Closed tarsius closed 4 years ago
I think my original intention was to have folks add (pull-request-merge . nil)
to the list; if that's not working, I'll have to investigate.
I didn't check again whether (... . nil)
works too, but vaguely remember that it didn't. But even if that works I still find this a bit weird. Is it really necessary to force users to use a setup function for configuration? And if so, then that should at least take care of it all. You could for example allow this:
(magithub-feature-autoinject
'((pull-request-merge . nil) ; don't enable
(commit-browse . t) ; enable
completion ; enable (backward compatibility)
(issues-section . t)
(pull-requests-section . t)
(status-checks-header . nil)))
We could make that work. How should multiple calls to autoinject
be handled, though? I'll see if I can work on this tonight, but I might be helping to write thank-you cards :-)
How should multiple calls to autoinject be handled, though?
I was wondering that too - including how does that work now :wink:
but I might be helping to write thank-you cards
Better do that then...
including how does that work now
Right now feature injection works on a symbol-by-symbol basis -- the t
and all
descriptors just map over all the symbols available in the feature list.
I have the following in my configuration because I don't want to use
pull-request-merge
andstatus-checks-header
:As a result I get warnings about not having configured the features that I don't want to use:
What am I supposed to do to disable a feature? I had to pretend that I had configured these features to make the warnings about not having done so go away: