#general
- w
witty-crayon-22786
07/29/2015, 9:28 PMok - w
witty-crayon-22786
07/29/2015, 9:28 PManywho, sorry. my ignorance is talking. - w
witty-crayon-22786
07/29/2015, 9:28 PMthanks a lot for the help - e
enough-analyst-54434
07/29/2015, 9:28 PMI am bothered by this ever having worked before - w
witty-crayon-22786
07/29/2015, 9:28 PMi will move this package - w
witty-crayon-22786
07/29/2015, 9:30 PMthere seem to be a fair number of nested thrift namespaces. - w
witty-crayon-22786
07/29/2015, 9:31 PMbeing in the same namespace but different targets will still trigger the issue, correct? - g
glamorous-carpenter-79
07/29/2015, 9:34 PM@enough-analyst-54434: Thank you! https://dl.bintray.com/pantsbuild/bin/build-support/bin/node/ - e
enough-analyst-54434
07/29/2015, 9:34 PMyw - e
enough-analyst-54434
07/29/2015, 9:35 PM@witty-crayon-22786: specifically only when one namespace is a strict prefix of another and these are in different targets - e
enough-analyst-54434
07/29/2015, 9:35 PMIf the rule of only code in leaf packages was followed - all good - w
witty-crayon-22786
07/29/2015, 9:36 PMok. so multiple targets for the same package is fine, because the `__init__.py`s happen to line up? - e
enough-analyst-54434
07/29/2015, 9:37 PM - e
enough-analyst-54434
07/29/2015, 9:37 PMBasically, namespace doctoring is done, but can;t trample init.py with stuff in it - e
enough-analyst-54434
07/29/2015, 9:39 PMIn order to add a namespace package to a non-empty init.py pants would need to collect global data on all packages in play 1st and union the contents of the init.pys then add the namespace to that - e
enough-analyst-54434
07/29/2015, 9:39 PMAnd typing that - I'd never do that - too much that could be wrong - e
enough-analyst-54434
07/29/2015, 9:39 PMbetter to use the global knowledge to fail the build - w
witty-crayon-22786
07/29/2015, 9:40 PMand the reason we can't stomp init.py is because the code might have sideeffects? - w
witty-crayon-22786
07/29/2015, 9:40 PMstomp/disallow - w
witty-crayon-22786
07/29/2015, 9:41 PM(and thrift generates them, but we know that) - e
enough-analyst-54434
07/29/2015, 9:41 PMBut that's not a but, its just a fact of life external code wil use non-empty init.py - e
enough-analyst-54434
07/29/2015, 9:41 PMPants has to allow for libraries to do this - e
enough-analyst-54434
07/29/2015, 9:41 PMand if it does, it must handle the merge case - w
witty-crayon-22786
07/29/2015, 9:42 PMsure... don't necessarily have to allow for package collisions between internal and 3rdparty though... - e
enough-analyst-54434
07/29/2015, 9:42 PMYou really do - e
enough-analyst-54434
07/29/2015, 9:42 PMYou're assuming a monorepo - w
witty-crayon-22786
07/29/2015, 9:42 PMheh, aw =( - w
witty-crayon-22786
07/29/2015, 9:42 PMyea, i am - e
enough-analyst-54434
07/29/2015, 9:42 PMYup - pants can encourage one, but really can;t require one imo - w
witty-crayon-22786
07/29/2015, 9:42 PMhm.