Webrev:
https://cr.opensolaris.org/action/browse/pkg/bpytlik/parallel-test-v1
Bugs:
18636 a single test suite run should be parallelizable
18571 test suite uses obsolete unittest function makeSuite
18616 fakeroot_repo should die
I've been working on this off and on for a while and I think it's
finally reliable and cleans up after itself well. In my experience, the
time savings can be pretty substantial (see the table below).
The default output is a bit verbose at this point, but I found it was
what I wanted to see. If you prefer the dots output, you can use -q (for
quiet). The -v and -p output options are also supported. With all output
though, instead of getting feedback in real time as each test finishes,
you'll be informed when each test class passes. If this really bothers
people, I can look into adding back real time response when only a
single process is being used.
I believe I've covered all of the other options appropriately, including
coverage. Being certain about having the right use of coverage was tough
because running coverage twice over the entire test suite using the gate
doesn't seem to reliably produce the exact same results. That said, the
numbers when using multiple processes seems to be reasonable from my
inspection.
Timing comparisons (all times in seconds):
Parallelism API CLI
Gate 693 5294
1 696 4861
2 346 2377
4 207 1293
6 173 978
8 178 843
Please let me know what you think.
Thanks,
Brock
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
