On Fri, Dec 19, 2014 at 11:34 AM, Gijs Kruitbosch <gijskruitbo...@gmail.com> wrote: > > On 19/12/2014 14:56, Ehsan Akhgari wrote: > >> Logging sufficiently is >> almost always enough to not have to use these timers, as those tests >> have demonstrated in practice. >> > > Who's working on improving the log output from tinderbox?
I'm not sure. > Timestamps get mashed [0], File a bug, I guess? > sometimes only the failing assertion is printed with no other logs at all > [1], This was done by Nathan Froyd, with the goal of reducing the log file size. This annoys me as well, but there is a remedy: SimpleTest.requestCompleteLog(). > ... it's not really such a happy world as you assert. > Not sure if you're complaining about the state of things or if you're suggesting that the usage of these types of timeouts is better. I hope that the discussion in this thread has proved the latter false, at least. If you have specific suggestions on how to make the logging story better, please file bugs, but I think holding back changes such as blacklisting well-known intermittent-failure prone until we get our logging story straight is not very reasonable. For one thing, requestCompleteLog() is *much* easier to use than any of these magical timeouts, if you need it. Cheers, Ehsan > ~ Gijs > > [0] https://bugzilla.mozilla.org/show_bug.cgi?id=1049545#c0 > [1] https://treeherder.mozilla.org/ui/logviewer.html#?repo= > fx-team&job_id=1494921 > -- Ehsan _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
