ()

For me dolphin does not crash with the latest strigi checkout from svn.

@Sven: Where is libstrigi SVN? The SourceForge page of the Strigi project lists version 0.6.x as latest version and seems somewhat outdated.

(In reply to comment #12)
> @Sven: Where is libstrigi SVN? The SourceForge page of the Strigi project lists
> version 0.6.x as latest version and seems somewhat outdated.

It's at: svn://anonsvn.kde.org/home/kde/trunk/kdesupport/strigi

I downloaded the latest strigi files, compiled and installed them. After restarting KDE, Dolphin keeps on crashing after closing it.

How I tested:
- info pane enabled
- hovering over a PDF to display meta data in tooltip
- closing Dolphin: crash

Ah, crap, forgot to give CMAKE the param "-DLIB_SUFFIX=64" for 64bit systems! Neverthesless: Installed again (this time in the correct directory /usr/lib64), restarted KDE and Dolphin keeps on crashing. :(

Created attachment 53932 [details]
New crash information added by DrKonqi

dolphin (1.5.9) on KDE Platform 4.5.80 (4.6 Beta1) using Qt 4.7.1

- What I was doing when the application crashed:
Dolphin crashes on exit (KDE 4.5.80). Usually if I open any file from it.

-- Backtrace (Reduced):
#13 0x0000000000407610 in _start ()

Then there is more to it, because that works here without crash. i586 here with openSUSE 11.3 and the KDE packages from KDE:Distro:Factory.

[Comment from a bug triager]
From bug 258723:
Opening and closing Dolphin does not result in a crash; however, ANY activity
in Dolphin will lead to a crash on exit - previewing files, opening files,
browsing folders, whatever.  Types of files don't matter.
This can be replicated EVERY time.

*** Bug 258723 has been marked as a duplicate of this bug. ***

*** Bug 258809 has been marked as a duplicate of this bug. ***

Bug 240013, which looks similar, has a Valgrind log (not very detailed though).

*** Bug 258934 has been marked as a duplicate of this bug. ***

[Comment from a bug triager]
Note: all those crashes are happening on OpenSuse installations ... (even different versions, 11.2, 11.3 and 11.4 devel).
May be the KDE SC 4.6 packages pulled some other package upgrade that causes this crash ?

Bug 259063 seems to be related to the NVidia driver. Is anyone else here using the NVidia driver ? (may be it is not related at all)

Thanks

Using nVidia here...

No NVIDIA here, running on integrated Intel Arrandale graphics with open source intel driver.

Running ati radeon r100 with nomodeset.

On Dec 7, 2010 2:19 AM, "Andreas Demmer" <mail@andreas-demmer.de> wrote:

https://bugs.kde.org/show_bug.cgi?id=257944




--- Comment #25 from Andreas Demmer <mail andreas-demmer de>  2010-12-07
08:19:40 ---
No NVIDIA here, running on integrated Intel Arrandale graphics with open
source
intel driver.

[Comment from a bug triager]
From bug 259371:
-- Information about the crash:
- What I was doing when the application crashed:
Closed Dolphin with X button after searching several folders .  This happens
frequently, but seldom if I simply open and exit dolphin quickly.  Any Dolphin
window sitting on the desktop for more than 10 minutes seems to crash on exit.
I have some samba shares mapped to the places panel, but was not using or
surfing those shares.  KDE4 4.6 beta2 from Opensuse.

*** Bug 259371 has been marked as a duplicate of this bug. ***

[Comment from a bug triager]
I have installed OpenSuse 11.4 on a 686 VM and I could reproduce the crash
under certain conditions (however, DrKonqi didn't appear). In the valgrind log
there were references to libstreams (from Strigi), but even installing the
debuginfo packages I couldn't get a full trace...

[Comment from a bug triager]
- From bug 259263:
I was viewing the contects of a DVD and opened konsole from the Tools menu to
launch a shell script on the disk. After about 15 minutes I closed Dolphin by
right clicking it on the task bar and selecting close. I attemted to reproduce
the bug twice but was unable to do so. This bug occurred twice yesterday under
similar conditions but not reported.

From bug 259063:
Steps to reproduce:
open dolphin
hover over some file(folder is not enough somehow)
close dolphin

*** Bug 259263 has been marked as a duplicate of this bug. ***

*** Bug 259064 has been marked as a duplicate of this bug. ***

*** Bug 259063 has been marked as a duplicate of this bug. ***

Created attachment 54411 [details]
New crash information added by DrKonqi

dolphin (1.5.9) on KDE Platform 4.5.85 (4.6 Beta2) using Qt 4.7.1

- What I was doing when the application crashed:

Dolphin immediately crashed when closing

-- Backtrace (Reduced):
#13 0x0000000000407610 in _start ()

Created attachment 54430 [details]
New crash information added by DrKonqi

dolphin (1.5.9) on KDE Platform 4.5.80 (4.6 Beta1) using Qt 4.7.1

- What I was doing when the application crashed:

just finished transfering files from my osx10.6.5 box to this one. once it finished, closed dolphin after unziping a file with ark. crashed. does this every time. running an ati radeon vivo 64mb with opengl and nomodeset on boot.

-- Backtrace (Reduced):
#18 0x0804e529 in _start ()

[Comment from a bug triager]
Another possible duplicate from Gwenview: bug 259350.
Other note: on my OpenSuse VM I could reproduce the crash invoking "kfile4 --dialog SomeLauncher.desktop", waiting 5 seconds or more and closing the dialog.

I should note as well, that I deleted my .kde4 folder recently, and the
issue still shows up. I will attempt the kfile command soon and report
back.

On Sat, Dec 11, 2010 at 11:08 AM, Dario Andres
<andresbajotierra@gmail.com>wrote:

> https://bugs.kde.org/show_bug.cgi?id=257944
>
>
> Dario Andres <andresbajotierra@gmail.com> changed:
>
>           What    |Removed                     |Added
>
> ----------------------------------------------------------------------------
>                  CC|                            |jos@vandenoever.info
>
>
>
>
> --- Comment #36 from Dario Andres <andresbajotierra gmail com>  2010-12-11
> 17:08:46 ---
> [Comment from a bug triager]
> Another possible duplicate from Gwenview: bug 259350.
> Other note: on my OpenSuse VM I could reproduce the crash invoking "kfile4
> --dialog SomeLauncher.desktop", waiting 5 seconds or more and closing the
> dialog.
>
> --
> Configure bugmail: https://bugs.kde.org/userprefs.cgi?tab=email
> ------- You are receiving this mail because: -------
> You are on the CC list for the bug.
>

(In reply to comment #36)
> Other note: on my OpenSuse VM I could reproduce the crash invoking "kfile4
> --dialog SomeLauncher.desktop", waiting 5 seconds or more and closing the
> dialog.

Note that bug 258031 comment 2 has a test file attached which crashes rdfindexer/xmlindexer with the same bt according to the reporter. Could you try that in your VM, Darío? In that case, it would pretty much look like a Strigi issue, I think.

I actually tried to find the source of Strigi 0.7.3 to build it and try it myself, but I was unable to find an official release. I couldn't find a 0.7.3 tag on KDE's svn server or on http://gitorious.org/strigi either.

