Me too, but I'm okay with killing the sdist targets as an intermediate step and reintroducing them with a non-pants build
On Friday, July 24, 2015, Brian Wickman <wick...@apache.org> wrote: > > > > On July 24, 2015, 5:57 p.m., Brian Wickman wrote: > > > src/main/python/apache/thermos/observer/BUILD, line 74 > > > < > https://reviews.apache.org/r/36700/diff/2/?file=1020365#file1020365line74> > > > > > > This puts an aurora dependency on thermos and I imagine causes a > cycle in the build graph (pants may just silently ignore it.) > > > > > > Let's just remove the thermos_observer entry point from this > package entirely by deleting the .with_binaries clause. > > > > Bill Farner wrote: > > This target (`src/main/python/apache/thermos/observer`) is a > top-level target, only used by `build-support/release/make-python-sdists`, > so i think we're safe from a cycle. I'm not strictly against removing the > `with_binaries`, but i assume it is there for a reason and something would > break? Or do we just not use the sdists [this way]? > > > > Brian Wickman wrote: > > Afaik nobody uses these in practice. They're only useful for when > you install the sdists via pip or pex directly. Instead we always build > pex binaries using pants which delegates to the python_binary target. > > > > Kevin Sweeney wrote: > > I answered someone's question in IRC with a gist ~2yrs ago > https://gist.github.com/kevints/8361414 so it does seem likely these are > used in practice. Okay with removing them but a NEWS entry would be good > form. > > > > Bill Farner wrote: > > Should i just kill the sdists entirely? > > I'd prefer to go in the opposite direction and kill pants. > > > - Brian > > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/36700/#review92944 > ----------------------------------------------------------- > > > On July 24, 2015, midnight, Bill Farner wrote: > > > > ----------------------------------------------------------- > > This is an automatically generated e-mail. To reply, visit: > > https://reviews.apache.org/r/36700/ > > ----------------------------------------------------------- > > > > (Updated July 24, 2015, midnight) > > > > > > Review request for Aurora, Kevin Sweeney and Brian Wickman. > > > > > > Bugs: AURORA-1381 > > https://issues.apache.org/jira/browse/AURORA-1381 > > > > > > Repository: aurora > > > > > > Description > > ------- > > > > The meaning of the targets under > `src/main/python/apache/thermos/cli/bin/BUILD` and > `src/main/python/apache/thermos/observer/bin/BUILD` changed recently, which > bit several people (including our own packaging). This removes targets > that Aurora users should not be consuming. > > > > > > Diffs > > ----- > > > > build-support/packaging/debian/rules > 23828c02b73f007393d2ed4ce69c010cebf07e57 > > build-support/packaging/rpm/aurora.spec > 0c4c106a1730ee7e4e62c7ba778e7f0e2cb44771 > > src/main/python/apache/thermos/BUILD > 8221aa0bd4efe5f519550cba716d6a564ba9ae44 > > src/main/python/apache/thermos/cli/bin/BUILD > f33c7f838a6a0932abc737d0ecf7ca3a59580e2e > > src/main/python/apache/thermos/cli/bin/thermos.py > fcef38d33e7ca2005f35c3bdb90ffee6aeade3af > > src/main/python/apache/thermos/observer/BUILD > 41db2cc2e11234c434f58f55abec0b9f308096be > > src/main/python/apache/thermos/observer/bin/BUILD > 0abe2ccfe9c5ccb509ad731125385900114ba352 > > src/main/python/apache/thermos/observer/bin/thermos_observer.py > 39d3994a6163746e853cd21fc4c3585edc2b54cb > > > > Diff: https://reviews.apache.org/r/36700/diff/ > > > > > > Testing > > ------- > > > > ./build-support/jenkins/build.sh > > > > ./src/test/sh/org/apache/aurora/e2e/test_end_to_end.sh > > (as far as it currently gets on master, anyhow, due to AURORA-1409) > > > > > > Thanks, > > > > Bill Farner > > > > > > -- Sent from Gmail Mobile
