gh-98586: add whatsnew entry covering limited-API outgoing vector calls, related documentation fix#99056
Conversation
PR python#98587 addressing issue python#98586 lacked a "what's new" entry. While making those changes, I noticed an inconsistency in how ``PY_VECTORCALL_ARGUMENTS_OFFSET`` is declared in the underlying Sphinx markup when compared to other macro constants like ``Py_TPFLAGS_HAVE_VECTORCALL``. This PR fixes that as well, which should connect a few currently broken cross-references
|
FYI, you should mention in the PR title (which is also the default commit message) what you are adding a What's New entry for... |
CAM-Gerlach
left a comment
There was a problem hiding this comment.
Seems like these should be changed to use equivalent roles from the appropriate domain, though I'm not sure why the Python domain has data/const but the C domain doesn't, despite C actually having "real" constants and Python only having a convention, nor what the reason was for using construct from the "wrong" domain originally...
@AA-Turner , any expert insight here?
Misc/NEWS.d/next/Core and Builtins/2022-10-24-10-30-30.gh-issue-98586.Tha5Iy.rst
Show resolved
Hide resolved
Co-authored-by: C.A.M. Gerlach <CAM.Gerlach@Gerlach.CAM>
|
@CAM-Gerlach -- I updated the PR title (sorry) and merged the change with the missing article. I am hesitant to make the domain-related changes you requested. Part of the reason is that I am following the pattern that generally seems to be used for macro constants. For example look at how |
CAM-Gerlach
left a comment
There was a problem hiding this comment.
Okay, that makes sense for now I guess. In that case, LGTM, aside from one comment
Co-authored-by: C.A.M. Gerlach <CAM.Gerlach@Gerlach.CAM>
Incorporated, thanks. |
Co-authored-by: Erlend E. Aasland <erlend.aasland@protonmail.com>
erlend-aasland
left a comment
There was a problem hiding this comment.
LGTM, thanks!
Related: we should really update the docs on best practice for writing exension modules using the Limited C API (there is already an issue for this).
PR #98587 addressing issue #98586 lacked a "what's new" entry.
While making those changes, I noticed an inconsistency in how
PY_VECTORCALL_ARGUMENTS_OFFSETis declared in the underlying Sphinx markup when compared to other macro constants likePy_TPFLAGS_HAVE_VECTORCALL. This PR fixes that as well, which should connect a few currently broken cross-references