#general
- w
witty-crayon-22786
07/30/2015, 9:23 PMbut we also have it installed as
rather thanjunit_tests
, which makes the lack of an implicit dep more conspicuousjava_tests - h
happy-kitchen-89482
07/30/2015, 9:23 PMhmm - h
happy-kitchen-89482
07/30/2015, 9:23 PMYes, we don't inject it - h
happy-kitchen-89482
07/30/2015, 9:24 PMwe manually add it everywhere - h
happy-kitchen-89482
07/30/2015, 9:24 PMand actually I'm OK with that - w
witty-crayon-22786
07/30/2015, 9:24 PMregarding "tests shouldn't know what logging impl is in use"... that's the reason why slf4j exists - e
enough-analyst-54434
07/30/2015, 9:24 PMis it :junit or 3rdparty:junit - h
happy-kitchen-89482
07/30/2015, 9:24 PMsince you are importing and using JUnit symbols explicitly - b
bored-art-40741
07/30/2015, 9:24 PM"manually" - w
witty-crayon-22786
07/30/2015, 9:24 PM"explicitly" - b
bored-art-40741
07/30/2015, 9:24 PMIt's 3rdparty:junit - w
witty-crayon-22786
07/30/2015, 9:24 PMyea - e
enough-analyst-54434
07/30/2015, 9:24 PMthen Tom / you should be able to tuck away the slf4j there with no shading - h
happy-kitchen-89482
07/30/2015, 9:24 PMall our test targets depend on 3rdparty:junit explicitly, as it turns out - b
bored-art-40741
07/30/2015, 9:24 PMAnd it's injected by buildgen, which is basically as manual as it gets at 4sq - h
happy-kitchen-89482
07/30/2015, 9:25 PM@bored-art-40741: it's injected in the normal way (buildgen matches the used symbols to its symbol maps) or do we do something special based on the target type? - b
bored-art-40741
07/30/2015, 9:25 PMThe former - h
happy-kitchen-89482
07/30/2015, 9:25 PMgreat - e
enough-analyst-54434
07/30/2015, 9:25 PMthats the issue then - e
enough-analyst-54434
07/30/2015, 9:25 PMprobably have a tool symbol map - h
happy-kitchen-89482
07/30/2015, 9:26 PMso yeah, I'll mention to Tom that he can add the slf4j dep as a dep of 3rdparty:junit - e
enough-analyst-54434
07/30/2015, 9:26 PMthis feels like a bug we conceptually solved with words at the pre-summit Square hack session @mammoth-autumn-2551 - h
happy-kitchen-89482
07/30/2015, 9:26 PMif that's the way we want to go - e
enough-analyst-54434
07/30/2015, 9:26 PMerm s/paul/patrick/ - b
bored-art-40741
07/30/2015, 9:26 PMAgreed, John - b
bored-art-40741
07/30/2015, 9:26 PMI think Eric's product refactor ties closely into this - b
bored-art-40741
07/30/2015, 9:26 PMTool resolution is still very specialized and it doesn't have to be - h
happy-kitchen-89482
07/30/2015, 9:28 PMso I guess making slf4j a compile-time dep isn't strictly the right thing to do. It would be nice to have separate concepts of compile-time and runtime deps - h
happy-kitchen-89482
07/30/2015, 9:28 PMBut for now it'll be fine - b
bored-art-40741
07/30/2015, 9:28 PMNote that we'll need to remove it from our global excludes, @happy-kitchen-89482, and that could be a problem