This bug is a split from https://bugs.mageia.org/show_bug.cgi?id=32231 which became unreadable For some users, guayadeque crashes immediately for some reason to investigate NB I's not the same problem as describe in https://bugs.mageia.org/show_bug.cgi?id=32532 for which a solution had been found
The following comments from bug 32231 explain what happens with the current version from core and with the update proposed to test in bug 32532 (which solved an other kind of crash for other users) https://bugs.mageia.org/show_bug.cgi?id=32231#c54 https://bugs.mageia.org/attachment.cgi?id=14160 https://bugs.mageia.org/show_bug.cgi?id=32231#c72 The reason why the Guayadeque rpms crashes on some 64bits systems and not on other has not yet an explanation
NB These users can't even use the solution of disabling fader and equalizer to prevent the crash as for https://bugs.mageia.org/show_bug.cgi?id=32532 as suggested by Jani (https://bugs.mageia.org/show_bug.cgi?id=32231#c21) Guaydeque crashes immediately for them (they can't get the interface of the program where to modify the playback preferences)
CC: (none) => andrewsfarm, geiger.david68210, j.alberto.vc, jani.valimaa, marja11, ngompa13
This useful part of the bug report created by Guayadeque itself https://bugs.mageia.org/attachment.cgi?id=14160 indicates that something is wrong for these users with the use of libwx_gtk3u_wxsqlite3-3.2.so.0 <stack> <frame level="0" function="Guayadeque::guMainApp::OnFatalException()" offset="0" address="0x56ee5a"/> <frame level="1" offset="0x1d8570" address="0x7f36a3fd8570" module="/lib64/libwx_baseu-3.2.so.0"/> <frame level="2" offset="0x36960" address="0x7f36a2664960" module="/lib64/libc.so.6"/> <frame level="3" function="sqlite3_initialize" offset="0x3f7" address="0x7f36a2bf1b17" module="/lib64/libwx_gtk3u_wxsqlite3-3.2.so.0"/> <frame level="4" offset="0x14827b" address="0x7f36a2ca427b" module="/lib64/libwx_gtk3u_wxsqlite3-3.2.so.0"/> <frame level="5" function="wxSQLite3Database::Open(wxString const&, wxMemoryBuffer const&, int)" offset="0x8b" address="0x7f36a2cb458b" module="/lib64/libwx_gtk3u_wxsqlite3-3.2.so.0"/> <frame level="6" function="wxSQLite3Database::Open(wxString const&, wxString const&, int)" offset="0x11c" address="0x7f36a2cb487c" module="/lib64/libwx_gtk3u_wxsqlite3-3.2.so.0"/> <frame level="7" function="Guayadeque::guDb::Open(wxString const&)" offset="0" address="0x5af1df"/> <frame level="8" function="Guayadeque::guDb::guDb(wxString const&)" offset="0" address="0x5af28c"/> <frame level="9" function="Guayadeque::guDbCache::guDbCache(wxString const&)" offset="0" address="0x5af720"/> <frame level="10" function="Guayadeque::guMainApp::OnInit()" offset="0" address="0x56e0ae"/> <frame level="11" function="wxEntry(int&, wchar_t**)" offset="0x6a" address="0x7f36a3f1ec8a" module="/lib64/libwx_baseu-3.2.so.0"/> <frame level="12" function="main" offset="0" address="0x561f28"/> <frame level="13" offset="0x236b7" address="0x7f36a26516b7" module="/lib64/libc.so.6"/> <frame level="14" function="__libc_start_main" offset="0x85" address="0x7f36a2651775" module="/lib64/libc.so.6"/> <frame level="15" function="_start" offset="0" address="0x567471"/> </stack>
Summary: Guayadeque crashes immediately when launched on some 64 bits systems => Guayadeque crashes immediately for some userswhen launched on some 64 bits systems
Summary: Guayadeque crashes immediately for some userswhen launched on some 64 bits systems => Guayadeque crashes immediately for some users when launched on their 64 bits systems
Please test with wxsqlite3-4.9.1-1.1.mga9 in core/updates_testing. SRPMS: wxsqlite3-4.9.1-1.1.mga9 RPMS: lib(64)wx_gtk3u_wxsqlite3_3.2_0-4.9.1-1.1.mga9 lib(64)wx_gtk3u_wxsqlite3_3.2-devel-4.9.1-1.1.mga9
Thanks Jani @ katnatek @ Thomas Andrews May you test it to verify if it corrects it for the current version and for the updated version ? (I will test it... but I'm not concerned by this bug ;)
I installed your update NB I use Guayadeque from a rpm built with mock and the corrected spec file https://bugs.mageia.org/attachment.cgi?id=14167 I renamed my /home/myusername/.guayadeque to /home/myusername/.guayadequebak so that guayadeque builds a new database As I thought no problem for me (I had not before) I then deleted this newly created /home/myusername/.guayadeque and renamed /home/myusername/.guayadequebak to /home/myusername/.guayadeque Guayadeque starts immediately with the old database and works perfectly But I don't know if this will change anything for the others PS maybe to really test we need to rebuild guayadeque with this new lib(64)wx_gtk3u_wxsqlite3_3.2-devel-4.9.1-1.1.mga9
Hi again Jani I have rebuilt guayadeque inside mock after having installed inside mock your last lib64wx_gtk3u_wxsqlite3_3.2_0-4.9.1-1.1.mga9 lib64wx_gtk3u_wxsqlite3_3.2-devel-4.9.1-1.1.mga9 and launched mock with --no-clean option So that it is now built upon the last version Then I installed guayadeque and tested it again the same way as in https://bugs.mageia.org/show_bug.cgi?id=32535#c6 No problem for me
One last question When the badly versioned of the guayadeque.rpm will be removed by sysadmins from Cauldron and Mageia9/core/updates_testing I may submit it again with the corrected spec file Will it be built for Mageia9/core/updates_testing with your last versions of lib(64)wx_gtk3u_wxsqlite3_3.2_0-4.9.1-1.1.mga9 lib(64)wx_gtk3u_wxsqlite3_3.2-devel-4.9.1-1.1.mga9 that are present in Mageia9/core/updates_testing ? Or with the version from Mageia9/core ?