[Comment from a bug triager]
OpenSuse updated Strigi to "0.7.3-42.1" and Valgrind still detects the crashes with my method using "kfile4".
I also tested xmlindexer on the file Frank mentioned (https://bugs.kde.org/attachment.cgi?id=54396). xmlindexer doesn't crash itself(there isn't a backtrace), however Valgrind still recognizes an invalid free/delete (I can't get full debuginfo yet)

I requested help to the opensuse-kde team, sending a mail to their mailing list. Hopefully a OpenSuse dev or expert user can help us in the near future

Thanks

[Comment from a bug triager]
Mh, just run "valgrind xmlindexer" on a folder with only one text file (with some random letters I just wrote) and it still detects the invalid free/delete on the end. So I don't even know if it is related to that crash, or if Valgrind is lying to me....

*** Bug 259541 has been marked as a duplicate of this bug. ***

Reproduced with Dolphin from the 4.5.86 snapshot.

but *only* if Dolphin is ran from konsole. Using krunner to start dolphin doesn't trigger the issue for me.

Valgrind didn't show anything helpful yet.

*** Bug 259646 has been marked as a duplicate of this bug. ***

*** Bug 259629 has been marked as a duplicate of this bug. ***

From bug 259719:
Start Konqueror
 F9 to get side bar open
Select a folder in home directory
Attempt Drag and Drop from Konqueror > Scribus 
Close Konqueror
Crash and Backtrace

*** Bug 259719 has been marked as a duplicate of this bug. ***

*** Bug 259824 has been marked as a duplicate of this bug. ***

*** Bug 259294 has been marked as a duplicate of this bug. ***

*** Bug 259014 has been marked as a duplicate of this bug. ***

(In reply to comment #42)
> but *only* if Dolphin is ran from konsole. Using krunner to start dolphin
> doesn't trigger the issue for me.

Looks like I can reproduce the crash every time now

Created attachment 54828 [details]
New crash information added by DrKonqi

dolphin (1.5.9) on KDE Platform 4.5.85 (4.6 Beta2) using Qt 4.7.1

- What I was doing when the application crashed:

Opened Dolphin, deleted 2 GB File, crash on closing.

-- Backtrace (Reduced):
#18 0x0804e529 in _start ()

*** Bug 260904 has been marked as a duplicate of this bug. ***

Created attachment 55156 [details]
New crash information added by DrKonqi

dolphin (1.5.9) on KDE Platform 4.5.85 (4.6 Beta2) using Qt 4.7.1

Dolphin crashes everytime. No matter what operation I do, it always shows me crash report saying it was having Segmentation Fault.

I am using openSUSE 11.3 with KDE 4.5.85 (KDE 4.6 Beta 2)

-- Backtrace (Reduced):
#13 0x0000000000407610 in _start ()

*** Bug 261127 has been marked as a duplicate of this bug. ***

Created attachment 55259 [details]
New crash information added by DrKonqi

dolphin (1.5.9) on KDE Platform 4.5.90 (4.6 RC1) using Qt 4.7.1

Still crashing after upgraded to 4.5.90 (RC1)

-- Backtrace (Reduced):
#13 0x0000000000407610 in launch (argc=5, _name=0x6aaa68 "/usr/bin/dolphin", args=<value optimized out>, cwd=0x0, envc=<value optimized out>, envs=<value optimized out>, reset_env=false, tty=0x0, 
    avoid_loops=false, startup_id_str=0x6aaac1 "teras;1293362871;699626;4089_TIME11543159") at /usr/src/debug/kdelibs-4.5.90/kinit/kinit.cpp:730
#14 0x0000000000408125 in handle_launcher_request (sock=8, who=<value optimized out>) at /usr/src/debug/kdelibs-4.5.90/kinit/kinit.cpp:1222
#15 0x000000000040889a in handle_requests (waitForPid=<value optimized out>) at /usr/src/debug/kdelibs-4.5.90/kinit/kinit.cpp:1415
#16 0x00000000004094f0 in main (argc=4, argv=0x7fffb26eb8d8, envp=0x7fffb26eb900) at /usr/src/debug/kdelibs-4.5.90/kinit/kinit.cpp:1903

Created attachment 55303 [details]
New crash information added by DrKonqi

dolphin (1.5.9) on KDE Platform 4.5.90 (4.6 RC1) using Qt 4.7.1

- What I was doing when the application crashed:

I just opened dolphin, browse to "My Documents" folder and opened one .pdf file with "Ocular". Then I just close Ocular and then dolphin and it crashes on close...

-- Backtrace (Reduced):
#14 0x0000000000400699 in _start () at ../sysdeps/x86_64/elf/start.S:113

*** Bug 261760 has been marked as a duplicate of this bug. ***

[Comment from a bug triager]
Renaming report and merging all the other KDE application crashes:

Plasma crashes: bug 258485, bug 261749, bug 261120
Digikam: bug 259781
Krusader: bug 258977
SystemSetting: bug 258877

*** Bug 257909 has been marked as a duplicate of this bug. ***

*** Bug 259781 has been marked as a duplicate of this bug. ***

*** Bug 258977 has been marked as a duplicate of this bug. ***

*** Bug 258877 has been marked as a duplicate of this bug. ***

*** Bug 258485 has been marked as a duplicate of this bug. ***

*** Bug 261749 has been marked as a duplicate of this bug. ***

*** Bug 261120 has been marked as a duplicate of this bug. ***

*** Bug 262043 has been marked as a duplicate of this bug. ***

*** Bug 262404 has been marked as a duplicate of this bug. ***

*** Bug 262407 has been marked as a duplicate of this bug. ***

*** Bug 240013 has been marked as a duplicate of this bug. ***

*** Bug 262023 has been marked as a duplicate of this bug. ***

*** Bug 262616 has been marked as a duplicate of this bug. ***

*** Bug 258562 has been marked as a duplicate of this bug. ***

*** Bug 262018 has been marked as a duplicate of this bug. ***

*** Bug 262352 has been marked as a duplicate of this bug. ***

*** Bug 259350 has been marked as a duplicate of this bug. ***

*** Bug 244983 has been marked as a duplicate of this bug. ***

*** Bug 251149 has been marked as a duplicate of this bug. ***

*** Bug 235420 has been marked as a duplicate of this bug. ***

*** Bug 238412 has been marked as a duplicate of this bug. ***

Bug 232602 may be the oldest report we have, so let's use it as an estimate to look for regressions. It is dated 2010-03-29, and uses KDE 4.4.69 (early KDE 4.5 alpha) from the same week.

Unfortunately, that's exactly the week after the "big strigi repo reorganization", which makes it much harder to check for changes. Some "post reorg" changes, such as those around 2010-03-21 might be possible candidates, too.

Since this bug seems only present on openSUSE, there could also be a change dated around that week in strigi or libstdc++ (which is part of the gcc package) within the openSUSE repos. Has the bug been reported to the openSUSE project?

Some time ago, I talked about this on #opensuse-kde; I also sent an email to a member of their community and he promised to forward it to the devel opensuse's mailing list.
I just found the mail: http://lists.suse.com/opensuse-kde/2010-12/msg00054.html

I can reproduce the same crash backtrace with "strigiclient" alone.
Just open it and close it -> crash.
strigiclient seems to have nothing to do with any KDE libs, so the problem seems to not KDE related.

