|
From: Carl L. <ce...@us...> - 2019-10-17 14:46:25
|
Philippe:
I have only seen the nightly Valgrind test results for MIPS since about
Sept 28 or to. I am guessing the various scripts to run the tests on
the other architectures is not running again. I am guessing the cron
jobs need to be restarted?
Carl Love
|
|
From: Philippe W. <phi...@sk...> - 2019-10-20 10:07:12
|
I see that we had only 2 changes since the 28 of September: * a change I did in callgrind (10 Oct) * a white space only change (19 Oct), done for testing. I see we had results produced based on the changes of 19 of Oct. Not clear what has happened the 10 of Oct. For the fact that we have every day results for mips: I suppose that the mips buildbot does not check if there was a change in the repository before launching the build. Normally, if the mips cronjob launches nightly/bin/nightly, we should only have results produced if there was a change pushed. Petar, any idea why we see mips builds everyday ? Thanks Philippe On Thu, 2019-10-17 at 07:46 -0700, Carl Love wrote: > Philippe: > > I have only seen the nightly Valgrind test results for MIPS since about > Sept 28 or to. I am guessing the various scripts to run the tests on > the other architectures is not running again. I am guessing the cron > jobs need to be restarted? > > Carl Love > > > > _______________________________________________ > Valgrind-developers mailing list > Val...@li... > https://lists.sourceforge.net/lists/listinfo/valgrind-developers |
|
From: Petar J. <mip...@gm...> - 2019-10-22 17:01:54
|
On Sun, Oct 20, 2019 at 12:07 PM Philippe Waroquiers <phi...@sk...> wrote: > > I see that we had only 2 changes since the 28 of September: > * a change I did in callgrind (10 Oct) > * a white space only change (19 Oct), done for testing. > > I see we had results produced based on the changes of 19 of Oct. > Not clear what has happened the 10 of Oct. > > For the fact that we have every day results for mips: > I suppose that the mips buildbot does not check if there > was a change in the repository before launching the build. > > Normally, if the mips cronjob launches nightly/bin/nightly, > we should only have results produced if there was a change pushed. > > Petar, any idea why we see mips builds everyday ? > The engineer responsible for Valgrind buildbot just extended the script to send the results after the build/test completion. I agree that it suboptimal, will ask him to change it. Petar |
|
From: John R. <jr...@bi...> - 2019-10-20 14:18:20
|
On 10/20/19 3:06 AM, Philippe Waroquiers wrote: <<snip>> > For the fact that we have every day results for mips: > I suppose that the mips buildbot does not check if there > was a change in the repository before launching the build. > > Normally, if the mips cronjob launches nightly/bin/nightly, > we should only have results produced if there was a change pushed. <<snip>> It would be good for testing to run weekly even if no change was pushed. Changes in the "external environment" (compilers, build utilities, operating systems) may cause bad results, and such problems should be detected despite no change in the repository. |
|
From: Petar J. <mip...@gm...> - 2019-10-22 17:10:23
|
> > It would be good for testing to run weekly even if no change was pushed. > Changes in the "external environment" (compilers, build utilities, > operating systems) may cause bad results, and such problems > should be detected despite no change in the repository. > > Btw, executing the whole test suite even if there are no new changes over the years have discovered flakiness of some of the tests. So, there is some value in it. Having said this, the mailing list should not be spammed for that reason though. Petar |