[08:20] <mchehab> @all, FYI, there's an small issue with the Jenkins task that is doing media-next builds... [08:20] <mchehab> when a fail happens, it is emitting 2 e-emails, instead of one [08:21] <mchehab> a "failed" one: https://www.linuxtv.org/cgi-bin/mailman/private/media-submaintainers/2021-May/000394.html [08:21] <mchehab> and after a couple of minutes: [08:21] <mchehab> a "back to normal" one: https://www.linuxtv.org/cgi-bin/mailman/private/media-submaintainers/2021-May/000395.html [08:22] <mchehab> probably, the back to normal one is broken [08:24] <mchehab> fixing it is not too trivial [08:24] <mchehab> the problem is that the jenkins job only do a real build when the changesets are different. If they're identical, it just exits with 0 as return code [08:25] <mchehab> (and an environment var that tells Jenkins to not keep the resuts) [08:26] <mchehab> I'll likely need to do something similar to the "patchwork" bulds, where the e-mail notification is done by the script, and not by Jenkins [08:26] <mchehab> Or to make the script to return the past build result when something bad happens [08:50] <mchehab> Fixed. It should now properly report the failed results to Jenkins [08:51] <mchehab> so, the "back to normal" email should only be sent when the media next tree can be built with success [08:55] <mchehab> the script still has one issue, though... if the tree fails to build, it is keeping some badly merged stuff, causing all subsequent PRs to also fail [08:55] <mchehab> so: [08:56] <mchehab> diff --cc fail_trees [08:56] <mchehab> index 000000000000,000000000000..b4bc2bc3cefb [08:56] <mchehab> new file mode 100644 [08:56] <mchehab> --- /dev/null [08:56] <mchehab> +++ b/fail_trees [08:56] <mchehab> @@@ -1,0 -1,0 +1,5 @@@ [08:56] <mchehab> ++git://linuxtv.org/hverkuil/media_tree.git tags/br-v5.14a [08:56] <mchehab> ++git://linuxtv.org/hverkuil/media_tree.git tags/br-v5.14-out1 [08:56] <mchehab> ++git://linuxtv.org/hverkuil/media_tree.git tags/br-v5.14b [08:56] <mchehab> ++git://linuxtv.org/hverkuil/media_tree.git tags/br-v5.13d [08:56] <mchehab> ++git://linuxtv.org/hverkuil/media_tree.git tags/br-v5.14c [08:56] <mchehab> diff --cc success_trees [08:56] <mchehab> index 000000000000,000000000000..6c4fbe627604 [08:56] <mchehab> new file mode 100644 [08:56] <mchehab> --- /dev/null [08:56] <mchehab> +++ b/success_trees [08:56] <mchehab> @@@ -1,0 -1,0 +1,4 @@@ [08:56] <mchehab> ++git://linuxtv.org/media_tree.git fixes [08:56] <mchehab> ++git://linuxtv.org/media_stage.git master [08:56] <mchehab> ++git://linuxtv.org/mchehab/experimental.git next [08:56] <mchehab> ++git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media tags/media/v5.13-1 [08:56] <mchehab> (I suspect that just hverkuil br-v5.14a PR failed above) [09:08] <pinchartl> I've seen that indeed, was wondering what it was about :-)