Even simpler: I get the same crash when running "strigicmd --help"

Hmmm ... I don't know what to make out of this:
When I compile strigi-0.7.3 (source from openSuse strigi-debugsource-0.7.3-47.3.i586) and run the newly compiled strigicmd --help from this compile directory, which pulls in its own compiled libs
/usr/src/packages/SOURCES/strigi-0.7.3/build/libstreamanalyzer.so.0.7
/usr/src/packages/SOURCES/strigi-0.7.3/build/libstreams.so.0
the crash backtrace does NOT appear.
- When I run the installed /usr/bin/strigicmd against my newly compiled strigi libs, the crash APPEARS

The only thing ccmake configure tells me I do not use is:
** CLucene1 not found. Support for CLucene Version 1+ backend is disabled
** Log4cxx not found. Support for advanced logging is disabled
 Java5 found FALSE

everything else is found.

Good news 
strigicmd --help
strigicmd:
    Program for creating and querying indices.
    This program is not meant for talking to the strigi daemon.

usage:
  strigicmd create [-j num] -t backend -d indexdir [-i include] [-x exclude] files/dirs
  strigicmd deindex -t backend -d indexdir files/dirs
  strigicmd get -t backend -d indexdir files
  strigicmd listFiles -t backend -d indexdir [parent dir]
  strigicmd listFields -t backend -d indexdir
  strigicmd query -t backend -d indexdir queries
[-q xesam_query_language_file]
  strigicmd update [-j num] -t backend -d indexdir [-i include] [-x exclude] files/dirs

with lastest kde factory build 4.6-RC2 under 11.4-factory
and those installed packages

S | Nom                        | Type   | Version    | Arch   | Repo           
--+----------------------------+--------+------------+--------+-----------------
i | libstrigi0                 | paquet | 0.7.3-47.6 | x86_64 | factory-kde-core
i | libstrigi0-32bit           | paquet | 0.7.3-47.6 | x86_64 | factory-kde-core
i | libstrigi0-debuginfo       | paquet | 0.7.3-47.6 | x86_64 | factory-kde-core
i | libstrigi0-debuginfo-32bit | paquet | 0.7.3-47.6 | x86_64 | factory-kde-core
i | strigi                     | paquet | 0.7.3-47.6 | x86_64 | factory-kde-core
i | strigi-debuginfo           | paquet | 0.7.3-47.6 | x86_64 | factory-kde-core
i | strigi-debugsource         | paquet | 0.7.3-47.6 | x86_64 | factory-kde-core
i | strigi-devel               | paquet | 0.7.3-47.6 | x86_64 | factory-kde-core

That does not necessarily tell you if the bug is gone, as I discovered that the crash can also only be visible with valgrind.
Probably because it is so late during the exit of the application.

So please check "valgrind strigicmd --help"

Thanks Martin, you kill my joy :D

here the things now
valgrind -v strigicmd --help
==7957== Memcheck, a memory error detector
==7957== Copyright (C) 2002-2010, and GNU GPL'd, by Julian Seward et al.
==7957== Using Valgrind-3.6.0 and LibVEX; rerun with -h for copyright info
==7957== Command: strigicmd --help
==7957== 
--7957-- Valgrind options:
--7957--    -v
--7957-- Contents of /proc/version:
--7957--   Linux version 2.6.37-rc7-desktop (geeko@buildhost) (gcc version 4.5.1 20101208 [gcc-4_5-branch revision 167585] (SUSE Linux) ) #1 SMP PREEMPT 2010-12-28 07:01:24 +0100
--7957-- Arch and hwcaps: AMD64, amd64-sse3-cx16
--7957-- Page sizes: currently 4096, max supported 4096
--7957-- Valgrind library directory: /usr/lib64/valgrind
--7957-- Reading syms from /usr/bin/strigicmd (0x400000)
--7957--   Considering /usr/lib/debug/.build-id/65/711e95222fd7b5f6d395f57d896a1f4b849da6.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /lib64/ld-2.11.3.so (0x4000000)
--7957--   Considering /usr/lib/debug/.build-id/5c/fc5a2c4891477ba3f389a7f24582df1496bd89.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /usr/lib64/valgrind/memcheck-amd64-linux (0x38000000)
--7957--   Considering /usr/lib/debug/usr/lib64/valgrind/memcheck-amd64-linux.debug ..
--7957--   .. CRC is valid
--7957--    object doesn't have a dynamic symbol table
--7957-- Reading suppressions file: /usr/lib64/valgrind/default.supp
--7957-- REDIR: 0x4015e90 (strlen) redirected to 0x38042017 (vgPlain_amd64_linux_REDIR_FOR_strlen)
--7957-- Reading syms from /usr/lib64/valgrind/vgpreload_core-amd64-linux.so (0x4a20000)
--7957--   Considering /usr/lib/debug/.build-id/36/c68b042fbd64d5afa4df68dc5d84c9bd7ac6bd.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so (0x4c22000)
--7957--   Considering /usr/lib/debug/.build-id/70/eb77800f4162281d776ad96802c8f73254374a.debug ..
--7957--   .. build-id is valid
==7957== WARNING: new redirection conflicts with existing -- ignoring it
--7957--     new: 0x04015e90 (strlen              ) R-> 0x04c26ee0 strlen
--7957-- REDIR: 0x4015de0 (index) redirected to 0x4c26b40 (index)
--7957-- REDIR: 0x4015e60 (strcmp) redirected to 0x4c274e0 (strcmp)
--7957-- Reading syms from /usr/lib64/libstreamanalyzer.so.0.7.3 (0x4e2c000)
--7957--   Considering /usr/lib/debug/.build-id/1c/57b9d22134021bf9d3afbb5422281eea4259d0.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /lib64/libz.so.1.2.5 (0x50ba000)
--7957--   Considering /usr/lib/debug/.build-id/04/27a08bb604a33f2973b2f8831e10518780a868.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /lib64/libbz2.so.1.0.6 (0x52d3000)
--7957--   Considering /usr/lib/debug/.build-id/b0/4e479095d5cf56565cb96cc0c383b92cf44f33.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /lib64/libpthread-2.11.3.so (0x54e2000)
--7957--   Considering /usr/lib/debug/.build-id/a6/054dbbb5b029f140440e854aba87d9c86fa120.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /usr/lib64/libstdc++.so.6.0.14 (0x56ff000)
--7957--   Considering /usr/lib/debug/.build-id/c5/6d4a89581212b96b9ae3fd172d2c5499fd0985.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /lib64/libc-2.11.3.so (0x5a08000)
--7957--   Considering /usr/lib/debug/.build-id/0d/950bde4b77aa25e40384b58280de0f1c77073b.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /lib64/libgcc_s.so.1 (0x5d75000)
--7957--   Considering /usr/lib/debug/.build-id/26/4f532e2222db3f73dba1814e7e5bda9136e173.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /usr/lib64/libstreams.so.0.7.3 (0x5f8b000)
--7957--   Considering /usr/lib/debug/.build-id/ef/d53261182d9d3172d8322a903edb8c4e81849d.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /usr/lib64/libxml2.so.2.7.8 (0x61c6000)
--7957--   Considering /usr/lib/debug/.build-id/47/5ad0e59f21b6cd41d0ef167c7e54e18450912e.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /lib64/libdl-2.11.3.so (0x651c000)
--7957--   Considering /usr/lib/debug/.build-id/1d/53250bcf26a610704c9e7eeb77747dea19806c.debug ..
--7957--   .. build-id is valid
--7957-- Reading syms from /lib64/libm-2.11.3.so (0x6720000)
--7957--   Considering /usr/lib/debug/.build-id/70/ba8a643d9be01b9c29d718c9672d6d0e07eb21.debug ..
--7957--   .. build-id is valid
--7957-- REDIR: 0x5a87000 (__GI_strrchr) redirected to 0x4c26960 (__GI_strrchr)
--7957-- REDIR: 0x5a85530 (__GI_strlen) redirected to 0x4c26ea0 (__GI_strlen)
--7957-- REDIR: 0x5a83a80 (strcmp) redirected to 0x4a20650 (_vgnU_ifunc_wrapper)
--7957-- REDIR: 0x5b142e0 (__strcmp_sse42) redirected to 0x4c27420 (strcmp)
--7957-- REDIR: 0x5a85500 (strlen) redirected to 0x4a20650 (_vgnU_ifunc_wrapper)
--7957-- REDIR: 0x5b15180 (__strlen_sse42) redirected to 0x4c26e80 (strlen)
--7957-- REDIR: 0x57bc220 (operator new(unsigned long)) redirected to 0x4c26030 (operator new(unsigned long))
--7957-- REDIR: 0x5a89190 (memcpy) redirected to 0x4c275e0 (memcpy)
--7957-- REDIR: 0x5a80c70 (calloc) redirected to 0x4c24bb0 (calloc)
--7957-- REDIR: 0x5a8a460 (strchrnul) redirected to 0x4c286b0 (strchrnul)
strigicmd:
--7957-- REDIR: 0x5a88880 (mempcpy) redirected to 0x4c287c0 (mempcpy)
    Program for creating and querying indices.
    This program is not meant for talking to the strigi daemon.

