On Monday June 01 2015 16:57:39 Artur Szostak wrote:

> > Did you handle deactivation or did you simply make it impossible.
> 
> The update of the configuration files is handled during the activation phase 
> and rolled-back during deactivation.
> So it should all be handled. Deactivating the port should get you back to the 
> same state as if the port was never installed in the first place. If this is 
> not the case then I should fix it :-)
> Of course, a user could always break something by making some very weird 
> modifications to the configuration files directly. But I think I have handled 
> all reasonable scenarios.

Yeah? How do you handle 

%> port install repositoryA
%> port install repositoryB
%> port deactivate|uninstall repositoryA

Do you roll back to the state from before installing repo A, removing repo B at 
the same time?

That's just the easiest case, there are probably more complicated scenarios 
where the user didn't touch any configuration file directly ...

> OK, so add-port-repository would ship with the MacPorts core?

Not necessarily, it could be installed through a port itself (cf. port-depgraph 
or port-whatsnew, 
http://svn.macports.org/repository/macports/contrib/port-whatsnew).


> Are we still talking about the same thing? I was thinking that implementing 
> the SAT-solver based solution might take a while.

Ah, the general implementation of that, sure, that's possible. I was just 
talking about the implementation of per-port priorities on top of the current 
(or revamped) per-repo priorities. And I could be wrong :)

R.
_______________________________________________
macports-dev mailing list
[email protected]
https://lists.macosforge.org/mailman/listinfo/macports-dev

Reply via email to