#general
- w
witty-crayon-22786
10/15/2015, 6:50 PMbecause it isn't used here - h
happy-kitchen-89482
10/15/2015, 6:50 PMyes, it's a ~placeholder for some more robust representation of a 'language' - h
happy-kitchen-89482
10/15/2015, 6:50 PMLet me poke at this again - w
witty-crayon-22786
10/15/2015, 6:50 PMk. will shipit without--source-root-patterns - h
happy-kitchen-89482
10/15/2015, 6:51 PMIt's possible I can make what you suggest work - h
happy-kitchen-89482
10/15/2015, 6:51 PMAs long as there's some way to say that
meanspy
and so onpython - w
witty-crayon-22786
10/15/2015, 6:51 PMsure. will ship either way without--source-root-patterns - h
happy-kitchen-89482
10/15/2015, 6:51 PMWhich means we can't really get away from enumerating languages somewhere - w
witty-crayon-22786
10/15/2015, 6:51 PM"As long as there's some way to say that
meanspy
and so on" -- yea, i think that would be a simple mappingpython - h
happy-kitchen-89482
10/15/2015, 6:52 PMIn the long run that would have to be something we read from plugins - w
witty-crayon-22786
10/15/2015, 6:52 PM@happy-kitchen-89482: you can avoid enumerating them if the majority of the time they map to themselves. - w
witty-crayon-22786
10/15/2015, 6:52 PMsure. - h
happy-kitchen-89482
10/15/2015, 6:52 PMbut that will have to wait until we have a "Language" as a first-class entity - w
witty-crayon-22786
10/15/2015, 6:52 PMbut i think the language "enumeration" will end up being the language impls. - h
happy-kitchen-89482
10/15/2015, 6:52 PMyes, definitely - h
happy-kitchen-89482
10/15/2015, 6:53 PMthis list of strings is a placeholder for something sensible - w
witty-crayon-22786
10/15/2015, 6:53 PMk, your call - c
cuddly-architect-27013
10/15/2015, 6:55 PMquestions about junit_run for anyone with insight: what is the purpose of trying to calculate test names from the targets? is that just to satisfy the runner’s requirement of tests to run? is there a downside of defaulting to all tests run and just passing through when a test is specified? - c
cuddly-architect-27013
10/15/2015, 6:56 PMseems like there’s a lot of complication in those mappings and i’m wondering if there may be a chance to simplify or if there’s some requirement i’m not clear on - w
witty-crayon-22786
10/15/2015, 6:58 PM@cuddly-architect-27013: what do you mean about test names? test class names? - w
witty-crayon-22786
10/15/2015, 6:58 PMto "run all tests", i believe the runner would need to scan the whole classpath, which would be very expensive. - c
cuddly-architect-27013
10/15/2015, 6:58 PMyeah, for example in def _calculate_tests_from_targets(self, targets): - c
cuddly-architect-27013
10/15/2015, 6:59 PMwe take a set of targets, and i think go through all files for that target, then convert the filenames to class names - c
cuddly-architect-27013
10/15/2015, 6:59 PMand i think this is ultimately what’s sent to the runner - c
cuddly-architect-27013
10/15/2015, 6:59 PMah, hmmm - w
witty-crayon-22786
10/15/2015, 7:00 PM@cuddly-architect-27013: yep. it gives the runner an explicit set of classes to check. - c
cuddly-architect-27013
10/15/2015, 7:01 PMi see, and that avoids junit having to scan for instances of @Test attribute i suppose - w
witty-crayon-22786
10/15/2015, 7:01 PMbut that's 15ish lines... is this causing problems elsewhere? - w
witty-crayon-22786
10/15/2015, 7:01 PM@cuddly-architect-27013: it does that, but using only the explicit set of classes - w
witty-crayon-22786
10/15/2015, 7:02 PMps, the runner is also in pants. changing the runner is a possibility.