Test 1 Install the previous testing guayadeque package with the testing lib(64)wx_gtk3u_wxsqlite3_3.2_0-4.9.1-1.1.mga9 The list maybe don't have all the wx requires because I have other application that also use that libraries LC_ALL=C urpmi guayadeque To satisfy dependencies, the following packages are going to be installed: Package Version Release Arch (medium "Core Release (Installer) (DVD1)") lib64gpod4 0.8.3 24.mga9 x86_64 lib64sgutils2_2 1.46 2.mga9 x86_64 libgpod 0.8.3 24.mga9 x86_64 (medium "QA Testing (64-bit)") guayadeque 0.4.8 4.git2023111> x86_64 lib64wx_gtk3u_wxsqlite3_3.2_0 4.9.1 1.1.mga9 x86_64 11MB of additional disk space will be used. 3.5MB of packages will be retrieved. Proceed with the installation of the 5 packages? (Y/n) y installing //home/katnatek/qa-testing/x86_64/guayadeque-0.4.8-4.git20231115.1.mga9.x86_64.rpm /mnt/MageiaDVD/x86_64/media/core/lib64gpod4-0.8.3-24.mga9.x86_64.rpm //home/katnatek/qa-testing/x86_64/lib64wx_gtk3u_wxsqlite3_3.2_0-4.9.1-1.1.mga9.x86_64.rpm /mnt/MageiaDVD/x86_64/media/core/lib64sgutils2_2-1.46-2.mga9.x86_64.rpm /mnt/MageiaDVD/x86_64/media/core/libgpod-0.8.3-24.mga9.x86_64.rpm Preparing... ########################################################################################### 1/5: lib64sgutils2_2 ########################################################################################### 2/5: libgpod ########################################################################################### 3/5: lib64gpod4 ########################################################################################### 4/5: lib64wx_gtk3u_wxsqlite3_3.2_0 ########################################################################################### 5/5: guayadeque Run guayadeque as users crash I will try rebuild guyadeque against this new wx library but, why don't like to disable as-needed ldflags to make it works ?
Hi before rebuilding with disabling as-needed ldflags (that I never had to use) I am still thinking there's an hidden dependency installed on my computer that is not on yours Would you please run this in a console : rpm -qa |sort > rpm.list and attach this file so that I can compare with mine : I will try to uninstall some rpms that are on my computers and not on yours and then launch guayadeque to find the culprit
(In reply to Philippe Didier from comment #10) > Hi > > before rebuilding with disabling as-needed ldflags (that I never had to use) > > I am still thinking there's an hidden dependency installed on my computer > that is not on yours > Would you please run this in a console : > > rpm -qa |sort > rpm.list > > and attach this file so that I can compare with mine : > I will try to uninstall some rpms that are on my computers and not on yours > and then launch guayadeque to find the culprit This is again starting to be a mess. Lets avoid fuss, and keep this simple. I'll provide a list of extra pkgs that are required if the pkg is built without -Wl,--as-needed.
Sorry for the mess Didn't know that there was a better way to find which required pkg is missing Thanks for your help ...
Test 2 Install guayadeque build from the testing srpm against the testing wxsqlite (I did have to build in my copr) Run as user and crash Log of my build of guyadeque https://download.copr.fedorainfracloud.org/results/katnatek/blogdrake/mageia-9-x86_64/06668017-guayadeque/builder-live.log.gz
(In reply to Jani Välimaa from comment #11) > (In reply to Philippe Didier from comment #10) > > Hi > > > > before rebuilding with disabling as-needed ldflags (that I never had to use) > > > > I am still thinking there's an hidden dependency installed on my computer > > that is not on yours > > Would you please run this in a console : > > > > rpm -qa |sort > rpm.list > > > > and attach this file so that I can compare with mine : > > I will try to uninstall some rpms that are on my computers and not on yours > > and then launch guayadeque to find the culprit > > This is again starting to be a mess. Lets avoid fuss, and keep this simple. > > I'll provide a list of extra pkgs that are required if the pkg is built > without -Wl,--as-needed. I avoid you that and say that 0
Perhaps this will help... guayadeque-0.4.8-4.git20231115.1.mga9.x86_64.rpm is what's currently installed. After deleting the .guayadeque folder, I ran it from the command line and saw this: [tom@localhost ~]$ guayadeque 11:59:18: Created the configuration folder 11:59:18: Created the lyrics folder 11:59:18: Created the collections folder 11:59:18: Created the Radios folder 11:59:18: Created the Jamendo folder 11:59:18: Created the Magnatune folder 11:59:18: Created the Podcasts folder 11:59:18: Created the Layouts folder 11:59:18: Created the default configuration file 11:59:18: Created the default equalizers file 11:59:18: Created the default lyrics sources file 11:59:18 AM: Initialized locale ( en_US ) sh: line 1: addr2line: command not found 11:59:29 AM: A debug report has been generated. It can be found in /tmp/guayadeque_dbgrpt-4209-20231119T115918 And includes the following files: guayadeque.xml: process context description Please send this report to the program maintainer, thank you! Aborted (core dumped) Looks to me like it's not finding the command "addr2line" anywhere. The missing dependency, perhaps?
