On 2023-06-24 16:58:15, Ilias Tsitsimpis wrote: > Control: forwarded -1 https://github.com/ndmitchell/hoogle/issues/359 > > Hi Iustin, > > On Sun, Apr 23, 2023 at 11:25PM, Iustin Pop wrote: > > AFAIK, both armel and armfh are low-powered/"slow" arches, but i386 is > > surprising. Maybe due to memory limits? > > > > I wonder if tests shouldn't simply be restricted to amd64, arm64, ppc64el > > and > > s390x? This should give it enough of architecture heterogeneity to catch > > any problems that simply do not appear on amd64 (mainstream arch) (yes, > > I've been bitten by this in one package of mine). > > > > This would be a cheap fix (one entry in the control file). Thoughts? > > (Anyone?) It seems this bug is the only one that prevents it from > > entering testing (and it's a leaf package). > > Unfortunately it's not just the tests that fail, hoogle is currently > completely broken on all 32-bit architectures. We need to either fix it, > or remove hoogle from these architectures.
Ooh, interesting. I agree with Neil's last comment in the upstream bug - Hoogle was even years ago when I last built it a memory/CPU hog, so I think restricting to 64 bit is better than not having it. My 2c. iustin
