-
-
Notifications
You must be signed in to change notification settings - Fork 10.8k
BLD: updates to the Meson build #23740
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 CI failure is in the meson build on macos_arm64. |
Hmm, I tested that locally, but in a conda env. Shouldn't be hard to figure out. The error was:
also something wrong with CMake in Homebrew it looks like. |
It's good that
I think we want to keep this unchanged, meaning that if no BLAS/LAPACK library was found, we use the fallback library and continue. We should at least emit a clear warning though, because for many users this continues to be a performance footgun. |
Do you need help debugging this? The wheel builder job seems to be workingThe numpy-1.25.0.dev0+1377.g276cc995c-cp39-cp39-macosx_11_0_arm64.whl built 4 days ago and uploaded to https://anaconda.org/scipy-wheels-nightly/numpy/files does have a 22 MB file named |
Not really, this is pretty routine by now, the problem should be that the I am bandwidth-constrained though and may not touch this for several days, so if you want to fix it, I won't complain:) |
close/reopen |
meson.build
Outdated
@@ -34,10 +35,11 @@ elif cc.get_id() == 'msvc' | |||
'when building with MSVC') | |||
endif | |||
endif | |||
if not cy.version().version_compare('>=0.29.34') | |||
error('SciPy requires Cython >= 0.29.34') |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is there a "project_name" macro we could use here instead?
error('SciPy requires Cython >= 0.29.34') | |
error('NumPy requires Cython >= 0.29.34') |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good catch, thanks. I'll see if the current CI run passes, then will push this with a few more tweaks.
Is there a "project_name" macro we could use here instead?
Not that I know of - but seems more complexity than it's worth anyway, and makes it harder to grep for the error string, so I'd prefer not to do that.
It looks like we should start moving some more CI jobs over to Meson soon. Also, the automatic wheel build trigger is a massive waste of jobs. This looks better now, I'll keep it like this and will add a comment that OpenBLAS is not used:
EDIT and the warning I wanted to see:
|
Mostly synced from SciPy, and also addresses some warnings that show up with Meson 1.1.0 [skip circle] [skip azp]
@andyfaff FYI I make some minor tweaks to |
A fun new error with
|
I think I saw it before at some point and I don't think the compiler is right, just guessing on the "this may happen" side of things. Anyway, initializing to silence the warning is OK. |
Indeed, the code wasn't broken but the compiler cannot know. I think using an uninitialized variable that is then used only within an if-statement controlled by some other not-directly-the-same-pointer-condition should always be avoided. |
This is good to go now. |
Thanks @rgommers |
Mostly synced from SciPy, and also addresses some warnings that show up with Meson 1.1.0