[PATCH] Search user plugin directories before site plugin directories

[Aaron Bentley]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

John Arbash Meinel wrote:

> And we can always change the import code to load in reverse discovery
> order.

I think we should.

> I think this initial patch should be merged, and then we can fix the
> "import last found non-conflicting plugin first" so that plugins get a
> chance to override commands at the right time.

Yes.  I've changed my vote.

> Actually, if you look at the "register_command()" function, if a plugin
> doesn't supply "decorate=True" it won't get precedence. So actually it is:
> 
> 1. "bar" is loaded
> 2, "bar" overrides the builtin diff
> 3. "foo" is loaded
> 4. "foo" tries to overload "bar"'s diff, but can't because diff is
>    already overloaded.

I used "diff" in that example because in order for "foo" to work
normally, it would have to supply decorate=True.

If someone else implemented a "zap" command, there's no reason to think
it would work like bzrtools zap, so failure is probably the right result.

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFF3zEB0F+nu1YWqI0RAlb+AJ9fEvl0tZ65vLVeppaRj4l08Fz3YwCeKrbh
eIgs4xwhb+VFJTiBHbGO2S4=
=giY+
-----END PGP SIGNATURE-----