Re: Trivial bugs hurting progress
On Wed, Mar 4, 2015 at 8:44 AM, Martin Packman martin.pack...@canonical.com wrote: On 04/03/2015, Andrew Wilkins andrew.wilk...@canonical.com wrote: Curtis and I have talked about also doing a ppc64 test run as part of the gating job, that gets us the map ordering stuff as a newer go would, and other gccgo quirks covered as well. We don't want to bump the compiler version yet I think, as we want to be able to build with the trusty toolchain still, and not accidentally let in regressions by depending on newer gos. I assume Core developers are like myself. I have run the test suite just like the lander. I know a failure to merge is probably a bad test. As CI is running more compilers, series, and archs, We could prevent regressions in master by using lesser combination that must pass. We could: - catch and fix map order bugs - skip tests that wont pass on an arch/compiler combinations - discover and fix real arch and series issues without ever breaking the build, never blocking other engineers from merging. We cannot achieve my ideal combination of precise, gcc-4.9+patches, and ppc64el or i386. Precise doesn't work with ppc64el, and it is hard to get a run 386 test quickly. -- Curtis Hovey Canonical Cloud Development and Operations http://launchpad.net/~sinzui -- Juju-dev mailing list Juju-dev@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev
Re: Trivial bugs hurting progress
On 04/03/15 19:05, Andrew Wilkins wrote: The bugs were fixed in this branch: https://github.com/juju/juju/pull/1738 - invalid printf-style formatting will cause go vet to fail My bad. I had deleted the pre-push hook script some time ago when go vet failed continuously, and I hadn't reinstated it. I have done so now. Tim -- Juju-dev mailing list Juju-dev@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev
Re: Trivial bugs hurting progress
On Thu, Mar 5, 2015 at 10:24 AM, Tim Penhey tim.pen...@canonical.com wrote: On 04/03/15 19:05, Andrew Wilkins wrote: The bugs were fixed in this branch: https://github.com/juju/juju/pull/1738 - invalid printf-style formatting will cause go vet to fail My bad. I had deleted the pre-push hook script some time ago when go vet failed continuously, and I hadn't reinstated it. I have done so now. Thanks. FYI (and others' I), there's a couple of ways to temporarily disable hooks or just vet: git push --no-verify or IGNORE_VET_WARNINGS=1 git push (obviously to be used judiciously) -- Juju-dev mailing list Juju-dev@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev
Re: Trivial bugs hurting progress
One option is to run the test suite with go 1.3 and just make sure that juju compiles with all the other compilers (1.2, gccgo, etc). That gives us a fast precommit check, which won't catch everything, but should catch most 1.2 compat bugs. And leave the full test suite runs as CI tests. John =:- On Thu, Mar 5, 2015 at 6:28 AM, Andrew Wilkins andrew.wilk...@canonical.com wrote: On Thu, Mar 5, 2015 at 10:24 AM, Tim Penhey tim.pen...@canonical.com wrote: On 04/03/15 19:05, Andrew Wilkins wrote: The bugs were fixed in this branch: https://github.com/juju/juju/pull/1738 - invalid printf-style formatting will cause go vet to fail My bad. I had deleted the pre-push hook script some time ago when go vet failed continuously, and I hadn't reinstated it. I have done so now. Thanks. FYI (and others' I), there's a couple of ways to temporarily disable hooks or just vet: git push --no-verify or IGNORE_VET_WARNINGS=1 git push (obviously to be used judiciously) -- Juju-dev mailing list Juju-dev@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev -- Juju-dev mailing list Juju-dev@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev
Re: Trivial bugs hurting progress
