Bug#702102: fails to upgrade (cowbuilder) chroot

Chris Knadle Chris.Knadle at coredump.us
Tue Mar 26 14:18:47 UTC 2013


On Tuesday, March 26, 2013 08:46:08, Michael Biebl wrote:
> Am 26.03.2013 09:48, schrieb Junichi Uekawa:
> > not enough information in the bug, 702811 seems to be a better bug.
> 
> Say what? Have you read the full bug report, including the analysis that
> it is because of /run/shm vs /dev/shm?

Related to this, a patch was incorporated from the following email to handle 
another bug.

   http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=700591#5

Based on the idea that it's somehow this patch that's causing this new 
problem, I'm "pulling at straws" to try to help figure out why.

You mentioned you have /run/shm but not /dev/shm.  Based on that I'm looking 
at the following logic from the patch:

+    if [ "$DEB_BUILD_ARCH_OS" = "linux" ] && [ "$USEDEVSHM" = "yes" ]; then
+       SHM_PATH="run/shm"                                                                                                               
+       [ ! -d "/$SHM_PATH" ] && SHM_PATH="dev/shm"

This looks to me like it's checking if /run/shm exists (and is a directory), 
and if it doesn't then blindly uses /dev/shm instead.  This seems reasonable 
but I'm wondering if there's some way for this test to fail.  If /run/shm were 
a /device/ rather than a directory or a softlink to a directory, that would be 
one way for that to happen.

  -- Chris

--
Chris Knadle
Chris.Knadle at coredump.us



More information about the Pbuilder-maint mailing list