Removed the updated guayadeque, and installed the one from core_release. Deleted the .guayadeque folder again, and ran the app once more from the command line, seeing the same result as in comment 15. I can run an strace if it would be helpful.
There're unused/extra libraries when building without -Wl,--as-needed. Warning: unused libraries in /usr/bin/guayadeque: libgstvideo-1.0.so.0 libgstaudio-1.0.so.0 libgstbase-1.0.so.0 libsqlite3.so.0 libz.so.1 libimobiledevice-1.0.so.6 libplist-2.0.so.3 Besides those 'urpmq --requires' shows also extra requires: libwx_baseu_net-3.2.so.0 The list of extra requires. I didn't check if they're pulled via other generated requires, though. lib64gstreamer-plugins-base1.0_0 lib64sqlite3_0 lib64zlib1 lib64imobiledevice1.0_6 lib64plist2.0_3 lib64wx_baseu_net3.2_0
A search in MCC for the filename "addr2line" reveals several packages that provide it, and none of them are installed on this system.
(In reply to Jani Välimaa from comment #17) > There're unused/extra libraries when building without -Wl,--as-needed. > > Warning: unused libraries in /usr/bin/guayadeque: > libgstvideo-1.0.so.0 > libgstaudio-1.0.so.0 > libgstbase-1.0.so.0 > libsqlite3.so.0 > libz.so.1 > libimobiledevice-1.0.so.6 > libplist-2.0.so.3 > > Besides those 'urpmq --requires' shows also extra requires: > libwx_baseu_net-3.2.so.0 > > The list of extra requires. I didn't check if they're pulled via other > generated requires, though. > lib64gstreamer-plugins-base1.0_0 > lib64sqlite3_0 > lib64zlib1 > lib64imobiledevice1.0_6 > lib64plist2.0_3 > lib64wx_baseu_net3.2_0 I guess I already have that, because no extra package was installed in my test in both arches
(In reply to Thomas Andrews from comment #18) > A search in MCC for the filename "addr2line" reveals several packages that > provide it, and none of them are installed on this system. are you sure of that? binutils-2.40-11.mga9.x86_64:/usr/bin/addr2line
Hi Thomas Nice catch Don't you have binutils installed ? it provides addr2line I have it installed on my compuer and it's required by lots of other programs !
@ katnatek I saw that you don't use plasma as desktop Plasma needs lib64imobiledevice1.0_6 lib64plist2.0_3 Since Guayadeque has acces to Apple's Ipod these rpms are needed Are they installed on your computer ?
(In reply to Philippe Didier from comment #22) > @ katnatek > I saw that you don't use plasma as desktop > Plasma needs > lib64imobiledevice1.0_6 > lib64plist2.0_3 > > Since Guayadeque has acces to Apple's Ipod these rpms are needed > > Are they installed on your computer ? rpm -q lib64imobiledevice1.0_6 lib64imobiledevice1.0_6-1.3.0-7.mga9 rpm -q lib64plist2.0_3 lib64plist2.0_3-2.2.0-4.mga9 BTW, I use for the moment lxqt but I have plasma X11 and also plasma wayland
(In reply to Jani Välimaa from comment #17) > There're unused/extra libraries when building without -Wl,--as-needed. > > Warning: unused libraries in /usr/bin/guayadeque: > libgstvideo-1.0.so.0 > libgstaudio-1.0.so.0 > libgstbase-1.0.so.0 > libsqlite3.so.0 > libz.so.1 > libimobiledevice-1.0.so.6 > libplist-2.0.so.3 > > Besides those 'urpmq --requires' shows also extra requires: > libwx_baseu_net-3.2.so.0 > > The list of extra requires. I didn't check if they're pulled via other > generated requires, though. > lib64gstreamer-plugins-base1.0_0 > lib64sqlite3_0 > lib64zlib1 > lib64imobiledevice1.0_6 > lib64plist2.0_3 > lib64wx_baseu_net3.2_0 I don't see that in my builds, perhaps because I add also %global optflags %{optflags} -flto=auto
@ Jani lib64wx_baseu_net3.2_0 is installed on my computer it is required by - audacity that is installed - opencpn that I maintain and is installed - lib64wxgtku3.2-devel-3.2.1-3 that is installed - lib64wx_gtk3u_wxsqlite3_3.2-devel-4.9.1-1.1. that is installed We apparently need to add lib64wx_baseu_net3.2_0 as Requires (this never appeared to me since it's installed for these 4 rpms !