usage:
  strigicmd create [-j num] -t backend -d indexdir [-i include] [-x exclude] files/dirs
  strigicmd deindex -t backend -d indexdir files/dirs
  strigicmd get -t backend -d indexdir files
  strigicmd listFiles -t backend -d indexdir [parent dir]
  strigicmd listFields -t backend -d indexdir
  strigicmd query -t backend -d indexdir queries
[-q xesam_query_language_file]
  strigicmd update [-j num] -t backend -d indexdir [-i include] [-x exclude] files/dirs
--7957-- REDIR: 0x57ba490 (operator delete(void*)) redirected to 0x4c25320 (operator delete(void*))
--7957-- REDIR: 0x5a7fff0 (free) redirected to 0x4c25690 (free)
==7957== Invalid read of size 4
==7957==    at 0x579F2C6: std::basic_string<char, std::char_traits<char>, std::allocator<char> >::~basic_string() (atomicity.h:46)
==7957==    by 0x5A3D914: __cxa_finalize (cxa_finalize.c:56)
==7957==    by 0x5F99775: ??? (in /usr/lib64/libstreams.so.0.7.3)
==7957==    by 0x5FBBF80: ??? (in /usr/lib64/libstreams.so.0.7.3)
==7957==    by 0x5A3D580: __run_exit_handlers (exit.c:78)
==7957==    by 0x5A3D5D4: exit (exit.c:100)
==7957==    by 0x5A26BC3: (below main) (libc-start.c:258)
==7957==  Address 0x697ee70 is 16 bytes inside a block of size 30 free'd
==7957==    at 0x4C2538E: operator delete(void*) (vg_replace_malloc.c:387)
==7957==    by 0x579F2D8: std::basic_string<char, std::char_traits<char>, std::allocator<char> >::~basic_string() (basic_string.h:237)
==7957==    by 0x5A3D580: __run_exit_handlers (exit.c:78)
==7957==    by 0x5A3D5D4: exit (exit.c:100)
==7957==    by 0x5A26BC3: (below main) (libc-start.c:258)
==7957== 
==7957== Invalid free() / delete / delete[]
==7957==    at 0x4C2538E: operator delete(void*) (vg_replace_malloc.c:387)
==7957==    by 0x579F2D8: std::basic_string<char, std::char_traits<char>, std::allocator<char> >::~basic_string() (basic_string.h:237)
==7957==    by 0x5A3D914: __cxa_finalize (cxa_finalize.c:56)
==7957==    by 0x5F99775: ??? (in /usr/lib64/libstreams.so.0.7.3)
==7957==    by 0x5FBBF80: ??? (in /usr/lib64/libstreams.so.0.7.3)
==7957==    by 0x5A3D580: __run_exit_handlers (exit.c:78)
==7957==    by 0x5A3D5D4: exit (exit.c:100)
==7957==    by 0x5A26BC3: (below main) (libc-start.c:258)
==7957==  Address 0x697ee60 is 0 bytes inside a block of size 30 free'd
==7957==    at 0x4C2538E: operator delete(void*) (vg_replace_malloc.c:387)
==7957==    by 0x579F2D8: std::basic_string<char, std::char_traits<char>, std::allocator<char> >::~basic_string() (basic_string.h:237)
==7957==    by 0x5A3D580: __run_exit_handlers (exit.c:78)
==7957==    by 0x5A3D5D4: exit (exit.c:100)
==7957==    by 0x5A26BC3: (below main) (libc-start.c:258)
==7957== 
==7957== 
==7957== HEAP SUMMARY:
==7957==     in use at exit: 59 bytes in 2 blocks
==7957==   total heap usage: 235 allocs, 235 frees, 15,833 bytes allocated
==7957== 
==7957== Searching for pointers to 2 not-freed blocks
==7957== Checked 251,304 bytes
==7957== 
==7957== LEAK SUMMARY:
==7957==    definitely lost: 59 bytes in 2 blocks
==7957==    indirectly lost: 0 bytes in 0 blocks
==7957==      possibly lost: 0 bytes in 0 blocks
==7957==    still reachable: 0 bytes in 0 blocks
==7957==         suppressed: 0 bytes in 0 blocks
==7957== Rerun with --leak-check=full to see details of leaked memory
==7957== 
==7957== ERROR SUMMARY: 8 errors from 2 contexts (suppressed: 4 from 4)
==7957== 
==7957== 2 errors in context 1 of 2:
==7957== Invalid free() / delete / delete[]
==7957==    at 0x4C2538E: operator delete(void*) (vg_replace_malloc.c:387)
==7957==    by 0x579F2D8: std::basic_string<char, std::char_traits<char>, std::allocator<char> >::~basic_string() (basic_string.h:237)
==7957==    by 0x5A3D914: __cxa_finalize (cxa_finalize.c:56)
==7957==    by 0x5F99775: ??? (in /usr/lib64/libstreams.so.0.7.3)
==7957==    by 0x5FBBF80: ??? (in /usr/lib64/libstreams.so.0.7.3)
==7957==    by 0x5A3D580: __run_exit_handlers (exit.c:78)
==7957==    by 0x5A3D5D4: exit (exit.c:100)
==7957==    by 0x5A26BC3: (below main) (libc-start.c:258)
==7957==  Address 0x697ee60 is 0 bytes inside a block of size 30 free'd
==7957==    at 0x4C2538E: operator delete(void*) (vg_replace_malloc.c:387)
==7957==    by 0x579F2D8: std::basic_string<char, std::char_traits<char>, std::allocator<char> >::~basic_string() (basic_string.h:237)
==7957==    by 0x5A3D580: __run_exit_handlers (exit.c:78)
==7957==    by 0x5A3D5D4: exit (exit.c:100)
==7957==    by 0x5A26BC3: (below main) (libc-start.c:258)
==7957== 
==7957== 
==7957== 6 errors in context 2 of 2:
==7957== Invalid read of size 4
==7957==    at 0x579F2C6: std::basic_string<char, std::char_traits<char>, std::allocator<char> >::~basic_string() (atomicity.h:46)
==7957==    by 0x5A3D914: __cxa_finalize (cxa_finalize.c:56)
==7957==    by 0x5F99775: ??? (in /usr/lib64/libstreams.so.0.7.3)
==7957==    by 0x5FBBF80: ??? (in /usr/lib64/libstreams.so.0.7.3)
==7957==    by 0x5A3D580: __run_exit_handlers (exit.c:78)
==7957==    by 0x5A3D5D4: exit (exit.c:100)
==7957==    by 0x5A26BC3: (below main) (libc-start.c:258)
==7957==  Address 0x697ee70 is 16 bytes inside a block of size 30 free'd
==7957==    at 0x4C2538E: operator delete(void*) (vg_replace_malloc.c:387)
==7957==    by 0x579F2D8: std::basic_string<char, std::char_traits<char>, std::allocator<char> >::~basic_string() (basic_string.h:237)
==7957==    by 0x5A3D580: __run_exit_handlers (exit.c:78)
==7957==    by 0x5A3D5D4: exit (exit.c:100)
==7957==    by 0x5A26BC3: (below main) (libc-start.c:258)
==7957== 
--7957-- 
--7957-- used_suppression:      2 U1004-ARM-_dl_relocate_object
--7957-- used_suppression:      2 glibc-2.5.x-on-SUSE-10.2-(PPC)-2a
==7957== 
==7957== ERROR SUMMARY: 8 errors from 2 contexts (suppressed: 4 from 4)

