On Wed, Sep 09, 2015 at 03:57:55PM -0400, Tom Stellard wrote: > On Wed, Sep 09, 2015 at 12:52:22PM -0700, Dylan Baker wrote: > > > I have a buildbot seti up to run piglit and the way it detects whether > > > or no there have been piglit regressions is by running piglit summary > > > console and then parsing the output. > > > > > > The CSV summary seems to only accept a single result file, unless I'm > > > doing something wrong: > > > > > > tstellar@localhost ~/piglit $ ./piglit summary csv -o out.csv > > > results/master--150727-/ results/master-BONAIRE-150223-/ > > > usage: piglit [-h] [-o <Output File>] <Input Files> > > > piglit: error: unrecognized arguments: results/master-BONAIRE-150223-/ > > > > > > -Tom > > > > Are you just wanting to see if there are regressions, or do you want to > > get a count and the tests that regressed as well? > > What I want is a list of changes, and yes/no for regressions. The > number of regressions does not matter. `piglit/piglit summary console > -d` gives me a list of changes and prints the summary so I can check > if there were regressions. > > Another solution I considered was to add an exit status to piglit > summary to indicate whether or not there were regressions. This would > be the best solution, but I wasn't sure if this would break other > people's use cases. > > -Tom
I think with my changes it would be trivial to add a '-r/--regressions' mode to piglit summary that would print only regressions in the change list, and could have an exit status for whether there were regressions or not. I'll add an extra patch to my series. Dylan
signature.asc
Description: Digital signature
_______________________________________________ Piglit mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/piglit
