Bug#969672: gnutls28: please upgrade to guile-3.0 soon, if feasible
On 2020-11-28 Rob Browning wrote: > Andreas Metzler writes: [...] > I also wonder if it could be something we're setting in the debian/rules > environment, since the tests work fine if I run: > make -C b4deb check > after the debian/rules build hangs and I C-c it. Hello, It is not debian/rules but somehow related to debhelper: make -C b4deb/guile/ check VERBOSE=1 TESTS="tests/reauth.scm" succeeds while dh_auto_test --builddirectory=b4deb/guile/ -- TESTS="tests/reauth.scm" hangs. ... It is not the environment but stdin. This also hangs: env GUILE_AUTO_COMPILE=0 make -C b4deb/guile/ check TESTS="tests/reauth.scm" VERBOSE=1 < /dev/null cu Andreas -- `What a good friend you are to him, Dr. Maturin. His other friends are so grateful to you.' `I sew his ears on from time to time, sure'
Bug#969672: gnutls28: please upgrade to guile-3.0 soon, if feasible
Andreas Metzler writes: > Abovementioned reproducer works on amd64. I can reproduce it now, but before I spend too much time trying to track it down, is it a requirement that debian packages be able to build and test in /dev/shm? I'll be a little surprised if that's feasible for all our packages, and my current guess is that maybe /dev/shm doesn't behave the way either gnutls or guile expects. I also wonder if it could be something we're setting in the debian/rules environment, since the tests work fine if I run: make -C b4deb check after the debian/rules build hangs and I C-c it. Thanks -- Rob Browning rlb @defaultvalue.org and @debian.org GPG as of 2011-07-10 E6A9 DA3C C9FD 1FF8 C676 D2C4 C0F0 39E9 ED1B 597A GPG as of 2002-11-03 14DD 432F AE39 534D B592 F9A0 25C8 D377 8C7E 73A4
Bug#969672: gnutls28: please upgrade to guile-3.0 soon, if feasible
On 2020-11-25 Rob Browning wrote: > Andreas Metzler writes: >> I think I have described how to reproduce this on amd64 in the other >> bugreport: >> (sid)ametzler@argenau:/$ rm -rf /tmp/GNUTLS/ /dev/shm/GNUTLS >> (sid)ametzler@argenau:/$ mkdir /tmp/GNUTLS/ /dev/shm/GNUTLS >> (sid)ametzler@argenau:/$ cd /dev/shm/GNUTLS >> (sid)ametzler@argenau:/dev/shm/GNUTLS$ apt source gnutls28 ; ln -s >> /dev/shm/GNUTLS/gnutls28-3.6.15 /tmp/GNUTLS/ ; cd /tmp/GNUTLS/gnutls28-3.6.15 >> (sid)ametzler@argenau:/tmp/GNUTLS/gnutls28-3.6.15$ dpkg-buildpackage -uc -us >> -j1 -B > Does it happen if you build in a more typical filesystem? It is the simplest setup I found. > It just built > fine on amdahl (arm64 porterbox), with a current sid chroot, and the -5 > package from experimental, via "fakeroot debian/rules binary". > I can't do what you describe above there because /dev/shm is 64M in the > porterbox chroots. Doesn't your own local development machine have a bigger tmpfs? Abovementioned reproducer works on amd64. cu Andreas -- `What a good friend you are to him, Dr. Maturin. His other friends are so grateful to you.' `I sew his ears on from time to time, sure'
Bug#969672: gnutls28: please upgrade to guile-3.0 soon, if feasible
Andreas Metzler writes: > I think I have described how to reproduce this on amd64 in the other > bugreport: > > (sid)ametzler@argenau:/$ rm -rf /tmp/GNUTLS/ /dev/shm/GNUTLS > (sid)ametzler@argenau:/$ mkdir /tmp/GNUTLS/ /dev/shm/GNUTLS > (sid)ametzler@argenau:/$ cd /dev/shm/GNUTLS > (sid)ametzler@argenau:/dev/shm/GNUTLS$ apt source gnutls28 ; ln -s > /dev/shm/GNUTLS/gnutls28-3.6.15 /tmp/GNUTLS/ ; cd /tmp/GNUTLS/gnutls28-3.6.15 > (sid)ametzler@argenau:/tmp/GNUTLS/gnutls28-3.6.15$ dpkg-buildpackage -uc -us > -j1 -B Does it happen if you build in a more typical filesystem? It just built fine on amdahl (arm64 porterbox), with a current sid chroot, and the -5 package from experimental, via "fakeroot debian/rules binary". I can't do what you describe above there because /dev/shm is 64M in the porterbox chroots. Thanks -- Rob Browning rlb @defaultvalue.org and @debian.org GPG as of 2011-07-10 E6A9 DA3C C9FD 1FF8 C676 D2C4 C0F0 39E9 ED1B 597A GPG as of 2002-11-03 14DD 432F AE39 534D B592 F9A0 25C8 D377 8C7E 73A4
Bug#969672: gnutls28: please upgrade to guile-3.0 soon, if feasible
