On 14 November 2014 09:26, Dylan Baker <[email protected]> wrote: > This restores the ability to create a summary from a partial set of > runs by using the same functionality as resume. > > Signed-off-by: Dylan Baker <[email protected]> > cc: [email protected]
Tested-by: Dave Airlie <[email protected]> wfm, please push! > --- > framework/results.py | 13 +++++++++++++ > 1 file changed, 13 insertions(+) > > diff --git a/framework/results.py b/framework/results.py > index 175a386..6ab22b1 100644 > --- a/framework/results.py > +++ b/framework/results.py > @@ -190,6 +190,19 @@ def load_results(filename): > # file descriptors > if not os.path.isdir(filename): > filepath = filename > + elif os.path.exists(os.path.join(filename, 'metadata.json')): > + # If the test is still running we need to use the resume code, since > + # there will not be a results.json file. > + # We want to return here since the results are known current (there's > + # an assert in TestrunResult.load), and there is no filepath > + # to pass to update_results > + # XXX: This needs to be run before searching for a results.json file > so > + # that if the new run is overwriting an old one we load the > + # partial and not the original. It might be better to just > delete > + # the contents of the folder if there is anything in it. > + # XXX: What happens if the tests folder gets deleted in the middle of > + # this? > + return TestrunResult.resume(filename) > else: > # If there are both old and new results in a directory pick the new > # ones first > -- > 2.1.3 > _______________________________________________ Piglit mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/piglit
