On Sat, 2020-11-28 at 11:07 -0800, tony mancill wrote: > On Sat, Nov 28, 2020 at 04:56:28PM +0000, Adam D. Barratt wrote: > > On Sat, 2020-11-28 at 08:01 -0800, tony mancill wrote: > > > On Sat, Nov 28, 2020 at 05:10:39PM +0200, Adrian Bunk wrote: > > > > On Wed, Nov 25, 2020 at 08:23:45PM -0800, tony mancill wrote: > > > > > -with_check = disabled for this upload > > > > > +# with_check = disabled for this upload > > > > > ... > > > > > > > > FTR, this change increases the build time on zero architectures > > > > by > > > > half a week of running tests. > > > > > > > > E.g. on mipsel-osuosl-01 that is currently building mips64el > > > > this > > > > adds 3 days 23 hours to the buildtime.[1] > > > > > > > > Which is a bit wasteful since failures are ignored and a 3 > > > > digit > > > > number of tests fail on all release architectures. > > [...] > > > Ugh, that is unfortunate. The "with_check" change is there to > > > keep > > > the package as close to the version in testing/unstable as > > > feasible. The only sourceful packaging difference is the build- > > > dep > > > on g++-8 instead of g++-10. > > > > > > Adam expressed concerns about the change as well. Do I need to > > > prepare another upload? > > > > Given that this weekend is the freeze for the point release, that > > might > > be safest, just in case the builds show any issues. It's at least > > not a > > regression from the version currently in buster. > > I assumed that I needed to bump the revision instead of overwriting > deb10u1; debdiff for the deb10u2 source package attached.
That's correct. > Let me know if that looks okay and I can proceed with the upload. As this is the second revision, it doesn't need the: + * Rebuild for Buster (Closes: #975728) but if it makes it easier to get the upload sorted more quickly then I also don't mind it being included. Please feel free to upload. Regards, Adam
