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
explicit references to numpy.random.mtrand.RandomState in the docs #10545
numpy 1.17.0 changed their random module docs, which lead our doc build fails. Those were fixed in gh-10540 by linking directly to numpy.random.mtrand.RandomState, which is probably undesired (note the mtrand, which is an implementation detail). Apparently, this is a numpy intersphinx regression, which is hopefully fixed on the numpy end at some point. When it is, we should roll back and link to numpy.random.RandomState
IMO, the whole issue is relatively minor. So it's up to RM's convenience. If backporting it (or not backporting it) makes your job easier, it's fine either way.
Activity
tylerjereddy commentedon Jul 30, 2019
Is our interim fix sufficiently bad that I shouldn't backport it to 1.3.x?
ev-br commentedon Jul 30, 2019
IMO, the whole issue is relatively minor. So it's up to RM's convenience. If backporting it (or not backporting it) makes your job easier, it's fine either way.
andyfaff commentedon Jan 21, 2020
@rgommers is this rollback-able now?
larsoner commentedon Jan 21, 2020
It looks like if we switch to using numpy.org/devdocs it should be
rgommers commentedon Jan 21, 2020
Indeed. And switching the intersphinx mapping in
doc/source/conf.pyis a useful change anywayj-bowhay commentedon Jan 19, 2023
close by #11396, have confirmed there are no other occurrences