Control: close -1 On Thu, Aug 08, 2019 at 09:33:18PM +0200, Paul Gevers wrote: > Hi Domenico, [also in To: to prevent the time lag] > > On 08-08-2019 00:35, Domenico Andreoli wrote: > >> On 30-07-2019 19:09, Debian Bug Tracking System wrote: > >>> * Autopkgtest on libc-bin-dbgsym instead of dwarves-dbgsym. Closes: > >>> #923717. > >> > >> Unless I am very much mistaken, this is not really fixing the issue. The > >> problem is that the unstable debug archive isn't available during > >> migration testing (that's bug #917528, to be fixed in debci). Hence, the > >> test will still fail if it needs the libc-bin-dbgsym package from > >> unstable. Please add the skip-not-installable restriction as I > >> recommended in my initial bug report. > > > > The "test" (much to be improved) just needs debug symbols for trying > > to analyze some binary. It's very simple test to prove that pahole does > > not die badly like when DWARF-4 has been introduced. > > > > At the moment does not really matter if such symbols are coming from > > testing or unstable as long as they match the version of the binary > > being analyzed. > > > > I thought to untangle a little bit the situation in using some binary > > different from dwarves itself, in that: > > > > dwarves/unstable + dwarves-dbgsym/testing => unsatisfiable > > > > Do you still think that skip-not-installable is better? Why? > > So this libc-bin-dbgsym has no relation with dwarves? Than I guess it
Indeed it has not any. > will work as you can always install the version from testing and no > relation will pull in the version from unstable. If that's the case, > just close the bug again. Otherwise, skip-not-installable *on top of* > other improvements will just mean that *if* any package is unsatisfiable > than the test is marked as skipped and as such adds a neutral result to > the final outcome. Thanks, I'll consider it. Dom -- 3B10 0CA1 8674 ACBA B4FE FCD2 CE5B CF17 9960 DE13