@ katnatek have you lib64wx_baseu_net3.2_0 installed on your computer ? If no : install it and test guayadeque again (after having deleted /home/yourusername/.guayadeque directory)
(In reply to katnatek from comment #20) > (In reply to Thomas Andrews from comment #18) > > A search in MCC for the filename "addr2line" reveals several packages that > > provide it, and none of them are installed on this system. > > are you sure of that? > binutils-2.40-11.mga9.x86_64:/usr/bin/addr2line I'm sure. It surprised me too that it wasn't installed, but, this happens to be a very basic install, with nothing much beyond what the original install iso put there.
(In reply to Thomas Andrews from comment #27) > (In reply to katnatek from comment #20) > > (In reply to Thomas Andrews from comment #18) > > > A search in MCC for the filename "addr2line" reveals several packages that > > > provide it, and none of them are installed on this system. > > > > are you sure of that? > > binutils-2.40-11.mga9.x86_64:/usr/bin/addr2line > > I'm sure. It surprised me too that it wasn't installed, but, this happens to > be a very basic install, with nothing much beyond what the original install > iso put there. Well that not is the cause rpm -q binutils binutils-2.40-11.mga9
(In reply to Philippe Didier from comment #26) > @ katnatek > have you lib64wx_baseu_net3.2_0 installed on your computer ? > If no : install it and test guayadeque again > (after having deleted /home/yourusername/.guayadeque directory) rpm -q lib64wx_baseu_net3.2_0 lib64wx_baseu_net3.2_0-3.2.1-3.mga9
I would like to know if LD_DEBUG shows anything interesting. $ LD_DEBUG=symbols LD_DEBUG_OUTPUT=guayadeque.log guayadeque To get info about LD_DEBUG one can use: $ LD_DEBUG=help cat
perhaps I need to compress some of them ls -lah guayadeque.log.22* -rw-r--r-- 1 katnatek katnatek 73M nov 19 12:41 guayadeque.log.22191 -rw-r--r-- 1 katnatek katnatek 33K nov 19 12:41 guayadeque.log.22214 -rw-r--r-- 1 katnatek katnatek 14K nov 19 12:41 guayadeque.log.22215 -rw-r--r-- 1 katnatek katnatek 125K nov 19 12:41 guayadeque.log.22216
(In reply to katnatek from comment #28) > (In reply to Thomas Andrews from comment #27) > > (In reply to katnatek from comment #20) > > > (In reply to Thomas Andrews from comment #18) > > > > A search in MCC for the filename "addr2line" reveals several packages that > > > > provide it, and none of them are installed on this system. > > > > > > are you sure of that? > > > binutils-2.40-11.mga9.x86_64:/usr/bin/addr2line > > > > I'm sure. It surprised me too that it wasn't installed, but, this happens to > > be a very basic install, with nothing much beyond what the original install > > iso put there. > > Well that not is the cause > rpm -q binutils > binutils-2.40-11.mga9 That doesn't surprise me. urpmq --whatrequires-recursive binutils > binutils.txt produces an extremely long list of packages, but guayadeque is not among them. Since it needs that command, it should be, shouldn't it? Or is it supposed to get it from somewhere else?
(In reply to Thomas Andrews from comment #32) > (In reply to katnatek from comment #28) > > (In reply to Thomas Andrews from comment #27) > > > (In reply to katnatek from comment #20) > > > > (In reply to Thomas Andrews from comment #18) > > > > > A search in MCC for the filename "addr2line" reveals several packages that > > > > > provide it, and none of them are installed on this system. > > > > > > > > are you sure of that? > > > > binutils-2.40-11.mga9.x86_64:/usr/bin/addr2line > > > > > > I'm sure. It surprised me too that it wasn't installed, but, this happens to > > > be a very basic install, with nothing much beyond what the original install > > > iso put there. > > > > Well that not is the cause > > rpm -q binutils > > binutils-2.40-11.mga9 > > That doesn't surprise me. urpmq --whatrequires-recursive binutils > > binutils.txt produces an extremely long list of packages, but guayadeque is > not among them. Since it needs that command, it should be, shouldn't it? Or > is it supposed to get it from somewhere else? Indeed binutils is required by lots of devel packages If you don't use to build packages then binutils might be not installed ! Install it and then launch again guayadeque You might not have again : addr2line: command not found and maybe no more problem ?
(In reply to Jani Välimaa from comment #30) > I would like to know if LD_DEBUG shows anything interesting. > > $ LD_DEBUG=symbols LD_DEBUG_OUTPUT=guayadeque.log guayadeque > > To get info about LD_DEBUG one can use: > $ LD_DEBUG=help cat I upload to blogdrake's server please tell me whe you download yo remove https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.22191 https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.22214 https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.22215 https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.22216 This are from the version compiled against testing version of lib64wx_gtk3u_wxsqlite3_3.2_0
(In reply to katnatek from comment #34) > (In reply to Jani Välimaa from comment #30) > > I would like to know if LD_DEBUG shows anything interesting. > > > > $ LD_DEBUG=symbols LD_DEBUG_OUTPUT=guayadeque.log guayadeque > > > > To get info about LD_DEBUG one can use: > > $ LD_DEBUG=help cat > > I upload to blogdrake's server please tell me whe you download yo remove > > https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.22191 > https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.22214 > https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.22215 > https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.22216 > > This are from the version compiled against testing version of > lib64wx_gtk3u_wxsqlite3_3.2_0 Please don't mix personal builds with official builds. We should always use official packages when debugging bugs so we know there're no users' own customizations etc.