Bruno, thanks for your valgrind log. It lacks the most important information, though: the debug information (including source line numbers) for the strigi package.

Hi Christophe, I don't know how to get it. As I work only with compiled rpm, I've also all debuginfo and debugsource installed as well. 

But discussing with packager on opensuse, with new version of strigi, he get that result 

valgrind -v strigicmd --help
==7100== Memcheck, a memory error detector
==7100== Copyright (C) 2002-2010, and GNU GPL'd, by Julian Seward et al.
==7100== Using Valgrind-3.6.0 and LibVEX; rerun with -h for copyright info
==7100== Command: strigicmd --help
==7100== 
--7100-- Valgrind options:
--7100--    -v
--7100-- Contents of /proc/version:
--7100--   Linux version 2.6.37-rc7-default (geeko@buildhost) (gcc version 4.5.1 20101208 [gcc-4_5-branch revision 167585] (SUSE Linux) ) #1 SMP 2010-12-28 07:01:24 +0100
--7100-- Arch and hwcaps: X86, x86-sse1-sse2
--7100-- Page sizes: currently 4096, max supported 4096
--7100-- Valgrind library directory: /usr/lib/valgrind
--7100-- Reading syms from /lib/ld-2.11.3.so (0x4000000)
--7100-- Reading syms from /usr/bin/strigicmd (0x8048000)
--7100-- Reading syms from /usr/lib/valgrind/memcheck-x86-linux (0x38000000)
--7100--    object doesn't have a symbol table
--7100--    object doesn't have a dynamic symbol table
--7100-- Reading suppressions file: /usr/lib/valgrind/default.supp
--7100-- REDIR: 0x4016e80 (index) redirected to 0x3803fcbf (???)
--7100-- Reading syms from /usr/lib/valgrind/vgpreload_core-x86-linux.so (0x4022000)
--7100--    object doesn't have a symbol table
--7100-- Reading syms from /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so (0x4025000)
--7100--    object doesn't have a symbol table
==7100== WARNING: new redirection conflicts with existing -- ignoring it
--7100--     new: 0x04016e80 (index               ) R-> 0x04029380 index
--7100-- REDIR: 0x4017020 (strlen) redirected to 0x40297c0 (strlen)
--7100-- Reading syms from /usr/lib/libstreamanalyzer.so.0.7.3 (0x4030000)
--7100-- Reading syms from /usr/lib/libstreams.so.0.7.3 (0x416f000)
--7100-- Reading syms from /lib/libz.so.1.2.5 (0x420c000)
--7100--    object doesn't have a symbol table
--7100-- Reading syms from /lib/libbz2.so.1.0.6 (0x4224000)
--7100--    object doesn't have a symbol table
--7100-- Reading syms from /usr/lib/libxml2.so.2.7.8 (0x4234000)
--7100--    object doesn't have a symbol table
--7100-- Reading syms from /lib/libdl-2.11.3.so (0x4381000)
--7100-- Reading syms from /lib/libpthread-2.11.3.so (0x4386000)
--7100-- Reading syms from /usr/lib/libstdc++.so.6.0.14 (0x43a1000)
--7100--    object doesn't have a symbol table
--7100-- Reading syms from /lib/libm-2.11.3.so (0x4490000)
--7100-- Reading syms from /lib/libc-2.11.3.so (0x44bb000)
--7100-- Reading syms from /lib/libgcc_s.so.1 (0x4627000)
--7100--    object doesn't have a symbol table
--7100-- REDIR: 0x45380f0 (strstr) redirected to 0x40224e0 (_vgnU_ifunc_wrapper)
--7100-- REDIR: 0x4537ce0 (__GI_strstr) redirected to 0x402b340 (strstr)
--7100-- REDIR: 0x4531810 (rindex) redirected to 0x40291f0 (rindex)
--7100-- REDIR: 0x4531430 (__GI_strlen) redirected to 0x40297a0 (__GI_strlen)
--7100-- REDIR: 0x4530d70 (strcmp) redirected to 0x4029d60 (strcmp)
--7100-- REDIR: 0x4531370 (strlen) redirected to 0x40224e0 (_vgnU_ifunc_wrapper)
--7100-- REDIR: 0x45313b0 (__strlen_sse2) redirected to 0x4029780 (strlen)
--7100-- REDIR: 0x4454d70 (operator new(unsigned int)) redirected to 0x4028980 (operator new(unsigned int))
--7100-- REDIR: 0x4532b10 (memcpy) redirected to 0x4029ee0 (memcpy)
--7100-- REDIR: 0x452e390 (calloc) redirected to 0x4027500 (calloc)
--7100-- REDIR: 0x4533c60 (strchrnul) redirected to 0x402af00 (strchrnul)
strigicmd:
--7100-- REDIR: 0x4532620 (mempcpy) redirected to 0x402b030 (mempcpy)
    Program for creating and querying indices.
    This program is not meant for talking to the strigi daemon.

