Bug#568607: libogg-vorbis-perl: dup syscall yields fd leak

2010-02-07 Thread Vincent Lefevre
On 2010-02-07 17:51:04 +0100, Vincent Lefevre wrote: > On 2010-02-07 16:45:48 +0200, Niko Tyni wrote: > > As you're clearly using Ogg::Vorbis, did you have a particular reason > > for choosing it over the others? Is there something it can do and > > the others can't? > > The only reason is that I

Bug#568607: libogg-vorbis-perl: dup syscall yields fd leak

2010-02-07 Thread Damyan Ivanov
-=| Niko Tyni, Sun, Feb 07, 2010 at 04:45:48PM +0200 |=- > However, I think this package (and libogg-vorbis-header-perl > too for that matter) should rather be removed from Debian as > it's clearly dead upstream and there are maintained alternatives > (libogg-vorbis-header-pureperl-perl and libogg-

Bug#568607: libogg-vorbis-perl: dup syscall yields fd leak

2010-02-07 Thread Vincent Lefevre
On 2010-02-07 16:45:48 +0200, Niko Tyni wrote: > As you're clearly using Ogg::Vorbis, did you have a particular reason > for choosing it over the others? Is there something it can do and > the others can't? The only reason is that I didn't know the other ones (I wonder whether they existed when I

Bug#568607: libogg-vorbis-perl: dup syscall yields fd leak

2010-02-07 Thread Niko Tyni
On Sun, Feb 07, 2010 at 02:09:09AM +0100, Vincent Lefevre wrote: > tags 568607 patch > thanks > > Actually, $ogg->clear solves the problem. But then, the clear method > should be called when the object is destroyed. I've attached a patch. > This patch calls ov_clear directly, but this seems to wor

Bug#568607: libogg-vorbis-perl: dup syscall yields fd leak

2010-02-06 Thread Vincent Lefevre
tags 568607 patch thanks Actually, $ogg->clear solves the problem. But then, the clear method should be called when the object is destroyed. I've attached a patch. This patch calls ov_clear directly, but this seems to work even when ov_clear has already been called (doing $ogg->clear twice is allo

Bug#568607: libogg-vorbis-perl: dup syscall yields fd leak

2010-02-06 Thread Vincent Lefevre
On 2010-02-06 21:59:24 +0100, Vincent Lefevre wrote: > With the following testcase, if I provide any (non-corrupt) ogg file > as an argument, I get a segmentation fault after the number 1021 is > printed. Actually the cause is that a file descriptor is created by > the dup syscall at each iteration

Bug#568607: libogg-vorbis-perl: dup syscall yields fd leak

2010-02-06 Thread Vincent Lefevre
retitle 568607 libogg-vorbis-perl: dup syscall yields fd leak -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Bug#568607: libogg-vorbis-perl: dup syscall yields fd leak

2010-02-06 Thread Vincent Lefevre
reassign 568607 libogg-vorbis-perl found 568607 libogg-vorbis-perl/0.05.ds1-1+b1 thanks On 2010-02-06 04:32:37 +0100, Vincent Lefevre wrote: > I have a script that uses XML::LibXML that takes more and more memory > and ends up with a segmentation fault (before the memory is exhausted, > though). >