(In reply to Jani Välimaa from comment #35) > (In reply to katnatek from comment #34) > > (In reply to Jani Välimaa from comment #30) > > > I would like to know if LD_DEBUG shows anything interesting. > > > > > > $ LD_DEBUG=symbols LD_DEBUG_OUTPUT=guayadeque.log guayadeque > > > > > > To get info about LD_DEBUG one can use: > > > $ LD_DEBUG=help cat > > > > I upload to blogdrake's server please tell me whe you download yo remove > > > > https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.22191 > > https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.22214 > > https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.22215 > > https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.22216 > > > > This are from the version compiled against testing version of > > lib64wx_gtk3u_wxsqlite3_3.2_0 > > Please don't mix personal builds with official builds. We should always use > official packages when debugging bugs so we know there're no users' own > customizations etc. Is directly build from the srpm in updates_testing, but I will unistall and install again the guayadeque in updates_testing
New urls of the test with the current guayadeque in updates_testing https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.31248 https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.31273 https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.31274 https://ftp.blogdrake.net/RPMS/tmp/guayadeque.log.31275 @Jani, please tell me when you download
(In reply to Philippe Didier from comment #33) > > Indeed binutils is required by lots of devel packages > If you don't use to build packages then binutils might be not installed ! > Install it and then launch again guayadeque > You might not have again : > addr2line: command not found > > and maybe no more problem ? Of course it wasn't that easy. I didn't think it would be. Guayadeque still crashes right away, but now it no longer complains about not finding addr2line. Philippe, I believe that the best way for you to examine why guayadeque is crashing on systems without all the devel packages on your machine would be to create a new, basic Mageia 9 guest in VirtualBox, and play with that yourself. I've already contributed too much to this mess of a bug. I'm not going to make it even worse by sending people off on wild goose chases because I don't know what I'm doing. I'll be back if it ever gets ready for QA, but until then I don't believe I can add anything useful.
CC: andrewsfarm => (none)
Thanks Thomas You give the specific explanation of the first crash that you encountered It was not so easy to suspect (I couldn't imagine that binutils is not installed on every system !) I will add Requires : binutils to the spec file We have now to find the second explanation of the crash that is now common to you and katnatek
Last tests I installed Mageia9 in a virtual machine : basic install from the iso (no rpm added) updated it binutils is not installed by default ! 1) I installed guayadeque from core launch it => crash immediately 2) I installed binutils Launch guayadeque add a path to some music OK add a play list OK play => crash 3) launch guayadeque again disable fader and equalizer no more crash when playing ok 4) Update guayadeque from updates_testing launch it enable fader and equalizer no more crash ok I can't reproduce the crash describe by katnatek It's not necessary to update lib64wx_gtk3u_wxsqlite3_3.2_0
Last tests I installed Mageia9 in a virtual machine : basic install from the iso (no rpm added) updated it Effectively, binutils is not installed by default ! (I will add this requires in the spec file of guayadeque!) 1) I installed guayadeque from core (0.4.7-1) launch it => crash immediately like for Thomas 2) I installed binutils Launch guayadeque add a path to some music OK add a play list OK play => crash 3) launch guayadeque again disable fader and equalizer no more crash when playing OK 4) Update guayadeque from updates_testing (0.4.8-4.git20231115.1) launch it enable fader and equalizer no more crash OK I can't reproduce the crash described by katnatek @ Jani PS It's not necessary to update lib64wx_gtk3u_wxsqlite3_3.2_0 from 4.9.1-1 to 4.9.1-1.1 for guayadeque : it works without this update Nevertheless, I updated it and it doesn't hurt !
See Also: (none) => https://bugs.mageia.org/show_bug.cgi?id=32231
See Also: (none) => https://bugs.mageia.org/show_bug.cgi?id=32532
Thank you everyone for all the testing, building etc. Especially Philippe. Comments 40 & 41 report the same sequence of crashing to working. Two things emerge: 1. It looks like guayadeque requires binutils. 2. Given the presenc of binutils, the version in updates_testing looks good. (In reply to Philippe Didier from comment #40) > 4) Update guayadeque from updates_testing > launch it > enable fader and equalizer > no more crash ok (In reply to Philippe Didier from comment #41) > 4) Update guayadeque from updates_testing (0.4.8-4.git20231115.1) > launch it > enable fader and equalizer > no more crash OK (In reply to Thomas Andrews from comment #32) > That doesn't surprise me. urpmq --whatrequires-recursive binutils > > binutils.txt produces an extremely long list of packages, but guayadeque is > not among them. $ urpmq --whatrequires-recursive binutils | grep guayadeque might have been easier. Conversely: $ urpmq --requires-recursive guayadeque | grep binutils returns nothing. Without the grep, the list is huge. Should the version in updates_testing be held in obeyance until the requires on binutils is added? Rather than yet another bug. Or this bug be closed OK along with its mate bug 32532.
