-
-
Notifications
You must be signed in to change notification settings - Fork 31.7k
bpo-43372: Re-generate frozen code for __hello__. #24708
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
The Check if generated files are up to date CI appears to run regen-all with an older Python version. |
That is Python 3.9.1 |
See nascheme#2 -- I think that should do. |
Doesn't seem to work, unfortunately:
That CI test is broken, IMHO. I wonder is we should spilt fixing the CI into a separate PR. Sort of a chicken and egg problem I guess since each PR will fail unless both applied. |
My attempt fails with:
|
Another option might be to disable those CI tests until we can fix them. They were added in this change, it seems:
Using some different version of Python to regen stuff does not seem safe to me. It just happened to mostly work until now. Using the freshly build Python (i.e. |
@hroncok it seems the issue is the
I'm going to try updating my PR and hope the github CI rules work. |
It passed. Thanks! |
It seems all is working now. The failure in "Tests / Ubuntu" seems unrelated to these changes (SSL test failed). Something I noticed: it seems inconsistent how |
The marshal format for code objects has changed with bpo-42246, commit 877df85. Update expected code sizes in ctypes test_frozentable.
bb273c2
to
ad42c73
Compare
Marking as "DO-NOT-MERGE". I think my other PR is a cleaner fix. Assuming that one is okay, we can close this one. |
Closing since this is not the best approach. #24759 seems better. |
It's not a bug, but it's made on purpose. For example, if you use Python 3.8 as PYTHON_FOR_REGEN, using PYTHON_FOR_REGEN to build the "stdlib module names" list would be wrong (Python 3.8 has different/less stdlib modules). You really must use the just built Python to build this list. Do you think that a comment is required in Makefile.pre.in to prevent future mistakes? |
As you noticed, Travis CI and the GitHub Actions use two steps. Example from .travis.yml:
At least, running "make regen-all" displays this message:
|
The marshaled data for code objects has changed with bpo-42246, commit
877df85. Update expected code sizes in ctypes test_frozentable.
https://bugs.python.org/issue43372