From: SourceForge.net <no...@so...> - 2005-08-05 05:43:31
|
Bugs item #1170716, was opened at 2005-03-25 21:11 Message generated for change (Comment added) made by fellaw You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=102435&aid=1170716&group_id=2435 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None Status: Open Resolution: None Priority: 5 Submitted By: A. Haferburg (ahaferburg) Assigned to: Nobody/Anonymous (nobody) Summary: executing a shell scripts gives a memory leak Initial Comment: Hi! I've been trying to run a configure script, which gave me a huge memory leak. I've then created an empty script like this: $ vi e :wq $ chmod +x e then I wrote a script f that calls this empty script 100 times. for((i=0;i<100;i++)) do ./e done I've never written a shell script before, but it seems to work. : -) Anyway, every time I call f, approximately 2700KB of virtual memory are gone. (I check this in the task manager.) environment: Windows 2000 SP4 gcc version 3.4.2 (mingw-special) GNU ld version 2.15.91 20040904 MinGW-3.2.0-rc-3.exe MSYS-1.0.11-2004.04.30-1.exe The configure script I executed was from the k3d project. The task manager said, I was using 137740KB of memory. I ran ./configure --without-docbook --without-gnome --without-gts --without-imagemagick --without-jpeg --without-openexr --without-plib --without-png --without-python --without-tiff Memory consumption climbed up to 243396KB, then I got the error messages below. I also noted that the process IDs for the stuff called from the script climbed up to values >30000, which seems unusual to me. If you need to know anything else or want me to try sth, please drop me a mail. (Does sf.net do this automatically if someone adds a comment to this bug?) Thanks A. Haferburg said error message: 0 [main] sh 31396 sync_with_child: child 31404(0x2EC) died before initiali zation with status code 0x80 9169 [main] sh 31396 sync_with_child: *** child state waiting for longjmp ./config.status: fork: Resource temporarily unavailable 13246 [main] sh 31396 sync_with_child: child 31408(0x350) died before initiali zation with status code 0x80 19424 [main] sh 31396 sync_with_child: *** child state waiting for longjmp ./config.status: fork: Resource temporarily unavailable ./config.status: /bin/sed: Permission denied 0 [main] sh 20420 sync_with_child: child 31416(0x268) died before initiali zation with status code 0x80 114 [main] sh 20420 sync_with_child: *** child state waiting for longjmp ./config.status: fork: Resource temporarily unavailable 0 [main] sh 356 sync_with_child: child 31436(0x2BC) died before initializa tion with status code 0x80 4152 [main] sh 356 sync_with_child: *** child state waiting for longjmp ./configure: fork: Resource temporarily unavailable 8789 [main] sh 356 sync_with_child: child 31440(0x260) died before initializa tion with status code 0x80 14492 [main] sh 356 sync_with_child: *** child state waiting for longjmp ./configure: fork: Resource temporarily unavailable ---------------------------------------------------------------------- Comment By: Hans Becker (fellaw) Date: 2005-08-05 07:43 Message: Logged In: YES user_id=1324353 I also can confirm this with WinXP SP1. I even tried within safe mode to exclude any running software which might have an influence. My memory went from 462 of 512 MB free down to 240 MB free when configuring SDL12 without any options. Make started normally afterwars, but soon memory went down to 12 MB free and make aborted with severe fork errors. Env. used was MinGW 4.1.1 + msysDTK 1.0.1 + msys 1.0.10. Also tried older versions of MinGW(3.1.0.1) and msys(1.0.8), no change. Haven't tried building cvs msys yet, because the changelog didn't mention anything related to this issue. The script above climbed about the same amount of memory ahaferburg stated. ---------------------------------------------------------------------- Comment By: Hugh Perkins (hughperkins) Date: 2005-06-02 05:01 Message: Logged In: YES user_id=844774 I believe I have a similar problem to this. Its a critical blocking issue for me, because it means I cant actually run *any* configure script to completion, at least not using cccl (which obviously doubles the number of forks). Can we bump up the priority a little please? ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=102435&aid=1170716&group_id=2435 |