Bug#1069488: parsinsert: FTBFS on armhf: make[1]: *** [debian/rules:30: override_dh_auto_test] Error 1
Étienne Mollier, on 2024-06-28: > Some further results on parsinsert: after rebuilding on all > release architectures, it turned out the other affected platform > was armel, in addition to armhf. This suggests possibly a > regression introduced by time_t migration to 64-bit size. Or > this could be something else I have no clue of yet. > > The other architectures: amd64, arm64, i386, mips64el, ppc64el, > s390x, and risv64, built all fine and would not require removal. I opened #1074598 to request removal of parsinsert from armel and armhf. Once this is done, the severity of the bug can be reduced, and in case someone comes up with a proper fix in the future, the issue can be closed and the package may be reintroduced for those architectures. -- .''`. Étienne Mollier : :' : pgp: 8f91 b227 c7d6 f2b1 948c 8236 793c f67e 8f0d 11da `. `' sent from /dev/pts/0, please excuse my verbosity `-on air: Queensrÿche - The Whisper signature.asc Description: PGP signature
Bug#1069488: parsinsert: FTBFS on armhf: make[1]: *** [debian/rules:30: override_dh_auto_test] Error 1
Some further results on parsinsert: after rebuilding on all release architectures, it turned out the other affected platform was armel, in addition to armhf. This suggests possibly a regression introduced by time_t migration to 64-bit size. Or this could be something else I have no clue of yet. The other architectures: amd64, arm64, i386, mips64el, ppc64el, s390x, and risv64, built all fine and would not require removal. Hope this helps, -- .''`. Étienne Mollier : :' : pgp: 8f91 b227 c7d6 f2b1 948c 8236 793c f67e 8f0d 11da `. `' sent from /dev/pts/1, please excuse my verbosity `- signature.asc Description: PGP signature
Bug#1069488: parsinsert: FTBFS on armhf: make[1]: *** [debian/rules:30: override_dh_auto_test] Error 1
Control: tags -1 + confirmed Hi, parsinsert is making a lot of assumptions on certain compiler optimisation behaviors, which have evolved in time. This has caused an increasing number of ftbfs bugs affecting a range of CPU architectures, for which the resolution consisted in dropping optimization levels, e.g. #964082. I could reproduce the present bug on armhf. This architecture was already built with -O1 optimisation, so I attempted a drop to -O0 which resolved the segmentation fault, but caused other computation errors anyway. parsinsert has not seen any newer upstream version for about twelve years, so I think we don't have much options but lookup on which CPU architectures build is broken, request for removal on such targets, then reduce the bug severity, and move on. At least the package still seems to operate properly on amd64, so we can still get away without a removal of the package. Have a nice day, :) -- .''`. Étienne Mollier : :' : pgp: 8f91 b227 c7d6 f2b1 948c 8236 793c f67e 8f0d 11da `. `' sent from /dev/pts/3, please excuse my verbosity `-on air: Genesis - .. In That Quiet Earth' signature.asc Description: PGP signature