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 2015-03-04 08:53 AM, Dimiter Naydenov wrote: If it's about catching map ordering issues, let's use go 1.3+ as an extra step, rather than gccgo, which is buggy and slow and randomly segfaults. It's likely we'll slow down merge gating considerably. It's not just about map ordering issues. It's about complete compatibility with GCC Go. For example, there are issues with test doubles that have had to be worked around in the past. When a GCC Go incompatibility is landed, we sometimes spend days with an unreleasable Juju before it gets fixed. It has been 3 days since master was releasable. Even though GCC Go is slow, it's worth it to avoid three days of being unreleasable. My 2ยข. Aaron -BEGIN PGP SIGNATURE- Version: GnuPG v1 iQEcBAEBAgAGBQJU9xieAAoJEK84cMOcf+9hDg0H/j6TYPNbI92arsWimT21bxve 1TOBvXtlGJCAwsiZOU//mEKwHXmChrm/a8F2YeWlMDAvk4+Uw4Wid5EViucZiNK9 +IhrpCqG835PMhyxVRbOVzSYoxsjrn1E1Kb5Q+tcOWqrjntoj4JrFMKeCIuANTVj yi56sl1jNNtOKV0SpmGnILuNNrFVXpy0i4P0Iz462V5sbQ5VtSXwtRc1rtK5OJA6 b47cv346z+DqPJahWNknG+yt2RaGQWzOfTST46B3kiU9CEKsH/U4P3nY9Ixx6hBz uH8FHpQpfHByK2FPTs9+thnViD30YUaPdFNUnT8FfRnEaNjpqaDEs4jgOafI8LE= =4tUi -END PGP SIGNATURE- -- Juju-dev mailing list Juju-dev@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev
Re: Trivial bugs hurting progress
On Wed, Mar 4, 2015 at 1:05 AM, Andrew Wilkins andrew.wilk...@canonical.com wrote: The bot is currently running (I think) Go 1.2. I'm running 1.4, Ian's running 1.3, and I'm sure Dave's running tip ;) Go 1.3+ made map iteration less deterministic, so these sorts of bugs are much more likely to occur after 1.2. It'd be good to either bump the compiler version to something more recent, unless we want to gate things on multiple compilers (maybe gc and gccgo, seeing as we currently use both). While we add our golang archive to all the slaves to ensure golang 1.2.1 is available, apt prefers to use the latest version. On Vivid, that is golang 1.3, and it very erratic success rate is probably because of tests aren't ready for a newish golang. Fixing the test suite will give allow us to make the vivid unit tests voting. Our test runner script accepts a PPA option. We have added newer golang and mongos to our experimental PPA, then setup a test to run with it to prove the Juju was ready for the newer versions. We can do this again. -- Curtis Hovey Canonical Cloud Development and Operations http://launchpad.net/~sinzui -- Juju-dev mailing list Juju-dev@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev
Re: Trivial bugs hurting progress
On 04/03/2015, Andrew Wilkins andrew.wilk...@canonical.com wrote: Hi all, Ian asked me to mail the list about a couple of bugs that managed to get past the bot; first so that we can all be mindful of these sorts of bugs, and second to highlight the fact that they could have been caught by the bot. The bugs were fixed in this branch: https://github.com/juju/juju/pull/1738 - random. iteration Map is This was being addressed as part of bug 1427149. https://bugs.launchpad.net/juju-core/+bug/1427149 http://reviews.vapour.ws/r/1048/ Your patch means that branch now needed updating. - invalid printf-style formatting will cause go vet to fail So, do we want to revisit making the bot fail on go vet errors? That's trivial flip. The bot is currently running (I think) Go 1.2. I'm running 1.4, Ian's running 1.3, and I'm sure Dave's running tip ;) Go 1.3+ made map iteration less deterministic, so these sorts of bugs are much more likely to occur after 1.2. It'd be good to either bump the compiler version to something more recent, unless we want to gate things on multiple compilers (maybe gc and gccgo, seeing as we currently use both). Curtis and I have talked about also doing a ppc64 test run as part of the gating job, that gets us the map ordering stuff as a newer go would, and other gccgo quirks covered as well. We don't want to bump the compiler version yet I think, as we want to be able to build with the trusty toolchain still, and not accidentally let in regressions by depending on newer gos. Martin -- Juju-dev mailing list Juju-dev@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev
