Stable / unstable buildbots
Hello everyone, What is the rationale behind the distinction between "stable" and "unstable" buildbots? I ask that because the OpenBSD buildbot has failed compiling 3.0 for quite some time, but since that buildbot was in the "unstable" bunch, it was not discovered until someone filed a bug report for it (see http://bugs.python.org/issue3696). Regards Antoine.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Aug 28, 2008, at 6:28 AM, Antoine Pitrou wrote:
What is the rationale behind the distinction between "stable" and "unstable" buildbots? I ask that because the OpenBSD buildbot has failed compiling 3.0 for quite some time, but since that buildbot was in the "unstable" bunch, it was not discovered until someone filed a bug report for it (see http://bugs.python.org/issue3696) .
Mostly, it was a determination that Neal and I made at Pycon, about which bots we should "trust" to judge the health of the trees. I don't think the current list needs to be set in stone, and in fact several of the "stable" bots have had simple svn or other non-tree related problems for a while. As for the "unstable" ones, I personally rarely check them. - -Barry -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (Darwin) iQCVAwUBSLaXKXEjvBPtnXfVAQJfqgQAiyoqqUNQWByKEalKmWpRyqlX56PJIzPB S5gqubUZOvUTjVvHgvUtePPmJwEntyE+WcWXrX2tRlmMiCwantjgYOuNJeqonuwE BVZRrIjU/mayM9gRiN9NhuGHfV/YiLKpte+DKzuYZwGW2TnaCM3ijWe2QuMxmgqZ MMAO1w+M1kE= =qIFo -----END PGP SIGNATURE-----
2008/8/28 Barry Warsaw <barry@python.org>:
bots we should "trust" to judge the health of the trees. I don't think the current list needs to be set in stone, and in fact several of the "stable" bots have had simple svn or other non-tree related problems for a while.
Maybe a good requisite to move a buildbot from unstable to stable is to find a champion for it. I mean, something that can test on that platform and cares enough about it to, or fix the issue himself/herself, or find who broke it and bother the responsible until it gets fixed. Regards, -- . Facundo Blog: http://www.taniquetil.com.ar/plog/ PyAr: http://www.python.org/ar/
Facundo Batista <facundobatista <at> gmail.com> writes:
Maybe a good requisite to move a buildbot from unstable to stable is to find a champion for it. I mean, something that can test on that platform and cares enough about it to, or fix the issue himself/herself, or find who broke it and bother the responsible until it gets fixed.
By that metric, I fear that the only remaining buildbots would be the Linux/Windows x86/x64 ones. I'm not sure anyone here, for example, cares really much about Sparc buildbots, apart from the fact that having red "stable" buildbots doesn't make Python look very good, and we try to avoid that. But then Python wouldn't be really cross-platform anymore, or just in a very theoretical way. Also, lacking stable buildbots means some chunks of the language or stdlib aren't even tested anymore. For example, the bigmem/bigaddrspace tests are completely broken in the py3k branch. Regards Antoine.
2008/8/28 Antoine Pitrou <solipsis@pitrou.net>:
By that metric, I fear that the only remaining buildbots would be the Linux/Windows x86/x64 ones. I'm not sure anyone here, for example, cares really
Note that I meant to "move from unstable to stable, starting from the actual state", not to "decide again which ones will be stables or not". I agree with you in all the other matters. -- . Facundo Blog: http://www.taniquetil.com.ar/plog/ PyAr: http://www.python.org/ar/
participants (3)
-
Antoine Pitrou -
Barry Warsaw -
Facundo Batista