usage:
  strigicmd create [-j num] -t backend -d indexdir [-i include] [-x exclude] files/dirs
  strigicmd deindex -t backend -d indexdir files/dirs
  strigicmd get -t backend -d indexdir files
  strigicmd listFiles -t backend -d indexdir [parent dir]
  strigicmd listFields -t backend -d indexdir
  strigicmd query -t backend -d indexdir queries
[-q xesam_query_language_file]
  strigicmd update [-j num] -t backend -d indexdir [-i include] [-x exclude] files/dirs
--7100-- REDIR: 0x452d850 (free) redirected to 0x4027fe0 (free)
--7100-- REDIR: 0x4452b00 (operator delete(void*)) redirected to 0x4027c70 (operator delete(void*))
==7100== 
==7100== HEAP SUMMARY:
==7100==     in use at exit: 0 bytes in 0 blocks
==7100==   total heap usage: 233 allocs, 233 frees, 10,946 bytes allocated
==7100== 
==7100== All heap blocks were freed -- no leaks are possible
==7100== 
==7100== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 3 from 3)
--7100-- 
--7100-- used_suppression:      3 U1004-ARM-_dl_relocate_object
==7100== 
==7100== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 3 from 3)

Made by him with the lastest git checkout of strigi .
So not yet released .

Created attachment 55789 [details]
valgrind --log-file=nocrash -v gwenview

@Comment #36: Gwenview 2.5.0, openSUSE 11.2 (i586), 4.5.95 (4.6 RC2)
Created a new folder with 1 png-file in dolphin, opened folder with gwenview
--> no crash

Created attachment 55790 [details]
valgrind --log-file=crash -v gwenview

@Comment #36: Gwenview 2.5.0, openSUSE 11.2 (i586), 4.5.95 (4.6 RC2)
Added a second png-file in dolphin, opened folder with 2 png-files with gwenview:

--> gwenview crashes

ad comment #89: As said in comment #84 I do also not get the crash when I compile the currently released strigi from source!

So I don't think it actually has something to do with the strigi source itself.
To prove that, I wanted to compare current git sources with current released sources, but 
git clone https://projects.kde.org/projects/kdesupport/strigi/strigidaemon/repository
gives:
Initialized empty Git repository in /usr/src/packages/SOURCES/strigi/repository/.git/
warning: remote HEAD refers to nonexistent ref, unable to checkout.

???

http://gitweb.kde.org/strigidaemon.git should explain that you have to use:

git clone git://anongit.kde.org/strigidaemon

Thanks. Altough http://websvn.kde.org/trunk/kdesupport/strigi/STRIGI_MOVED_TO_GIT?view=markup
says different.

I checked the code of strigidaemon, libstreams and libstreamanalyzer and I see no change (and there are really just very few changes) which would cause already the start of strigicmd with the help output to fail (or no longer fail).

Probably the "bug" is in some other library, although strigicmd links only against a few libs. Or the compiler settings (e.g. CMAKE_BUILD_TYPE, optimizer flags, etc.) are different in the delivered rpm than what I used.

Created attachment 55892 [details]
New crash information added by DrKonqi

dolphin (1.5.9) on KDE Platform 4.5.85 (4.6 Beta2) using Qt 4.7.1

- What I was doing when the application crashed:

I was on my way closing applications before laptop shutdown.
Dolpĥin was the last close application, when crash occurred.

-- Backtrace (Reduced):
#10 0xb6126f27 in __libc_message (do_abort=2, fmt=0xb61ef0b4 "*** glibc detected *** %s: %s: 0x%s ***\n") at ../sysdeps/unix/sysv/linux/libc_fatal.c:186
#11 0xb612cf6b in malloc_printerr (action=<value optimized out>, str=<value optimized out>, ptr=0x86b6ef0) at malloc.c:6267
#12 0xb6131ab9 in __libc_free (mem=0x86b6ef0) at malloc.c:3700
#13 0xb62ddb1f in operator delete (ptr=0x86b6ef0) at ../../../../libstdc++-v3/libsupc++/del_op.cc:44
#14 0xb62be73b in std::basic_string<char, std::char_traits<char>, std::allocator<char> >::_Rep::_M_destroy(std::allocator<char> const&) () at ../../../../libstdc++-v3/src/parallel_settings.cc:1

Created attachment 55986 [details]
New crash information added by DrKonqi

dolphin (1.6) on KDE Platform 4.5.95 (4.6 RC2) using Qt 4.7.1

- What I was doing when the application crashed:

I Just Opened Dolphin, Inmediatly I Closed it and it Crashed, I did nothing, and is the same if I work with it, only when I press the "close botton" on the top of the window, it Crash, In fact just send the "crash window" but apparently it closed fine

Sorry for my poor english, :P

-- Backtrace (Reduced):
#18 0x0804e529 in _start ()

Created attachment 56013 [details]
New crash information added by DrKonqi

dolphin (1.6) on KDE Platform 4.5.95 (4.6 RC2) using Qt 4.7.1

- What I was doing when the application crashed:

Just closing Dolphin, after reviwing a subdirectoriy.

-- Backtrace (Reduced):
#13 0x0000000000407610 in _start ()

Created attachment 56023 [details]
New crash information added by DrKonqi

dolphin (1.6) on KDE Platform 4.5.95 (4.6 RC2) using Qt 4.7.1

- What I was doing when the application crashed:

I was closing Dolphin and the application crashed (SEGFAULT signal).

-- Backtrace (Reduced):
#18 0x0804e529 in _start ()

*** Bug 263089 has been marked as a duplicate of this bug. ***

*** Bug 263371 has been marked as a duplicate of this bug. ***

*** Bug 263432 has been marked as a duplicate of this bug. ***

(In reply to comment #98)
> Created an attachment (id=56023) [details]
> New crash information added by DrKonqi
> 
> dolphin (1.6) on KDE Platform 4.5.95 (4.6 RC2) using Qt 4.7.1
> 
> - What I was doing when the application crashed:
> 
> I was closing Dolphin and the application crashed (SEGFAULT signal).

My mistake: there's not a SEGFAULT signal. Is an ABORT signal.

> 
> -- Backtrace (Reduced):
> #18 0x0804e529 in _start ()

*** Bug 263492 has been marked as a duplicate of this bug. ***

@Bruno check that the debuginfos for libstrigi0 and strigi exactly match the base packages (version and build numbers).

I was able to reproduce this once with Dolphin so far.
Steps to reproduce: 
* enable metadata sidebar (F11) 
* hover a text file
* quit Dolphin

This gave an identical backtrace to comment #87, and I do have the correct debuginfo for libstrigi0 and strigi itself installed.  Multiple valgrind runs of dolphin show the same error as c#87 but no crash.

The interesting thing is that Nepomuk is completely disabled on this user, so why is libstreams involved at all? KMetaInfo?

Same double-delete in libstreams.so.0.7.3 when running strigicmd --help in valgrind.

