Rename antsibull project to antsibull-build #632
Conversation
This project can be published to PyPI to maintain backwards compatibility.
felixfontein
force-pushed
the
antsibull-stub
branch
4 times, most recently
from
07ab88e to
f391773
Compare
felixfontein
force-pushed
the
antsibull-stub
branch
from
f391773 to
123fbcf
Compare
|
@gotmax23 I rebased your commits and added two more to rename the Python module accordingly, and to bump the stub's version to 0.67.0 since antsibull 0.66.0 is already there. Does this look good to you? I'd like to get this merged, the repo renamed, and antsibull-build 0.66.0 released soon so I can start working on antsibull-build again (ansible-community/ansible-build-data#478 (comment)). |
|
|
||
| The filename can be a relative path separated with '/' to access subdirectories. | ||
| See https://docs.python.org/3/library/pkgutil.html#pkgutil.get_data for details. | ||
| """ | ||
| data = pkgutil.get_data("antsibull.data", filename) | ||
| data = pkgutil.get_data("antsibull_build.data", filename) |
There was a problem hiding this comment.
Note to self: we should switch this to use importlib.resources.
There was a problem hiding this comment.
Adding to your note: similar code also exists in antsibull-docs :)
| env: | ||
| # Since antsibull was renamed to antsibull-build, paths during the bytecompile test | ||
| # are too long and bytecompile fails with "OSError: AF_UNIX path too long". | ||
| COVERAGE_RELEASE_FORCE_TMPDIR: /tmp |
There was a problem hiding this comment.
Google returns no results for COVERAGE_RELEASE_FORCE_TMPDIR. How does it change pycoverage's behavior?
There was a problem hiding this comment.
It doesn't. This is interpreted by noxfile.py.
Co-authored-by: Maxwell G <[email protected]>
noxfile.py
Outdated
| if os.environ.get("COVERAGE_RELEASE_FORCE_TMPDIR"): | ||
| tmp_env["TMPDIR"] = os.environ["COVERAGE_RELEASE_FORCE_TMPDIR"] |
There was a problem hiding this comment.
This seems redundant if COVERAGE_RELEASE_FORCE_TMPDIR=/tmp that's set in CI is already the default. Also, probably better not to use the COVERAGE_ namespace if the coverage program is not actually reading this envvar.
Co-authored-by: Maxwell G <[email protected]>
|
Thank you! |
|
@gotmax23 thanks for suggesting and preparing this! |
Rebased version of #629.
Closes #629.
Fixes #627.