Ronan Pigott <rpig...@berkeley.edu> added the comment:
Any reason a change like this could not be added to a 3.8.x release? Like I pointed out in the PR, pydoc's use of MANPAGER is undocumented and, as far as I'm concerned, incorrect. A lot of software references idiosyncratic <PROGRAM>_PAGER vars, but pydoc reuses MANPAGER instead of something like PYDOC_PAGER. I don't think it's such a bad idea to reuse MANPAGER considering the two programs serve a similar function, but I think matching man's interpretation of MANPAGER is then the only reasonable behavior pydoc can take. If it does not, pydoc is by-default broken on plenty of otherwise fine setups. As it stands, pydoc is broken _because_ of undocumented behavior. A user reading the docs carefully still might not recognize that pydoc is preferring their MANPAGER value over PAGER. For that reason, I'd call this a bugfix. The bug here could be considered that pydoc references MANPAGER _at all_. However, given what I think is the intent of the original code - to allow pydoc to benefit from a user's MANPAGER customization - I've tried to update pydoc to reflect that intent in the PR, rather than remove the feature. I think this change could be eligible for 3.8.x. Thoughts? ---------- _______________________________________ Python tracker <rep...@bugs.python.org> <https://bugs.python.org/issue39400> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