Re: Trivial bugs hurting progress
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 4.03.2015 15:44, Martin Packman wrote: On 04/03/2015, Andrew Wilkins andrew.wilk...@canonical.com wrote: Hi all, Ian asked me to mail the list about a couple of bugs that managed to get past the bot; first so that we can all be mindful of these sorts of bugs, and second to highlight the fact that they could have been caught by the bot. The bugs were fixed in this branch: https://github.com/juju/juju/pull/1738 - random. iteration Map is This was being addressed as part of bug 1427149. https://bugs.launchpad.net/juju-core/+bug/1427149 http://reviews.vapour.ws/r/1048/ Your patch means that branch now needed updating. - invalid printf-style formatting will cause go vet to fail So, do we want to revisit making the bot fail on go vet errors? That's trivial flip. The bot is currently running (I think) Go 1.2. I'm running 1.4, Ian's running 1.3, and I'm sure Dave's running tip ;) Go 1.3+ made map iteration less deterministic, so these sorts of bugs are much more likely to occur after 1.2. It'd be good to either bump the compiler version to something more recent, unless we want to gate things on multiple compilers (maybe gc and gccgo, seeing as we currently use both). Curtis and I have talked about also doing a ppc64 test run as part of the gating job, that gets us the map ordering stuff as a newer go would, and other gccgo quirks covered as well. We don't want to bump the compiler version yet I think, as we want to be able to build with the trusty toolchain still, and not accidentally let in regressions by depending on newer gos. If it's about catching map ordering issues, let's use go 1.3+ as an extra step, rather than gccgo, which is buggy and slow and randomly segfaults. It's likely we'll slow down merge gating considerably. Just my 0.02c. Dimiter Martin - -- Dimiter Naydenov dimiter.nayde...@canonical.com Juju Core Sapphire team http://juju.ubuntu.com -BEGIN PGP SIGNATURE- Version: GnuPG v1 iQEcBAEBAgAGBQJU9w5NAAoJENzxV2TbLzHw+lsH/jDFDXb+APntIyq1WUzIfRb2 cbg3vurBAqvIcT6qt6pMTW3uEf1Q9daksvqSRFhC8Nbz70bIeLJ3vLNHzn6gRy/d z9QiyJQIXQE8208bmQLYdOkbLZAfURzT8I6PTFzioI7HkL7qA3suWod1pnEwSvcQ L+bw3+H/gym9bUGqDG2zG1p6u+Z0JH4lxMBomcxsaU1WJyBFN0HVi0xinJN5IX7k YZoZokaD6C0pu30r756wMy1PigImm/+G/lxLqzlnpL32iZFZhQOnWPZMxMVOtfIM AGzSAJtdy149jcnfLHJqeb4yNugTKUxPrARrFVkBbYKau4I09eGOes7yKq7j8fA= =8QrH -END PGP SIGNATURE- -- Juju-dev mailing list Juju-dev@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev
Trivial bugs hurting progress
Hi all, Ian asked me to mail the list about a couple of bugs that managed to get past the bot; first so that we can all be mindful of these sorts of bugs, and second to highlight the fact that they could have been caught by the bot. The bugs were fixed in this branch: https://github.com/juju/juju/pull/1738 - random. iteration Map is - invalid printf-style formatting will cause go vet to fail The bot is currently running (I think) Go 1.2. I'm running 1.4, Ian's running 1.3, and I'm sure Dave's running tip ;) Go 1.3+ made map iteration less deterministic, so these sorts of bugs are much more likely to occur after 1.2. It'd be good to either bump the compiler version to something more recent, unless we want to gate things on multiple compilers (maybe gc and gccgo, seeing as we currently use both). As for formatting: there is a pre-push hook that runs go vet to catch, amongst other things, printf formatting errors. It would seem that the bot is not currently failing these, as they occasionally end up in master. If you're not using the hook, please be extra careful; this slows people down. If you want to set up the hook, you can find instructions under https://github.com/juju/juju/blob/master/CONTRIBUTING.md#local-clone. Thanks, Andrew -- Juju-dev mailing list Juju-dev@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev