On 22/02/2021 17:21, Ian Jackson wrote: > Jan Beulich writes ("Re: Stable ABI checking (take 2)"): >> On 22.02.2021 15:03, Andrew Cooper wrote: >> +1 for option 2, fwiw. > I'm in favour of option 2.
Option 2 it is then. > > Andrew Cooper writes ("Re: Stable ABI checking (take 2)"): >> As far as RPM is concerned, splitting the two is important, as %build >> and %check are explicitly separate steps. I have no idea what the deb >> policy/organisation is here. > The reason why distro build systems like to distinguish "build" from > "check" (run tests) is that often the tests are time-consuming (or > have intrusive dependencies or other practical problems). > > IMO if the ABI check is very fast there is no reason not to run it by > default. (We have configure to deal with the dependency issue.) It turns out that libxl causes abi-dumper to churn for ~4s or so, which isn't ideal. All the other libraries are in the noise. If we are going for a check-by-default policy, then we obviously need to exclude the non-stable libraries by default. However, to fix problems pertaining to the unstable libraries, downstreams do need to be able to invoke checking on the other libraries as well. I'll have a think as to how best to make that happen. ~Andrew