Context Navigation

Changes between Version 1 and Version 2 of Ticket #29301, comment 8

-              v1
+              v2
 (B) If we take the interpretation that management commands are ultimately in control of their arguments, I would expect Django to back off if sees that a management command added its own argument with a conflicting name. Adding new default arguments would then be a minor change.
 Before the original patch for this ticket was merged, add_arguments() was allowed even more leeway than (B), potentially removing default arguments outright, setting a custom formatter, or doing other unusual things, even though those possibilities were not documented. I don't think we should necessary support arbitrary actions in add_arguments().
+(C) Before the original patch for this ticket was merged, add_arguments() was allowed even more leeway than (B), potentially removing default arguments outright, setting a custom formatter, or doing other unusual things, even though those possibilities were not documented. I don't think we should necessary support arbitrary actions in add_arguments().
 I'm personally +0 on approach (B). It would involve a small additional patch where Django checks whether each default argument already existed before adding its own.
+I'm personally +1 on approach (B). It would involve a small additional patch where Django checks whether each default argument already existed before adding its own.
 ---