Add sage-generate-meson script to sage-setup
#37008
Conversation
|
As it has been explained before, the point of this script was to create the initial boilerplate code. I see value in creating a script that adds missing python and cython files. But that would require to parse the existing meson files instead of overwriting them. Something similar is already supported for other targets, just not for python. https://mesonbuild.com/Rewriter.html So I am sure the meson team would be happy to accept your PR. |
It wouldn't. Even just overwriting + then editing often has value. |
|
With the |
Yes, the idea is to separate handwritten bits from autogenerated bits. |
|
Documentation preview for this PR (built with commit ec43bc5; changes) is ready! 🎉 |
|
I think we now have pressing needs to move to meson. This stuff looks OK, but it cannot be used in isolation. Can it even be tested without other bits for sagemath-standard? |
From PR 'Compile everything with meson' (sagemath#36524 at 46bc6b4 (2023-12-31)
…ything with meson (sagemath#36524 at 46bc6b4 (2023-12-31) squashed on 10.3.beta4
….build files from PR Compile everything with meson (sagemath#36524 at 46bc6b4 (2023-12-31) squashed on 10.3.beta4
…son.build files from PR Compile everything with meson (sagemath#36524 at 46bc6b4 (2023-12-31) squashed on 10.3.beta4
Well, it generates files that can be compared with what has been committed to #36524. |
The script was taken from:
... where it was called
generate-meson.pyand intended to be removed before merging. #36524 (comment)Here we make it available as a script installed by
sage_setupbecause what was useful once for generating code will be useful later for updating code, in particular as will be necessary for modularization concerns, as discussed in #36524 (comment)Various improvements have been made. When invoked with the switch
--monolithic, the output maintains compatibility with #36524.Author: @tobiasdiez, @mkoeppe
📝 Checklist
⌛ Dependencies