CC: (none) => lewyssmith
Hi Lewis I made a comment https://bugs.mageia.org/show_bug.cgi?id=32532#c17 Indeed there will be a new rpm in updates_testing : I made a mistake with the version numbering : 0.4.8-4.git20231115.1 is not right, nevertheless that brought a solution to bug 32532 We were waiting for sysadmins to remove this rpm from Cauldron and Mageia9/core/updates_testing before submitting a better number : that should have been : 0.4.7-3.git20231115.1 But in the mean time I took the opportunity to modify again the spec and add Requires : binutils to correct partly this bug 32535... and the version will be now 0.4.7-3.git20231115.2 It's not yet submitted to Cauldron and Mageia9/core/updates_testing ! nor tested When bug 32532 will be closed as resolved for its own cause (as I think it may with 0.4.7-3.git20231115.2) this won't resolve totally this bug 32535 nor the bug 32536 that will remain open Nevertheless I will add a comment here to inform that a new rpm is provided when it will have been submitted
(In reply to Philippe Didier from comment #43) > Hi Lewis > I made a comment https://bugs.mageia.org/show_bug.cgi?id=32532#c17 > > Indeed there will be a new rpm in updates_testing : > I made a mistake with the version numbering : 0.4.8-4.git20231115.1 is not > right, > nevertheless that brought a solution to bug 32532 > > > We were waiting for sysadmins to remove this rpm from > Cauldron and > Mageia9/core/updates_testing > before submitting a better number : that should have been : > 0.4.7-3.git20231115.1 > > But in the mean time I took the opportunity to modify again the spec and add > Requires : binutils to correct partly this bug 32535... and the version will > be now 0.4.7-3.git20231115.2 > No, it's still 0.4.7-3.git20231115.1. Nothing was pushed to build system yet, so all changes until the submit should done without bumping the release. Please revert the release bump in SVN, and don't forget to use SILENT in commit msg so it won't end up to %changelog.
That is all for me I ask to Jose Manual to test 0.4.8-4.git20231115.1 and work for him in VBOX and in real hardware. I don't found another solution for my x86_64 and i586 systems, as long as you and Jani don't like the idea of disable as needed then close at less bug#32536 as wontfix @Jani, can I remove the guyadeque logs from the blogdrake server?
Removing '-Wl,--as-needed' linker flag is OK, if it's the last resort. However it's still a workaround until we figure out what's really causing the crash. Logs can be removed.
> > > > But in the mean time I took the opportunity to modify again the spec and add > > Requires : binutils to correct partly this bug 32535... and the version will > > be now 0.4.7-3.git20231115.2 > > > > No, it's still 0.4.7-3.git20231115.1. Nothing was pushed to build system > yet, so all changes until the submit should done without bumping the release. > > Please revert the release bump in SVN, and don't forget to use SILENT in > commit msg so it won't end up to %changelog. Ok Sorry ! since I had added the Requires : binutils I thought it was necessary to bump the release I'm gonna revert it
(In reply to katnatek from comment #45) > That is all for me I ask to Jose Manual to test 0.4.8-4.git20231115.1 and > work for him in VBOX and in real hardware. > > I don't found another solution for my x86_64 and i586 systems, as long as > you and Jani don't like the idea of disable as needed then close at less > bug#32536 as wontfix > The worst bug to diagnose is the one that doesn't appear on all computers We already found a first explanation thanks to Thomas : we add Requires : binutils The second explanation to find is : What explains the crash for you and for Thomas since he has installed binutils ? This crash doesn't appear for me even in a VM where I installed a basic Mageia9 (see https://bugs.mageia.org/show_bug.cgi?id=32535#c41) this crash doesn't appear for Jose Manual... I am still thinking about an hidden dependency for you and Thomas about bug 32536 (similar to the old bug 29742) maybe it's possible to create a conditional spec file !
> > No, it's still 0.4.7-3.git20231115.1. Nothing was pushed to build system > yet, so all changes until the submit should done without bumping the release. > > Please revert the release bump in SVN, and don't forget to use SILENT in > commit msg so it won't end up to %changelog. Hi Jani I was on the way to revert the release bump but I saw that you kept it when adding - define _GUREVISION_ to have proper release string in app title bar is it necessary to revert the release bump now ?
@ Jani You apparently applied this change only for Cauldron ? Do I need only to revert the release bump to Mageia9 ?
(In reply to Philippe Didier from comment #48) > The worst bug to diagnose is the one that doesn't appear on all computers > We already found a first explanation thanks to Thomas : > we add Requires : binutils > > The second explanation to find is : > What explains the crash for you and for Thomas since he has installed > binutils ? > > This crash doesn't appear for me even in a VM where I installed a basic > Mageia9 > (see https://bugs.mageia.org/show_bug.cgi?id=32535#c41) > this crash doesn't appear for Jose Manual.. Sorry the typo is "Jose Manuel" > > I am still thinking about an hidden dependency for you and Thomas > > about bug 32536 (similar to the old bug 29742) maybe it's possible to create > a conditional spec file ! I think this and 32536 is the same bug, but who knows 2 bugs with the same "solution" not sound weird to me anymore XD
@Jani & Philippe , with the amount of warnings building this without ninja https://download.copr.fedorainfracloud.org/results/katnatek/blogdrake/mageia-9-x86_64/06671952-guayadeque/builder-live.log.gz (use the search function of your browser to go to 30% and start to scroll down) I think could a bad use of a function or a missing include in somewhere, why crash for me and Thomas is beyond of my current abilities
