> How would that impact running multiple different instances of
> autopkgtest locally given keeping the data after a run? Not having the
> name of the package for which a run seems to make things more confusing:
> you'd need to track which temporary directory was used for which run, or
> go dig in the debian/control file.

Yes, this would be a bit harder to tell which dir holds a given package,
but "grep Source: ...path.../*/debian/control" is not that complicated.

> Seems like the right location here for gains really is to shorten the
> test harness' directory; from say 'autopkgtest-virt-lxc' to just 'lxc'?
> The same logic appears to seemlessly apply to other harnesses; like
> qemu. We have the potential to gain as much as 17 characters here,
> without sacrificing in clarity of what things where used for in /tmp.

The maximal gain is much more, the package names and versions can get
really long:

 udd=> select max(length(source)) , max(length(source_version)) from packages; 
  max | max 
 -----+-----
   56 |  69
 (1 row)

I think it would be reasonable to cut from the test harness directory, too, but
xnox liked this approach better.

> Also, have you discussed this with upstream / directly with pitti?
>

I initiated the discussion in the Debian bug, but got little response so far.
Note that pitti is looking for help with the package:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=846328


** Bug watch added: Debian Bug tracker #846328
   http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=846328

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1680577

Title:
  Autopkgtest fails on s390x due to long PATH in test config

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/autopkgtest/+bug/1680577/+subscriptions

-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to