On May 9, 2012, at 2:08 PM, Ryan Schmidt wrote:
> On May 9, 2012, at 09:48, Daniel J. Luke wrote:
> 
>> On May 7, 2012, at 2:32 PM, Ryan Schmidt wrote:
>>> As far as I know things are working correctly the way they are today. What 
>>> problem are you trying to solve? It sounds like you're saying MacPorts 
>>> should compare the Portfile in the ports tree with the one in the archive, 
>>> and if they differ, ignore the archive and build from source. If so, I see 
>>> no reason to do that. If a Portfile change would result in a port needing 
>>> to be rebuilt, the committer would have increased the revision.
>> 
>> human error?
> 
> Sure. You know of a way to take humans out of the equation?

automation wherever possible

>>> And if not, then there's no reason not to use an available archive. For 
>>> example, just because someone decides to add a modeline or adjust a port's 
>>> whitespace or formatting is no reason to discard an archive built from the 
>>> previous Portfile.
>> 
>> it might be nice to automate things to the point where any change to a 
>> portfile re-creates the archive (especially if the buildbots aren't 
>> overloaded).
> 
> That doesn't entirely help. Unless the revision is increased, "port outdated" 
> will not tell a user that they should rebuild, nor will rebuilding do 
> anything unless the user forces it.
> 
> The status quo remains: maintainers must remember to increase the revision 
> when they commit a change that users should receive, same as it's always 
> been. I don't think making maintainers remember this simple rule is too much 
> to ask.

sure, revbumping is important.

I was more saying that in cases where the maintainer makes a change, but 
doesn't revbump, it might make sense to just rebuild the archive 
(unconditionally) so that the Portfile in the archive always (or almost always) 
matches the current Portfile.

--
Daniel J. Luke                                                                  
 
+========================================================+                      
  
| *---------------- [email protected] ----------------* |                      
    
| *-------------- http://www.geeklair.net -------------* |                      
    
+========================================================+                      
  
|   Opinions expressed are mine and do not necessarily   |                      
    
|          reflect the opinions of my employer.          |                      
    
+========================================================+



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

Reply via email to