Hi all! On Mageia Linux Cauldron on x86-64, fotoxx crashes upon starting it from the terminal in a new UNIX user account on LXDE. It generates a segfault such as this one: <<< zappcrash: segment fault fotoxx(_Z9zappcrashPKcz+0xe3) [0x4a4dd3 (null) fotoxx(_Z11get_secondsv+0) [0x4a55a0 (null) /lib64/libpthread.so.0(+0xf000) [0x7fd5400a5000 (null) /lib64/libc.so.6(fflush+0x26) [0x7fd53f53ee96 (null) fotoxx(_Z15exiftool_serverPPc+0x23d) [0x45924d (null) fotoxx(_Z9initzfuncPv+0x483) [0x42dd93 (null) /lib64/libglib-2.0.so.0(+0x4858b) [0x7fd54056c58b (null) /lib64/libglib-2.0.so.0(g_main_context_dispatch+0x135) [0x7fd54056ba05 (null) /lib64/libglib-2.0.so.0(+0x47d38) [0x7fd54056bd38 (null) /lib64/libglib-2.0.so.0(g_main_loop_run+0x72) [0x7fd54056c132 (null) /lib64/libgtk-3.so.0(gtk_main+0x85) [0x7fd541fd6b25 (null) fotoxx(main+0x4c14) [0x427144 (null) /lib64/libc.so.6(__libc_start_main+0xf5) [0x7fd53f4f7975 (null) fotoxx() [0x427409 (null) >>> Regards, -- Shlomi Fish
Assignee: bugsquad => juan.baptisteTarget Milestone: --- => Mageia 3
Hi all, just a note: I was able to reproduce this problem with a fotoxx compiled from source. valgrind detected some errors in fotoxx, and I notified the maintainer, who eventually told me he was going to look into fixing them. He also said this: <<<<<<<<< I did look at the source code for one of the errors and found something that could be related. In the program f.meta.cc there is a functioin named exiftool_server() which starts out like this: char ** exiftool_server(char **inputs) // revised API v.12.06 { int ii, cc, err; static int fcf = 1; // first call flag static FILE *fid1 = 0, *fid2 = 0; // exiftool input, output files char *pp, command[100]; static char *outrecs[100]; // pointers for up to 99 output recs. char outrec[exif_maxcc]; // single output record if (! inputs) // kill exiftool process { fid1 = fopen(exiftool_input,"a"); // (also orphaned process) if (fid1) { fprintf(fid1,"-stay_open\nFalse\n"); // tell it to exit fclose(fid1); fflush(fid1); ***** THIS MUST BE WRONG ******** } remove(exiftool_input); if (fid2) pclose(fid2); fid2 = 0; fcf = 1; // start exiftool process if called again return 0; } I flagged the suspected call to fflush() which is unnecessary after fclose() and in fact the FID argument is invalid after fclose(). It may be that the newest C library is sensitive to this whereas the older one was not. This is sheer speculation but maybe worth a test to see if this is really the problem. Can you remove this line and compile and test again? thanks Mike >>>>>>>>> Regards, -- Shlomi Fish
Hi, I just updated to latest version 12.10.1, please test and report back if the crash still occurs.
Hi Juan, (In reply to comment #2) > Hi, > > I just updated to latest version 12.10.1, please test and report back if the > crash still occurs. thanks! Not I'm getting a different problem: shlomif@telaviv1:~$ fotoxx =========== fotoxx Mon Oct 8 11:09:49 language: en_GB fotoxx v.12.10.1 last file sync: Jan 01 00:00:00 2000 exiftool version: 9.01 /bin/xdg-open which: no dcraw in (/home/shlomif/apps/perl/perlbrew/bin:/home/shlomif/apps/apache-maven/apache-maven-2.1.0//bin:/home/shlomif/Download/unpack/graphics/fop/fop-0.93:/home/shlomif/apps/perl/modules/local/bin:/home/shlomif/apps/perl/modules/bin:/home/shlomif/apps/latemp/bin:/home/shlomif/apps/test/quadpres/bin:/home/shlomif/bin:/home/shlomif/apps/apache-maven/apache-maven-2.1.0//bin:/bin:/usr/bin:/usr/games:/usr/lib64/qt4/bin) /bin/ufraw And after that a dialogue is displayed that complains about the lack of dcraw and that I need to install it. It may need to be a dependency.
Ok, added dcraw as requires, please update and test again.
Hi, (In reply to comment #4) > Ok, added dcraw as requires, please update and test again. working fine now. Resolving as FIXED.
Status: NEW => RESOLVEDResolution: (none) => FIXED