Your message dated Tue, 28 May 2013 07:28:52 +0200 with message-id <[email protected]> and subject line Re: timing problems using libfuse-dev has caused the Debian Bug report #644995, regarding timing problems using libfuse-dev to be marked as done.
This means that you claim that the problem has been dealt with. If this is not the case it is now your responsibility to reopen the Bug report if necessary, and/or fix the problem forthwith. (NB: If you are a system administrator and have no idea what this message is talking about, this may indicate a serious mail system misconfiguration somewhere. Please contact [email protected] immediately.) -- 644995: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=644995 Debian Bug Tracking System Contact [email protected] with problems
--- Begin Message ---Package: libfuse-dev Version: 2.8.6-1 Severity: normal Dear Maintainer, * What led up to the situation? I implemented malloc using treedb (a package of mine in SourceForge) To handle processes that forked other processes, I implemented a page-wise copy on write system using libfuse-dev * What exactly did you do (or not do) that was effective (or ineffective)? The tests malloc-test.sh and malloc-fork-test.sh work if I provide the "-d" option (enable debug output + foreground operation), but fail if I don't. * What was the outcome of this action? I was hoping to make the treedb-malloc-daemon completely silent and to be able to use libtreedb-malloc with regular programs via LD_PRELOAD. The debugging output will affect some programs, so I can't do it. * What outcome did you expect instead? A completely silent drop-in replacement for malloc. for memory intensive programs it would 1. use cache memory instead of consuming all available memory 2. avoid paging all your existing programs out 3. avoid grinding your system to a halt as it chomped away on swap space I think we've all come across programs like that. I thought this might be a neat way to get them to play nice. Steps to reproduce 1. git clone git://v3c.git.sourceforge.net/gitroot/v3c/v3c 2. git clone git://treedb.git.sourceforge.net/gitroot/treedb/treedb 3. cd v3c-2.0.0-01 4. ./v3c-tryout treedb-1.2.0-01 tryout This creates a directory called "tryout" in the parent directory that contains the v3c and treedb directories, unpacks the version of v3c used by the specified version of treedb, builds + installs v3c, builds + installs treedb, without needing elevated (root) privileges. To "try out" the results from the v3c directory, type .../tryout/enter-env and your prompt will indicate you're in the "sandbox" cd to treedb-1.2.0-01 inside the sandbox and type "make check" All the tests should pass. Now edit the two malloc-*,sh scripts to not pass the "-d" option to treedb- malloc-daemon, and the tests fail. To see what's going on a bit more clearly, you can also 1. change -DTREEDB_DEBUG=1 and -DV3C_DEBUG=1 in libtreedb_malloc_@PACKAGE_API_VERSION@_la_CPPFLAGS in v3c/3-comet/Makefile.am 2. touch v3c/3-comet/treedb-malloc.h 3. make check again You'll see some treedb messages about the memory operations I left these disabled as, like I said, I wanted the drop-in to be silent. I don't have enough spell potions to handle the task of debugging a libfuse program (like treedb-malloc-daemon) that talks to a kernel module while being used by a forking process! -- System Information: Debian Release: wheezy/sid APT prefers proposed-updates APT policy: (500, 'proposed-updates'), (500, 'testing'), (500, 'stable') Architecture: amd64 (x86_64) Kernel: Linux 3.0.0-1-amd64 (SMP w/2 CPU cores) Locale: LANG=en_IE.UTF-8, LC_CTYPE=en_IE.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Versions of packages libfuse-dev depends on: ii libfuse2 2.8.6-1 ii libselinux1-dev [libselinux-dev] 2.1.0-1 libfuse-dev recommends no packages. Versions of packages libfuse-dev suggests: ii fuse 2.8.6-1 -- no debconf information
--- End Message ---
--- Begin Message ---Version: 2.9.2 this should be fixed with above upstream version, closing. -- Address: Daniel Baumann, Donnerbuehlweg 3, CH-3012 Bern Email: [email protected] Internet: http://people.progress-technologies.net/~daniel.baumann/
--- End Message ---
