#general
- e
enough-analyst-54434
08/15/2015, 1:32 AM😕 - w
witty-crayon-22786
08/15/2015, 1:33 AMah - e
enough-analyst-54434
08/15/2015, 1:34 AMpita - e
enough-analyst-54434
08/15/2015, 1:34 AMNo good way I can see around this except being very careful that global options have namespaces by hand in their names - w
witty-crayon-22786
08/15/2015, 2:03 AMwell, we wanted to kill DEFAULT, right? - e
enough-analyst-54434
08/15/2015, 2:12 AMYeah - that will be the best way - e
enough-analyst-54434
08/15/2015, 2:13 AMNot kill it, but not use it - leave it for users to use if they want to use ConfigParser's DEFAULT inheritance feature - e
enough-analyst-54434
08/15/2015, 5:21 AMAnd Benjy filed exactly what you and I said already: https://github.com/pantsbuild/pants/issues/1803 - h
happy-kitchen-89482
08/16/2015, 1:13 PMI'm looking at making some test.junit options advanced, and there are some coverage-related ones we might be able to get rid of - h
happy-kitchen-89482
08/16/2015, 1:14 PMspecifically, can we kill
,--coverage-console
,--coverage-xml
and just always emit all three?--coverage-html - h
happy-kitchen-89482
08/16/2015, 1:14 PMI can't imagine that's a performance hit compared to the time it takes to run the tests in the first place - h
happy-kitchen-89482
08/16/2015, 2:37 PMI'll push this under the assumption that it's OK (and with jsirois' shipit), can always change/revert if needed. - a
ancient-stone-50795
08/16/2015, 2:39 PM@happy-kitchen-89482: That sounds fine to me. - h
happy-kitchen-89482
08/16/2015, 3:35 PM@enough-analyst-54434: how far out is your change to regularize optionable deps? - e
enough-analyst-54434
08/16/2015, 3:45 PMThat's on my list today. I need to clear my deck for some integration work in the Medium mono repo Monday am. - h
happy-kitchen-89482
08/16/2015, 3:53 PMcool, then I will hold off on my change to nest subsystem help under the depending task's help - h
happy-kitchen-89482
08/16/2015, 3:54 PMbecause that needs to be able to iterate over an optionable's dependencies, and I think your change will provide a generic way to do that without having to special-case tasks and subsystems separately? - h
happy-kitchen-89482
08/16/2015, 3:54 PMI need to be able to say, for any optionable, "give me all the optionables scoped to you" - h
happy-kitchen-89482
08/16/2015, 3:55 PMe.g., a task will give all of what we currently call "per-task subsystems" it uses (but not the global ones) - h
happy-kitchen-89482
08/16/2015, 3:55 PMIs that going to fall out of your change? - e
enough-analyst-54434
08/16/2015, 3:59 PMYes - e
enough-analyst-54434
08/16/2015, 3:59 PMUm, no on the task bit - e
enough-analyst-54434
08/16/2015, 4:00 PMFor any task or subsystem, with 1 call you get its dep graph - period - e
enough-analyst-54434
08/16/2015, 4:00 PMYou could filter out deps that are global though - e
enough-analyst-54434
08/16/2015, 4:06 PMI'll build in a query method - is_global or is_original on Task/Subsystem or Optionable - h
happy-kitchen-89482
08/16/2015, 4:49 PMthat's fine - h
happy-kitchen-89482
08/16/2015, 4:49 PMnote that "dep graph" is misleading, because of the scoping - h
happy-kitchen-89482
08/16/2015, 4:50 PMif two different tasks have a "per-task subsystem" then each get their own independent instance of that subsystem, and so on transitively - h
happy-kitchen-89482
08/16/2015, 4:50 PMSo, for example, we would never need to topo-sort this "dep graph" - e
enough-analyst-54434
08/16/2015, 4:59 PMNo - its definitely a graph