On Jun 7, 2007, at 9:13 PM, Ryan Schmidt <ryandesign@macports.org> wrote:

...

This would require no periodic checking, other than the `port sync` which has to happen anyway. Nor would there be any possibility of the user having problems in the intermittent period between the server update and their check.

Why do we even have two commands -- sync and selfupdate? Why can't we just have a single command which always does the right thing?

What if (for development/testing purposes) you had a branch of base from macports-1.4.40 installed, say, and you did a selfupdate? Wouldn't that either fail or overwrite your branch with the standard macports-1.4.42? Yet you might want to update the portfiles even though you were running with a branched base (and at least for me, "port sync" is a lot faster than "svn update" for grabbing the latest portfiles). To me, that provides justification for a "port sync" separate from "port selfupdate" -- I have used "port sync" for this reason. Of course, I could be mistaken about what happens with a "port selfupdate" when your current installation is a branch, but I've never been brave enough to try it, lest things end up in a muddled state. Regards, Glen