(In reply to katnatek from comment #51) > (In reply to Philippe Didier from comment #48) > > The worst bug to diagnose is the one that doesn't appear on all computers > > We already found a first explanation thanks to Thomas : > > we add Requires : binutils > > > > The second explanation to find is : > > What explains the crash for you and for Thomas since he has installed > > binutils ? > > > > This crash doesn't appear for me even in a VM where I installed a basic > > Mageia9 > > (see https://bugs.mageia.org/show_bug.cgi?id=32535#c41) > > this crash doesn't appear for Jose Manual.. > Sorry the typo is "Jose Manuel" > > > > I am still thinking about an hidden dependency for you and Thomas > > > > about bug 32536 (similar to the old bug 29742) maybe it's possible to create > > a conditional spec file ! > > I think this and 32536 is the same bug, but who knows 2 bugs with the same > "solution" not sound weird to me anymore XD Not all bug 29742 appeared only on 32bits in year 2021 and is similar to bug 32536 : it was induced by the Mageia's build_flags building directly from source didn't crash Building directly with a modified buildd where the Mageia's build_flags had been added https://bugs.mageia.org/attachment.cgi?id=13054 induced a crash https://bugs.mageia.org/show_bug.cgi?id=29742#c33 There's something that doesn't work anymore for i586 systems with Mageia's rpm The solution is perhaps a spec file with something conditional like this ifarch 1586 disable what you need endif bug 32535 appears for you and Thomas with the an other reason than the fader problem that the rpm from updates_testing corrected, and for an other reason than the problem with 32 bits systems That seems specific to both of you The workaround doesn't allow to find where to dig to find what is the culprit I persist to think that a rpm is missing for you
(In reply to katnatek from comment #52) > @Jani & Philippe , with the amount of warnings building this without ninja > https://download.copr.fedorainfracloud.org/results/katnatek/blogdrake/mageia- > 9-x86_64/06671952-guayadeque/builder-live.log.gz (use the search function of > your browser to go to 30% and start to scroll down) The next interesting points come until 44%, 51%, 64%, 78%, 79%, 85%, 86%, 87%
(In reply to Philippe Didier from comment #53) > (In reply to katnatek from comment #51) > > > I think this and 32536 is the same bug, but who knows 2 bugs with the same > > "solution" not sound weird to me anymore XD > > Not all > bug 29742 appeared only on 32bits in year 2021 and is similar to bug 32536 : > it was induced by the Mageia's build_flags > building directly from source didn't crash > Building directly with a modified buildd where the Mageia's build_flags had > been added https://bugs.mageia.org/attachment.cgi?id=13054 > induced a crash > https://bugs.mageia.org/show_bug.cgi?id=29742#c33 > > There's something that doesn't work anymore for i586 systems with Mageia's > rpm > > The solution is perhaps a spec file with something conditional like this > > ifarch 1586 > disable what you need > endif > > bug 32535 appears for you and Thomas with the an other reason than the fader > problem > that the rpm from updates_testing corrected, and for an other reason than > the problem with 32 bits systems > > That seems specific to both of you > The workaround doesn't allow to find where to dig to find what is the > culprit > I persist to think that a rpm is missing for you The reason that make me thing is the same is the age of my 2 systems, I find information that the x86_64 (Dell Optiplex 755) and the i586 (Compaq Presario C700 Notebook, Model 30D9) are too close in age (mid 2007, and in a moment of 2008) The main reason that make me doubt about missing require(s) is that nothing more were installed in my test disabling as needed so the missing require(s) must be still missing and thee application must fail also
(In reply to katnatek from comment #52) > @Jani & Philippe , with the amount of warnings building this without ninja > https://download.copr.fedorainfracloud.org/results/katnatek/blogdrake/mageia- > 9-x86_64/06671952-guayadeque/builder-live.log.gz (use the search function of > your browser to go to 30% and start to scroll down) I think could a bad use > of a function or a missing include in somewhere, why crash for me and Thomas > is beyond of my current abilities this happens with or without ninja All these warnings about deprecated declarations don't prevent to build and to work They are known by the guayadeque developer and don't need absolutely to be corrected I already created patches for some of them and applied them but it doesn't bring anything better in the end. No emergency to modify the source code
(In reply to katnatek from comment #55) > (In reply to Philippe Didier from comment #53) > The reason that make me thing is the same is the age of my 2 systems, I find > information that the x86_64 (Dell Optiplex 755) and the i586 (Compaq > Presario C700 Notebook, Model 30D9) are too close in age (mid 2007, and in a > moment of 2008) > > The main reason that make me doubt about missing require(s) is that nothing > more were installed in my test disabling as needed so the missing require(s) > must be still missing and thee application must fail also You got the explanation ! You're certainly right it's a problem with the age of your systems same age (2009) than the system for which I created bug 29742 (compaq mini 311) That leads you now to bug 32536 : it's not a dependency problem !!! and it's a flag problem ! If Thomas has an old system too we can close this
CC: (none) => andrewsfarm
@ Thomas Is your computer an old x86_64 ? Some of them were not strictly compatible with 64bits programs ! If yes it's the bug 32536 for you too !!! and we may close this one...
(In reply to Philippe Didier from comment #58) > @ Thomas > Is your computer an old x86_64 ? > Some of them were not strictly compatible with 64bits programs ! > > If yes it's the bug 32536 for you too !!! and we may close this one... inxi -F provide some information also about the age (at less of the bios)
@Philippe Remember I recommend this 2 lines, the second is enough but the first compensate any loss in the optimizations %global optflags %{optflags} -flto=auto %global _disable_ld_as_needed 1
(In reply to Philippe Didier from comment #50) > @ Jani > You apparently applied this change only for Cauldron ? > Do I need only to revert the release bump to Mageia9 ? Release bump needs to be reverted in both, mga9 and cauldron.