I have just submitted a newer strigi snapshot (as 0.7.3.99) to openSUSE KDE:Distro:Factory that apparently does not crash or exhibit the double-delete valgrind showed.  Please test with 0.7.3.99 when it becomes available.

Created attachment 56182 [details]
Valgrind run of strigicmd after update 

List of strigi installed packages 
S | Name                       | Type    | Version    | Arch   | Repository      
--+----------------------------+---------+------------+--------+-----------------
i | libstrigi0                 | package | 0.7.3-50.5 | x86_64 | factory-kde-core
i | libstrigi0-32bit           | package | 0.7.3-50.5 | x86_64 | factory-kde-core
i | libstrigi0-debuginfo       | package | 0.7.3-50.5 | x86_64 | factory-kde-core
i | libstrigi0-debuginfo-32bit | package | 0.7.3-50.5 | x86_64 | factory-kde-core
i | strigi                     | package | 0.7.3-50.5 | x86_64 | factory-kde-core
i | strigi-debuginfo           | package | 0.7.3-50.5 | x86_64 | factory-kde-core
i | strigi-debugsource         | package | 0.7.3-50.5 | x86_64 | factory-kde-core
i | strigi-devel               | package | 0.7.3-50.5 | x86_64 | factory-kde-core

log produced by 
valgrind -v --leak-check=full --show-reachable=yes --trace-children=yes --track-origins=yes --log-file=/home/bruno/tmp/KDE_KCRASH/kde_test_bug257944.log strigicmd --help

Created attachment 56183 [details]
@Will : Dolphin run with preview, +F11 doesn't crash at close time 

Full screen-shot. Even with strigi runnning indexing files, dolphin doesn't crash at close time.

@Comment #91: libstreamanalyzer/lib/streamanalyzer.cpp: line 431 from
> git clone git://anongit.kde.org/libstreamanalyzer:

cerr << "could not reset stream of " << idx.path().c_str()
                        << " from pos " << input->position()
                        << " to 0 after reading with " << sea->name()
                        << ": " << sea->error().c_str() << endl;

Lot of error messages:
gwenview(8822)/kdecore (kdelibs): findExe(): returning  "/usr/bin/update-mime-database"
could not skip chunk size 99141
could not reset stream of /home/hrd/Bilderrahmen/screenshot_4.png from pos 65536 to 0 after reading with PngEndAnalyzer:
could not skip chunk size 137483

> dir /home/hrd/Bilderrahmen/screenshot_4.png
lrwxrwxrwx 1 hrd users 59 17. Aug 17:12 /home/user/Bilderrahmen/screenshot_4.png -> /usr/share/wallpapers/City_at_Night/contents/screenshot.png

From all wallpapers links '3 png links' out of 37 and 'all 3 jpg links' don't produce the "could not reset stream of " error message on konsole.

> gwenview
...
KCrash: Application 'gwenview' crashing...

> ps jax | grep gwenview (this socket is stopped)
... S     1000   0:00 kdeinit4: kio_file [kdeinit] file local:/tmp/ksocket-hrd/klauncherRT2526.slave-socket local:/tmp/ksocket-hrd/gwenviewEt9663.slave-socket

This is openSUSE 11.2 with strigi-0.7.3-50.3.
Kdiff3 shows lots of diffs from strigi-0.7.3-50.3.src.rpm to the 5 'git clone git://anongit.kde.org/...'.

Created attachment 56193 [details]
> gwenview |& gvim -

No gwenview crash anymore with strigi-0.7.3.99-51.1 !

"could not reset stream" message still present.

Created attachment 56234 [details]
Valgrind run of strigicmd after update ( strigi 0.7.3.99 )

seem fixed
just added to monitor

Thanks for the update. I am still interested to see the valgrind log of the crash _with_ debugging sources installed. Comment #108 just says:

==11372==    by 0x5F98775: ??? (in /usr/lib64/libstreams.so.0.7.3)

so it also lacks the important part.

@Christoph I was also unable to get valgrind to resolve that frame, even with the debuginfo installed.  

@Bruno perhaps *try* installing the -debugsource also, although this is just superstition - the debuginfo should be enough for valgrind to get source file line numbers.

Perhaps some optimisation present prevents valgrind resolving things.

I've just check : 
rpm -qf /usr/lib64/libstreams.so.0.7.3 is in strigi and I already do the 2 last valgrind with the -debugsource too :-) 

S | Name                       | Type    | Version       | Arch   | Repository      
--+----------------------------+---------+---------------+--------+-----------------
i | libstrigi0                 | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | libstrigi0-32bit           | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | libstrigi0-debuginfo       | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | libstrigi0-debuginfo-32bit | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | strigi                     | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | strigi-debuginfo           | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | strigi-debugsource         | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | strigi-devel               | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core

I've just check : 
rpm -qf /usr/lib64/libstreams.so.0.7.3 is in strigi and I already do the 2 last valgrind with the -debugsource too :-) 

S | Name                       | Type    | Version       | Arch   | Repository      
--+----------------------------+---------+---------------+--------+-----------------
i | libstrigi0                 | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | libstrigi0-32bit           | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | libstrigi0-debuginfo       | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | libstrigi0-debuginfo-32bit | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | strigi                     | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | strigi-debuginfo           | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | strigi-debugsource         | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core
i | strigi-devel               | package | 0.7.3.99-51.1 | x86_64 | factory-kde-core

If some valgrind guru have another option I can run, just ping it :-)

@Comment #113:
==11372==    by 0x5F98775: ??? (in /usr/lib64/libstreams.so.0.7.3)

That was before the big strigi commit to 0.7.3.99-51.1 which handles new's and corresponding delete's of objects with a factory method. The buggy packages (0.7.3-50.5) are not on 'http://download.opensuse.org/repositories/KDE:/Distro:/Factory/openSUSE_11.2/' anymore.

Well, I wasn't really clear. While it is nice to know that upgrading fixed it, I still don't know what the actual bug/problem was. I am interested in _what_ caused the bug, and how it got fixed, so that we can avoid those problems in the future.

Additionally, I don't just want to resolve the bug as DOWNSTREAM, when the problem was in our repository.

So if anyone still can reproduce the crash with valgrind, and is able to provide the log with the relevant debugging information enabled, please share :)

Created attachment 56270 [details]
Output from valgrind

Output from:
valgrind -v --leak-check=full --show-reachable=yes --trace-children=yes --track-origins=yes --log-file=krash.log strigicmd --help

These packages are installed:

rpm -qa --last | grep strigi
strigi-debuginfo-0.7.3-6.1                    Fri 21 Jan 2011 10:31:18 AM EET
strigi-debugsource-0.7.3-6.1                  Fri 21 Jan 2011 10:31:14 AM EET
libstrigi0-debuginfo-0.7.3-6.1                Wed 12 Jan 2011 12:30:21 PM EET
strigi-0.7.3-6.1                              Wed 12 Jan 2011 12:22:59 PM EET
libstrigi0-0.7.3-6.1                          Wed 12 Jan 2011 12:21:54 PM EET

Operating System is openSUSE 11.4Milestone5+current Factory updates

Those were all the debug packages that I could install and there's still no useful libstreams information.  Anything else I could do to help?

zypper search strigi
Loading repository data...
Reading installed packages...

