Mathias Gug wrote:
> Now that kerberos 1.7 has been uploaded to unstable I've looked into
> syncing the krb5 package in Ubuntu. Bringing 1.7 in the archive requires
> a rebuild of all packages depending on libkrb53. The list of rdepends
> breaks down to 42 source packages in main, 101 in universe and 2 in
> multiverse for a grand total of 145 source packages to upload to the
> archive once krb5 has been built and published.
> How should this transition be tracked:
> 1. File 1 bug in LP and open a task for 145 packages?
> 2. File 145 bugs in LP (one for each package) and tag all of them with
> 3. Use a wiki page?
> Which option is best suited for this use case?
I tend to prefer to just use NBS for this sort of thing. Ask for
rebuild volunteers in #ubuntu-devel, and split the package set into
chunks of 30 or so for testing simple rebuilds. It shouldn't take more
than a day or two for 5-7 people to push the necessary uploads (and it's
fairly easy to script automated rebuild tests). Those that don't work
as a simple rebuild might deserve bugs for tracking, but it's a bit of
overhead to file a massive number of bugs, and then try to close them
all, especially as we'll probably get some of the fixes for free from
ubuntu-devel mailing list
Modify settings or unsubscribe at: