mass change depends port:dbus to path:bin/dbus-daemon:dbus

[pixilla@…]

I would like the dbus-devel port to be able to satisfy dbus dependencies in other ports.

The attached patch file mass-patch-depends-dbus.diff changes all dbus dependent ports from port: style depends to path: style.

Does this seem reasonable?

[pixilla@…]

mass-patch-depends-dbus.diff is against r80648.

[mk@…]

Cc Me!

[mk@…]

Replying to pixilla@…:

mass-patch-depends-dbus.diff is against r80648.

What do you mean here? I don't see a relation of python/py25-pgsql/Portfile's r80648 to dbus.

[pixilla@…]

Replying to mk@…:

Replying to pixilla@…:

mass-patch-depends-dbus.diff is against r80648.

What do you mean here? I don't see a relation of python/py25-pgsql/Portfile's r80648 to dbus.

r80648 is the svn revision I patched.

[dports@…]

Sounds fine to me, though note the following nitpicking:

path:bin/dbus-daemon:dbus should be written as bin:dbus-daemon:dbus (unless there's a reason not to). But I'd actually be inclined to write the dependency as lib:dbus:dbus since it's presumably the library that these ports are depending on.

[mk@…]

Replying to dports@…:

Sounds fine to me, though note the following nitpicking: path:bin/dbus-daemon:dbus should be written as bin:dbus-daemon:dbus (unless there's a reason not to). But I'd actually be inclined to write the dependency as lib:dbus:dbus since it's presumably the library that these ports are depending on.

bin: and lib: are dangerous, since they satisfy any executable or library found in the path. If you have multiple MacPorts setups on a system it is essential that the correct prefix is being used. So, the only workable way is to use path here! I am inclined to update the Guide accordingly. (This was the outcome of a thread on the dev mailing list a while ago.)

[mk@…]

Replying to pixilla@…:

Replying to mk@…:

Replying to pixilla@…:

mass-patch-depends-dbus.diff is against r80648.

What do you mean here? I don't see a relation of python/py25-pgsql/Portfile's r80648 to dbus.

r80648 is the svn revision I patched.

Ah, ok, I guess I now understand that you actually meant that you diffed your local tree against that revision. Me being stupid. Sorry! :-)

[mk@…]

Replying to mk@…:

I am inclined to update the Guide accordingly. (This was the outcome of a thread on the dev mailing list a while ago.)

I updated chapter 5.4.1:  http://guide.macports.org/#reference.dependencies.types

[mk@…]

So, since dports thinks it looks fine, I guess you could proceed, pixilla, and move ahead with committing this mass patch.

I just wonder whether it would be advisable to also revbump every single patched port...

[pixilla@…]

Replying to mk@…:

So, since dports thinks it looks fine, I guess you could proceed, pixilla, and move ahead with committing this mass patch. I just wonder whether it would be advisable to also revbump every single patched port...

This patch does not change any port installed files so a rev-bump would provide no advantage.

[pixilla@…]

Replying to pixilla@…:

Replying to mk@…:

So, since dports thinks it looks fine, I guess you could proceed, pixilla, and move ahead with committing this mass patch.

I am awaiting permission from michaelld (qt4-mac and qt4-mac-devel) and afb (Terminal and xfconf) as these ports are not openmaintainer.