On 2020-11-24 Rob Browning wrote: > Hmm, didn't see this until now (don't think I received a mail from the > bug tracker). > Andreas Metzler writes: >> a full build fails reproducibly for me when building in /dev/shm. > Building in /dev/shm? And on what architecture, etc. >> I did a test upload to experimental yesterday, it also failed on >> arm64, since a guile test did hang. > If we think it's arm64 specific, I can try that on one of the > porterboxes -- think a a sid chroot should suffice? Hello Rob, I think I have described how to reproduce this on amd64 in the other bugreport: (sid)ametzler@argenau:/$ rm -rf /tmp/GNUTLS/ /dev/shm/GNUTLS (sid)ametzler@argenau:/$ mkdir /tmp/GNUTLS/ /dev/shm/GNUTLS (sid)ametzler@argenau:/$ cd /dev/shm/GNUTLS (sid)ametzler@argenau:/dev/shm/GNUTLS$ apt source gnutls28 ; ln -s /dev/shm/GNUTLS/gnutls28-3.6.15 /tmp/GNUTLS/ ; cd /tmp/GNUTLS/gnutls28-3.6.15 (sid)ametzler@argenau:/tmp/GNUTLS/gnutls28-3.6.15$ dpkg-buildpackage -uc -us -j1 -B cu Andreas -- `What a good friend you are to him, Dr. Maturin. His other friends are so grateful to you.' `I sew his ears on from time to time, sure'
Bug#969672: gnutls28: please upgrade to guile-3.0 soon, if feasible
Hmm, didn't see this until now (don't think I received a mail from the bug tracker). > Andreas Metzler writes: > a full build fails reproducibly for me when building in /dev/shm. Building in /dev/shm? And on what architecture, etc. > I did a test upload to experimental yesterday, it also failed on > arm64, since a guile test did hang. If we think it's arm64 specific, I can try that on one of the porterboxes -- think a a sid chroot should suffice? Thanks -- Rob Browning rlb @defaultvalue.org and @debian.org GPG as of 2011-07-10 E6A9 DA3C C9FD 1FF8 C676 D2C4 C0F0 39E9 ED1B 597A GPG as of 2002-11-03 14DD 432F AE39 534D B592 F9A0 25C8 D377 8C7E 73A4
Bug#969672: gnutls28: please upgrade to guile-3.0 soon, if feasible
On 2020-10-03 Rob Browning wrote: > Andreas Metzler writes: > > I would love to, but guile-3.0 >= 3.0.4-1+b1 seems to be broken. See > > 969640. > Assuming I did it right (guile-3.0-dev was installed along with all > the other build deps, and guile-2.2-dev wasn't), a "fakeroot > debian/rules binary" just built without trouble here using the current > 3.6.15-4 package. > I wonder if the problem has been resolved, or maybe it's intermittent, > arch dependent, or something else. And was the failure local, or on the > buildds? > I can double-check in a sid sbuild chroot later. (When it succeeded, I > was just using the system I had handy, which is somewhere between > bullseye and sid (x86_64).) Hello, a full build fails reproducibly for me when building in /dev/shm. I did a test upload to experimental yesterday, it also failed on arm64, since a guile test did hang. cu Andreas -- `What a good friend you are to him, Dr. Maturin. His other friends are so grateful to you.' `I sew his ears on from time to time, sure'
Bug#969672: gnutls28: please upgrade to guile-3.0 soon, if feasible
Andreas Metzler writes: > I would love to, but guile-3.0 >= 3.0.4-1+b1 seems to be broken. See > 969640. Assuming I did it right (guile-3.0-dev was installed along with all the other build deps, and guile-2.2-dev wasn't), a "fakeroot debian/rules binary" just built without trouble here using the current 3.6.15-4 package. I wonder if the problem has been resolved, or maybe it's intermittent, arch dependent, or something else. And was the failure local, or on the buildds? I can double-check in a sid sbuild chroot later. (When it succeeded, I was just using the system I had handy, which is somewhere between bullseye and sid (x86_64).) -- Rob Browning rlb @defaultvalue.org and @debian.org GPG as of 2011-07-10 E6A9 DA3C C9FD 1FF8 C676 D2C4 C0F0 39E9 ED1B 597A GPG as of 2002-11-03 14DD 432F AE39 534D B592 F9A0 25C8 D377 8C7E 73A4
Bug#969672: gnutls28: please upgrade to guile-3.0 soon, if feasible
Control: reassign 969672 guile-gnutls Control: forcemerge 964284 969672 On 2020-09-06 Rob Browning wrote: > Source: gnutls28 > Severity: wishlist > Please migrate to guile-3.0 as soon as it's feasible. If we can, I'd > like to have the option to drop guile-2.2 from bullseye, so that we > won't have to maintain two versions throughout that release. I would love to, but guile-3.0 >= 3.0.4-1+b1 seems to be broken. See 969640. cu Andreas
