You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[Doc]: clarify when and how to use boilerplate.py #27190
Per this discussion #27158 (comment), I (maybe other folk) are a bit confused on what the "do not edit" comments in pyplot mean given that boilerplate.py is never run automatically. Consensus seems to be that the process should be when changing the signature of a pyplot function, run boilerplate.py and commit those changes but that's not documented anywhere.
Suggested improvement
More clearly document using boilerplate.py by:
changing the Autogen message to "generated by biolerplate.py, do not edit"
# Autogenerated by boilerplate.py. Do not edit as changes will be lost."""
change the comment at the top of pyplot to something like "plotting signatures are generated by boilerplate.py, rerun when changing signatures and commit the changes
# Note: The first part of this file can be modified in place, but the latter
# part is autogenerated by the boilerplate.py script.
add something like (note: this rst may be incorrect) to the coding guideline.
When a function in the :doc:`pyplot API<doc/api/pyplot_summary>`, run :file:`lib/matplotlib/test_pyplot.py::test_pyplot_up_to_date`. If the test fails and and you intended to change those signatures, run boilerplate.py and commit the changes
Consensus seems to be that the process should be when changing the signature of a pyplot function, run boilerplate.py and commit those changes but that's not documented anywhere.
This is generally true, though there are functions in there that are not part of the automatically generated set. An easier indicator is "if the lib/matplotlib/tests/test_pyplot.py::test_pyplot_up_to_date test fails", but one has to clarify that with "and you intended to change those signatures".
Activity
[-][Doc]: clarify when and how to use biolerplate.py[/-][+][Doc]: clarify when and how to use boilerplate.py[/+]QuLogic commentedon Oct 25, 2023
This is generally true, though there are functions in there that are not part of the automatically generated set. An easier indicator is "if the
lib/matplotlib/tests/test_pyplot.py::test_pyplot_up_to_datetest fails", but one has to clarify that with "and you intended to change those signatures".DOC: improve description of boilerplate.py
DOC: improve description of boilerplate.py
DOC: improve description of boilerplate.py
DOC: improve description of boilerplate.py