(In reply to Jani Välimaa from comment #46) > Removing '-Wl,--as-needed' linker flag is OK, if it's the last resort. > However it's still a workaround until we figure out what's really causing > the crash. > Someone having a crash should install debug packages and run the app through gbd to get a backtrace. It might give some clues.
(In reply to Jani Välimaa from comment #61) > (In reply to Philippe Didier from comment #50) > > @ Jani > > You apparently applied this change only for Cauldron ? > > Do I need only to revert the release bump to Mageia9 ? > > Release bump needs to be reverted in both, mga9 and cauldron. Done for both of them now
(In reply to Philippe Didier from comment #57) > (In reply to katnatek from comment #55) > > (In reply to Philippe Didier from comment #53) > > > The reason that make me thing is the same is the age of my 2 systems, I find > > information that the x86_64 (Dell Optiplex 755) and the i586 (Compaq > > Presario C700 Notebook, Model 30D9) are too close in age (mid 2007, and in a > > moment of 2008) > > > > The main reason that make me doubt about missing require(s) is that nothing > > more were installed in my test disabling as needed so the missing require(s) > > must be still missing and thee application must fail also > > You got the explanation ! > You're certainly right it's a problem with the age of your systems > same age (2009) than the system for which I created bug 29742 (compaq mini > 311) > > That leads you now to bug 32536 : it's not a dependency problem !!! > and it's a flag problem ! > > If Thomas has an old system too we can close this I looked at the specifications of the Dell Optiplex 755 : the cache memory of some of the processors used for this computtr was very little Perhaps an explanation for the stack overflow when using sqlite to build the database...
(In reply to Philippe Didier from comment #58) > @ Thomas > Is your computer an old x86_64 ? > Some of them were not strictly compatible with 64bits programs ! > > If yes it's the bug 32536 for you too !!! and we may close this one... Yes, the system in question is a Dell Dimension e520. It began life as a Dell Dimension e310, but just to see if it could be done I bought a motherboard and needed daughterboard on eBay that converted (with some case modifications) to the e520. That model originally shipped with a 64-bit P4, but the motherboard I bought had had BIOS upgrades to use a Core2Quad Q6600. This is the first x86_64 application from Mageia that it has been unable to use, possibly because of its age.
(In reply to Thomas Andrews from comment #65) > (In reply to Philippe Didier from comment #58) > > @ Thomas > > Is your computer an old x86_64 ? > > Some of them were not strictly compatible with 64bits programs ! > > > > If yes it's the bug 32536 for you too !!! and we may close this one... > > Yes, the system in question is a Dell Dimension e520. It began life as a > Dell Dimension e310, but just to see if it could be done I bought a > motherboard and needed daughterboard on eBay that converted (with some case > modifications) to the e520. > > That model originally shipped with a 64-bit P4, but the motherboard I bought > had had BIOS upgrades to use a Core2Quad Q6600. > > This is the first x86_64 application from Mageia that it has been unable to > use, possibly because of its age. OK So, this is not a missing dependency... but maybe something related with the stack management with these old processors I don't know if we must use what katnatek proposes : %global optflags %{optflags} -flto=auto %global _disable_ld_as_needed 1 only to be compatible for these old 64bits processors (nevertheless it may be useful to add this as a conditional pat in the spec files for i586 processors like bug 32536)
I hope we are at least circumscribing the reason of this particular bug that doesn't appear on other computers than Thomas'one and katnatek's one
A solution appeared for bug 32536 and may be good here too May you test it with this rpm for Mageia9 64bits https://download.copr.fedorainfracloud.org/results/katnatek/blogdrake/mageia-9-x86_64/06675361-guayadeque/guayadeque-0.4.7-3.git20231115.1.1bdk_mga9.x86_64.rpm https://bugs.mageia.org/show_bug.cgi?id=32536#c21 simply adding to the last spec file from Mageia svn %global _disable_ld_as_needed 1 seems enough We might resolve 4 bugs in one time !
LC_ALL=C urpmi https://download.copr.fedorainfracloud.org/results/katnatek/blogdrake/mageia-9-x86_64/06675361-guayadeque/guayadeque-0.4.7-3.git20231115.1.1bdk_mga9.x86_64.rpm warning: /var/cache/urpmi/partial/guayadeque-0.4.7-3.git20231115.1.1bdk_mga9.x86_64.rpm: Header V4 RSA/SHA256 Signature, key ID 20b7fd3d: NOKEY The following package has bad signature: /var/cache/urpmi/partial/guayadeque-0.4.7-3.git20231115.1.1bdk_mga9.x86_64.rpm: Invalid signature (NOT OK (no key): /var/cache/urpmi/partial/guayadeque-0.4.7-3.git20231115.1.1bdk_mga9.x86_64.rpm: Header V4 RSA/SHA256 Signature, key ID 20b7fd3d: NOKEY) Do you want to continue installation ? (y/N) y installing guayadeque-0.4.7-3.git20231115.1.1bdk_mga9.x86_64.rpm from /var/cache/urpmi/partial Preparing... ################################################################# 1/1: guayadeque ################################################################# Test on my Mageia 9 x86_64, Not surprise, can run without crash Uninstalling waiting the official package
See Also: (none) => https://bugs.mageia.org/show_bug.cgi?id=32550
resolved by bug 32550
Resolution: (none) => FIXEDStatus: NEW => RESOLVED