S | Name                       | Summary                                    | Type   
--+----------------------------+--------------------------------------------+--------
i | libstrigi0                 | Lightweight and fast desktop search engine | package
  | libstrigi0-32bit           | Lightweight and fast desktop search engine | package
i | libstrigi0-debuginfo       | Debug information for package libstrigi0   | package
  | libstrigi0-debuginfo-32bit | Debug information for package libstrigi0   | package
i | strigi                     | Lightweight and fast desktop search engine | package
i | strigi-debuginfo           | Debug information for package strigi       | package
i | strigi-debugsource         | Debug sources for package strigi           | package
  | strigi-devel               | Lightweight and fast desktop search engine | package

*** Bug 264006 has been marked as a duplicate of this bug. ***

Created attachment 56367 [details]
New crash information added by DrKonqi

dolphin (1.5.9) on KDE Platform 4.5.80 (4.6 Beta1) using Qt 4.7.1

- What I was doing when the application crashed:

Everytime i close Dolphin the program reports a crash

-- Backtrace (Reduced):
#13 0x0000000000407610 in _start ()

Created attachment 56536 [details]
New crash information added by DrKonqi

dolphin (1.6) on KDE Platform 4.5.95 (4.6 RC2) using Qt 4.7.1

- What I was doing when the application crashed:

Every time I close the Dolphin, it crashes, no matter what I am doing.

-- Backtrace (Reduced):
#13 0x0000000000407610 in _start ()

Please do not add any backtraces unless you updated your KDE and strigi to 4.6.0 and 0.7.3.99. Those packages are available from the openSUSE KDE repos.

What's the status, it didn't happen lately for me on opensuse, as Sven already said, anybody experience crashes on close with Strigi 0.7.3.99 and KDE 4.6?

Otherwise we can close this.

Seems fine for me as well.

Seems fine for me, too.

--- Comment #125 from Karsten König <remur gmx net> 2011-02-04 19:48:25 ---
> What's the status, it didn't happen lately for me on opensuse, as Sven already
> said, anybody experience crashes on close with Strigi 0.7.3.99 and KDE 4.6?
>
> Otherwise we can close this.
>
Ok for me also

strigiclient still crashes on exit with the same valgrind log using KDE 4.6.0 and strigi 0.7.3.99 on openSuse 11.3:

strigi-debuginfo-0.7.3.99-52.1.i586
libstrigi0-debuginfo-0.7.3.99-52.1.i586
strigi-devel-0.7.3.99-52.1.i586
strigi-debugsource-0.7.3.99-52.1.i586
libstrigi0-0.7.3.99-52.1.i586
strigi-0.7.3.99-52.1.i586

==7741== Invalid read of size 4
==7741==    at 0x4FF10F1: std::basic_string<char, std::char_traits<char>, std::allocator<char> >::~basic_string() (atomicity.h:46)
==7741==    by 0x507CB4A: __cxa_finalize (in /lib/libc-2.11.2.so)
==7741==    by 0x527AAB3: ??? (in /usr/lib/libstreamanalyzer.so.0.7.3)
==7741==    by 0x52DD46F: ??? (in /usr/lib/libstreamanalyzer.so.0.7.3)
==7741==    by 0x400ED1C: _dl_fini (in /lib/ld-2.11.2.so)
==7741==    by 0x507C7CE: __run_exit_handlers (in /lib/libc-2.11.2.so)
==7741==    by 0x507C82C: exit (in /lib/libc-2.11.2.so)
==7741==    by 0x5064C15: (below main) (in /lib/libc-2.11.2.so)
==7741==  Address 0x5878238 is 8 bytes inside a block of size 49 free'd
==7741==    at 0x40266AD: operator delete(void*) (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so)
==7741==    by 0x4FF105A: std::string::_Rep::_M_destroy(std::allocator<char> const&) (new_allocator.h:95)
==7741==    by 0x4FF1109: std::basic_string<char, std::char_traits<char>, std::allocator<char> >::~basic_string() (basic_string.h:237)
==7741==    by 0x507C7CE: __run_exit_handlers (in /lib/libc-2.11.2.so)
==7741==    by 0x507C82C: exit (in /lib/libc-2.11.2.so)
==7741==    by 0x5064C15: (below main) (in /lib/libc-2.11.2.so)

Martin are you sure you have libstrigi0-debuginfo-0.7.3.99-52.1.i586 installed? Because it shows ??? for libstreamanalyzer function calls

The valgrind output isn't of much use =/

Created attachment 57016 [details]
output from valgrind -v --leak-check=full --show-reachable=yes --trace-children=yes --track-origins=yes --log-file=krash.log strigicmd --help

Seems all right for me in openSUSE 11.4 Factory (RC1)

(In reply to comment #130)
> Martin are you sure you have libstrigi0-debuginfo-0.7.3.99-52.1.i586 installed?
> Because it shows ??? for libstreamanalyzer function calls

Yes, I'm sure. What I posted is the output of rpm -qa|grep strigi
Also, rpm -qa|grep strigi | xargs rpm -q --verify
does not give any error.

I've now updated to libstrigi0-debuginfo-0.7.3.99-57.3 and libstrigi0-0.7.3.99-57.3 and I still get valgrind errors
"Invalid free() / delete / delete[]" on exit of strigiclient
and it still shows
 ??? (in /usr/lib/libstreamanalyzer.so.0.7.3)

wait, strigiclient? This is something different, this bugreport is about dolphin and other kde workspace stuff (I know strigiclient is not unrelated), but I am pretty sure most people meant a crash on closing dolphin.

So please make a new bugreport so this doesn't get mixed with maybe misleading information in here.
btw don't get your hopes up about a quick fix for strigi stuff, the dev is currently swamped with work, also strigi has it's own bugtracker, rather post there to make sure he will see it
http://sourceforge.net/tracker/?group_id=171000&atid=856302

Please look at comment #80, #81 and #82. I got the impression the problem has the same source. Probably I'm wrong.

I can open and close strigiclient without a crash but dolphin used to crash for me before the 0.7.3.99 strigi from opensuse as well.
No idea what would be the best way forward

> --- Comment #135 from Karsten König <remur gmx net>  2011-02-09 19:42:27 ---
> I can open and close strigiclient without a crash but dolphin used to crash for
> me before the 0.7.3.99 strigi from opensuse as well.

did you check with valgrind ?
Without it, you might not see the problem.

Oh right, valgrind catches a broken free, but that doesn't happen in with dolphin so I still think it's a seperate issue, btw I also get no method names like you, bah.

Created attachment 57327 [details]
New crash information added by DrKonqi

dolphin (1.6) on KDE Platform 4.5.95 (4.6 RC2) using Qt 4.7.1

- What I was doing when the application crashed:

Closing dolphin, crashes randomly, openSUSE 11.3 KDE 4.6

-- Backtrace (Reduced):
#14 0x0000000000400699 in _start ()

Please update your system, no opensuse-kde maintained repository holds 4.6rc2 anymore.

*** Bug 270891 has been marked as a duplicate of this bug. ***

*** Bug 274724 has been marked as a duplicate of this bug. ***

Since we have not been able to find the cause for these crashes, and they were only happening on openSUSE factory packages, I mark this bug as a downstream bug.