Bug#792053: [pkg-go] Bug#792053: prometheus: FTBFS w/ test suite errors
Hi Aaron, On 25/08/15 17:13, Aaron M. Ucko wrote: > I'm happy to confirm that 0.15.1+ds-1 fixes the original errors. > However, new ones have cropped up, per > https://buildd.debian.org/status/logs.php?pkg=prometheus&ver=0.15.1%2Bds-1 : > TestEvictAndLoadChunkDescsType0 fails on armhf, and > TestEvictAndLoadChunkDescsType1 fails on armel and i386. Could you > please take another look? I have created an upstream bug to track this: https://github.com/prometheus/prometheus/issues/1031 I would have created a new bts bug, as it is a different issue.. But not a big deal :) > Incidentally, I did not explicitly request an update to a new upstream > release (though I certainly don't object!), so best practice would have > been to note in the changelog what this bug report was actually about. > You'll get a chance yet. ;-) Yes, sorry, I completely forgot to better explain this. The new upstream release was not because of this bug, but since I had already prepared it, and it included the fix to this bug, I just lumped all together. Thanks for the report! -- Martín Ferrari (Tincho)
Bug#792053: prometheus: FTBFS w/ test suite errors
found 792053 0.15.1+ds-1 notfixed 792053 0.15.1+ds-1 thanks "Aaron M. Ucko" writes: > Builds of prometheus on those architectures on which its build > dependencies are available (so far just armel, armhf, and i386, aside > from amd64) have been failing with assorted test suite errors: I'm happy to confirm that 0.15.1+ds-1 fixes the original errors. However, new ones have cropped up, per https://buildd.debian.org/status/logs.php?pkg=prometheus&ver=0.15.1%2Bds-1 : TestEvictAndLoadChunkDescsType0 fails on armhf, and TestEvictAndLoadChunkDescsType1 fails on armel and i386. Could you please take another look? Incidentally, I did not explicitly request an update to a new upstream release (though I certainly don't object!), so best practice would have been to note in the changelog what this bug report was actually about. You'll get a chance yet. ;-) Thanks! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu
Bug#792053: [pkg-go] Bug#792053: Bug#792053: Bug#792053: prometheus: FTBFS w/ test suite errors
Upstream has been very helpful digging into this issue. We have been able to pin this down to a difference in the go compiler. If I use upstream's Makefile to download the go runtime from golang.org and then use that to build the package, the errors go away. Now, it might not actually be a bug in the compiler, as this seems it could be related to undefined behaviour when casting to unsigned values (the upstream bug has more details). -- Martín Ferrari (Tincho) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#792053: [pkg-go] Bug#792053: Bug#792053: prometheus: FTBFS w/ test suite errors
forwarded 792053 https://github.com/prometheus/prometheus/issues/887 thanks On 10/07/15 20:34, Aaron M. Ucko wrote: > Great, thanks. As I said, I suspect at least some of the errors are due > to buildd quirks rather than architectural differences, so you might > need to install sbuild and perhaps also adjust its configuration to > match the autobuilders more closely. (I'm not familiar with the setup > details myself, so I can't offer any specifics here, sorry.) It was not needed. I just tried to rebuild the package in abel.d.o (armel), and I got the same test failures. I have opened a bug upstream to see if they know what can be causing this. Thanks again. -- Martín Ferrari (Tincho) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#792053: [pkg-go] Bug#792053: prometheus: FTBFS w/ test suite errors
Martín Ferrari writes: > This is pretty weird.. There are a few precision mismatches, but many > instances of "expected 1, got 0". I will get some 32 bit machine and > take a look. Great, thanks. As I said, I suspect at least some of the errors are due to buildd quirks rather than architectural differences, so you might need to install sbuild and perhaps also adjust its configuration to match the autobuilders more closely. (I'm not familiar with the setup details myself, so I can't offer any specifics here, sorry.) > Thanks for reporting! Thanks for following up right away! -- Aaron M. Ucko, KB1CJC (amu at alum.mit.edu, ucko at debian.org) http://www.mit.edu/~amu/ | http://stuff.mit.edu/cgi/finger/?a...@monk.mit.edu -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#792053: [pkg-go] Bug#792053: prometheus: FTBFS w/ test suite errors
Hi Aaron, On 10/07/15 19:26, Aaron M. Ucko wrote: > Builds of prometheus on those architectures on which its build > dependencies are available (so far just armel, armhf, and i386, aside > from amd64) have been failing with assorted test suite errors: > > https://buildd.debian.org/status/logs.php?pkg=prometheus&ver=0.14.0%2Bds-1 > > Architectural differences might play a part here (all attempted > builds were 32-bit), but I suspect buildd configuration quirks may > also be relevant, so I've gone with a severity of serious. This is pretty weird.. There are a few precision mismatches, but many instances of "expected 1, got 0". I will get some 32 bit machine and take a look. Thanks for reporting! -- Martín Ferrari (Tincho) -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#792053: prometheus: FTBFS w/ test suite errors
Source: prometheus Version: 0.14.0+ds-1 Severity: serious Justification: fails to build from source Builds of prometheus on those architectures on which its build dependencies are available (so far just armel, armhf, and i386, aside from amd64) have been failing with assorted test suite errors: https://buildd.debian.org/status/logs.php?pkg=prometheus&ver=0.14.0%2Bds-1 Architectural differences might play a part here (all attempted builds were 32-bit), but I suspect buildd configuration quirks may also be relevant, so I've gone with a severity of serious. Could you please take a look? Thanks! -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
