<https://github.com/pantsbuild/pants/issues/7444> ...
# development
a
https://github.com/pantsbuild/pants/issues/7444 this ticket has been dormant for a while. is this doc intended to resolve it?
h
Sort of, I suppose. We already have 6 V2 Python linters now.
python-eval
is going through deprecation cycles to be removed and
python-checkstyle
will be superceded by the Flake8 implementation +
flake8-pantsbuild
plugin The only ones Toolchain has zero plans to add is the thrift linters -- But, yes, generally, we realized that we aren’t comfortable having our users use the V2 linters yet until we fix the performance issue. The first experience of
./pants fmt ::
should be magical, and instead it’s sitting waiting for 2 minutes for something that previously took 30 seconds. (Unless you have remote caching enabled - then it will likely be magical)
a
The only ones Toolchain has zero plans to add is the thrift linters
thrift gen is the only blocker for twitter adopting the v2 python backend internally, so i'm likely going to trickle this out over the next month
🎉 1
h
#7020 looks related and solving the same problem, but for V1
python-checkstyle
instead of the V2 linters. Good example of why we think allowing this alternative caching scheme is warranted
thrift gen is the only blocker for twitter adopting the v2 python backend internally
Note that you can start adopting the linters now without adopting the rest of the Python backend! We made a change so that you can use
pants.backend.python.lint.*
without needing to register
pants.backend.python
a
#7020 was written in v2 at one point, then reverted