On 08/19/2009 08:09 PM, Toshio Kuratomi wrote:
On 08/19/2009 07:10 PM, Jesse Keating wrote:
> The host xz wouldn't be used to produce any rpms, the rpm
inside the
> chroot would. Does this come into play when initing the buildroot?
>
You're right, this wouldn't come into play unless it's a decompression
bug. And if that's so it would generate an error from the buildsystem
while trying to create the buildroot instead of a corrupted payload in
the built rpms. So not as severe. I'm checking to be sure it isn't a
decompression bug now.
Confirmed -- the compressor is the issue here, not the decompressor. So
we don't need to update the builders at this time. Since rawhide is the
only release building with xz payloads we don't need to worry about
buildroot overrides either.
It *is* possible that some of the packages built before this xz package
was put into the buildroot are corrupt::
http://koji.fedoraproject.org/koji/buildinfo?buildID=127510
(build finished at 2009-08-17 10:32:19) I don't know if this is
something releng wants to check for.
-Toshio