On 13/12/17 at 10:33 +0000, Holger Levsen wrote: > On Wed, Dec 13, 2017 at 10:49:37AM +0800, Paul Wise wrote: > > > would that be something for the reproducible project? > > Best ask on their mailing lists, but I guess it would significantly > > increase their resource usage to build three times instead of two. > > also: the more problems (unrelated to reproducible builds) we are > finding, the more we are being distracted from our goal, so this would > not only be draining on computing ressources, but also on our attention > span. > > (also we currently again have a modified gcc (and 3 other packages) > when testing unstable…) > > archive-wide rebuilds are still useful.
On my side, I have very little time for Debian stuff those days, and got a bit demotivated lately from doing archive rebuilds after interactions after my last bunch of bug filing. I do plan to return to doing archive rebuilds at some point though. However, regarding "fails to build twice in a row", I'm not sure of the relevance those days, where it's fairly easy to use e.g. git to return to a clean state when hacking on packages. Probably the severity of those bugs should be discussed on -devel@ before doing a MBF. If there's agreement that such a MBF is a good idea, someone is willing to do the bug filing work, I can do a rebuild on the AWS ressources and providing a list of failing packages + logs (that's only a few hours of work). But I won't file bugs myself (and that's the big part of the work). Lucas
