#development
- a
average-vr-56795
10/17/2018, 6:28 PMAre we picturing python test runs executing as independent processes (so they parallelise), or globbed together (so they merge output)? - w
witty-crayon-22786
10/17/2018, 6:28 PM3-5 more mins - w
witty-crayon-22786
10/17/2018, 6:29 PMindependent process invokes - a
average-vr-56795
10/17/2018, 6:29 PMđ - w
witty-crayon-22786
10/17/2018, 6:34 PMthe independent process invokes is closest to our current
,--no-fast
environment--chroot - w
witty-crayon-22786
10/17/2018, 6:34 PMand is most cacheable - w
witty-crayon-22786
10/17/2018, 6:36 PM@average-vr-56795: https://github.com/pantsbuild/pants/pull/6644 - a
average-vr-56795
10/17/2018, 7:13 PMAaaaaargh python fmt and lint - w
witty-crayon-22786
10/17/2018, 7:14 PMheh... what now - a
average-vr-56795
10/17/2018, 7:16 PMCopy code
âŚ* Checking lint FAILURE: 3 Python Style issues found. You may try `./pants fmt <targets>`
âŚ$ ./pants fmt --changed-parent=masterCopy code
âŚ* Checking lint FAILURE: 3 Python Style issues found. You may try `./pants fmt <targets>` - a
average-vr-56795
10/17/2018, 7:16 PMSo now I run lint without
and find my errors-q - a
average-vr-56795
10/17/2018, 7:16 PMI would perhaps suggest that the script should run lint in such a way that it actually outputs whatâs wrong⌠- a
average-vr-56795
10/17/2018, 7:17 PMOr better yet, use a formatter which will auto-fix anything it complains about đ - w
witty-crayon-22786
10/17/2018, 7:17 PMah, yea. blargh. @strong-crayon-70379 opened a ticket about that - a
average-vr-56795
10/17/2018, 7:21 PM@witty-crayon-22786 Is https://github.com/pantsbuild/pants/pull/6646 about what youâre expecting? - w
witty-crayon-22786
10/17/2018, 7:22 PM@average-vr-56795:
actually exists...TestResult - w
witty-crayon-22786
10/17/2018, 7:23 PMas does the static code we should use - a
average-vr-56795
10/17/2018, 7:23 PMThe model is pretty different - w
witty-crayon-22786
10/17/2018, 7:23 PM=x - w
witty-crayon-22786
10/17/2018, 7:24 PMso the testrunner_task_mixin code isn't reusable at all? it looked extractable - a
average-vr-56795
10/17/2018, 7:25 PMIt does partitioning and gloms together all the tests, assuming that anything in a partition which isnât in a failed list succeeded - a
average-vr-56795
10/17/2018, 7:26 PMThe previous version also assumes that the testrunner output directly to the console when it was running, and doesnât track stdoutw- 2
- 3
- w
witty-crayon-22786
10/17/2018, 7:28 PMi'll take a look. i had something else in mind, but it might be the right approach. - a
average-vr-56795
10/17/2018, 7:29 PMThanks đ - a
average-vr-56795
10/17/2018, 7:30 PMThe âHow should a console_rule propagate a status codeâ is also interesting - a
average-vr-56795
10/17/2018, 7:30 PMAre we assuming that at most one console_rule will be run in an invocation? - w
witty-crayon-22786
10/17/2018, 7:30 PM@average-vr-56795: exceptions should work - w
witty-crayon-22786
10/17/2018, 7:31 PM...i suppose we could have them yield a return code. but we can currently run multiple goals sequentially. so we'd be in a "take the max return code" situation there - w
witty-crayon-22786
10/17/2018, 7:32 PMwhat would the usecase be there? - a
average-vr-56795
10/17/2018, 7:33 PMIf I run two independent console_rules, and one succeeds, and one fails, I want the whole invocation to fail