[Python] Update list of Python files without .py suffix in .pep8 file

[practicalswift]

Changes:

• Add ld (test/Driver/Inputs/filelists/ld)
• Remove pre-commit-benchmark (deleted)

[jrose-apple]

:-/ It seems wrong to have to update this every time we need a new helper tool for tests. Is there a way to get it to detect files from their #! line? Maybe "files in test/* and utils/* with no extension with a first line matching #!.+python"?

[practicalswift]

@jrose-apple Yes, I agree that is problematic.

Unfortunately the .pep8 format is quite limited so I'm afraid the filename directive is the only inclusion mechanism available.

The way I'm finding Python files without a .py suffix is almost exactly what you suggested :-) This is what I'm using in my local "check-Swift-repo-consistency"-routine:

$ git grep '#!.*python' | cut -f1 -d: | sort | egrep -v '.py$' | sort
docs/scripts/ns-html2rst
test/Driver/Inputs/filelists/ld
utils/build-script
utils/gyb
utils/line-directive
utils/recursive-lipo
utils/submit-benchmark-results
utils/update-checkout
utils/viewcfg

Turns out test/Driver/Inputs/filelists/ld is the only Python file under test/ that is missing a .py extension.

Possible solutions:

• Should we simply skip linting Python files without .py suffix under test/?
• If not, should we standardize on having a .py suffix for all Python files under test/?

[jrose-apple]

Well, the reason ld doesn't is because it's acting like the real ld, but I suppose we could change the test so that it symlinks it into a temporary directory (ln -s %S/Inputs/filelists/ld.py %t/ld) rather than just pointing at the existing one.

[practicalswift]

@jrose-apple Yes, ld is a special case. The symlink solution sounds smart. Please let me know if you want me to make any changes to this PR to make it merge:able :-)

[jrose-apple]

Okay, I moved ld to fake-ld.py in 839dfc0. Everything else seems reasonable.

[practicalswift]

@jrose-apple PR updated accordingly. Let me know if any further changes are needed :-)

[jrose-apple]

Ah, now there are conflicts because of your other change. Can you rebase and try again?

[practicalswift]

@jrose-apple Rebased :-)

[jrose-apple]

I hate to keep bugging you but you snuck the "E114" change into this commit. Was that intentional?

[practicalswift]

@jrose-apple Good you noticed that :-) That change belongs in #1292, so removed it from this PR. Also added the remaining missing files. Let me know if any changes are needed :-)

[practicalswift]

Rebased again :-)

[jrose-apple]

…are the .in files valid Python? Valid enough to check with pep8? I mean, it's too bad that we wouldn't get any checking otherwise, but…

[practicalswift]

@jrose-apple Yes, they are all passing flake8. They are 100 % Python with "@PATH_TO_DRIVER_LIBRARY@" strings as the only magic :-)

$ flake8 ./benchmark/scripts/Benchmark_DTrace.in ./benchmark/scripts/Benchmark_GuardMalloc.in ./benchmark/scripts/Benchmark_RuntimeLeaksRunner.in
$ head -1 ./benchmark/scripts/Benchmark_DTrace.in ./benchmark/scripts/Benchmark_GuardMalloc.in ./benchmark/scripts/Benchmark_RuntimeLeaksRunner.in
==> ./benchmark/scripts/Benchmark_DTrace.in <==
#!/usr/bin/env python

==> ./benchmark/scripts/Benchmark_GuardMalloc.in <==
#!/usr/bin/env python

==> ./benchmark/scripts/Benchmark_RuntimeLeaksRunner.in <==
#!/usr/bin/env python
$ egrep '@.*@' ./benchmark/scripts/Benchmark_DTrace.in ./benchmark/scripts/Benchmark_GuardMalloc.in ./benchmark/scripts/Benchmark_RuntimeLeaksRunner.in
./benchmark/scripts/Benchmark_DTrace.in:DRIVER_LIBRARY_PATH = "@PATH_TO_DRIVER_LIBRARY@"
./benchmark/scripts/Benchmark_GuardMalloc.in:sys.path.append("@PATH_TO_DRIVER_LIBRARY@")
./benchmark/scripts/Benchmark_RuntimeLeaksRunner.in:sys.path.append("@PATH_TO_DRIVER_LIBRARY@")

[shahmishal]

@swift-ci Please test

[practicalswift]

@shahmishal Build seems to have passed 👍 Let me know if any further changes are needed :-)

[jrose-apple]

Thanks for sticking this one out. :-)