Hi Andreas, I like the idea to always print a summary. This is the information that's most important. The current (old?) way of dumping the stats differences was hard for a human to parse anyway. Adding a note in the scons output as to how to see more information whenever a test doesn't pass would be good too.
Cheers, Jason On Tue, May 31, 2016 at 11:29 AM Andreas Sandberg <[email protected]> wrote: > Hi Everyone, > > Andreas (the other one) just pointed out that the regress script used to > print a summary of the test results the second time you run it. As of > earlier today, this is no longer the case. > > Instead of printing a summary, it now prints the same results as the first > time the test was run. That is, you¹ll get both the passed/failed/changed > status /and/ an output diff if the test failed. Every time. > > Andreas pointed out that this is not the ideal behaviour since it makes it > hard to get an overview of the tests. These are some of the solutions I > can think of at them moment: > > * Always print a summary and rely on the user querying the pickled > results for more information. E.g., by running: tests.py show --format > summary `find build/ -name status.pickle` > > * Same behaviour as today, but require users to create the summary > themselves. E.g., by running: tests.py show --format summary `find build/ > -name status.pickle` > > * Same behaviour as today, but hack regress.py to print the summary after > all tests have been run. > > In terms of implementation effort, the first option is definitely easiest. > We might be able to make scons print a warning/error that suggests which > command to run to query the test state if one more more tests fail. > > Cheers, > Andreas > > IMPORTANT NOTICE: The contents of this email and any attachments are > confidential and may also be privileged. If you are not the intended > recipient, please notify the sender immediately and do not disclose the > contents to any other person, use it for any purpose, or store or copy the > information in any medium. Thank you. > > _______________________________________________ > gem5-dev mailing list > [email protected] > http://m5sim.org/mailman/listinfo/gem5-dev > -- Jason _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
