upgrades fail after leopard update

[Alan Batie]

Ryan Schmidt wrote:

> MacPorts just has no way to know. That's why it's an 
> excellent suggestion to add the OS version, Xcode version and processor 
> architecture to the unique identifier to try to make it truly unique.

We're in violent agreement ;-)  I was thinking of this future, sorry 
about that...

> You never want "port upgrade installed"; you want "port upgrade 
> outdated", if anything. "port upgrade" does not upgrade ports that are 
> not outdated, so saying "port upgrade installed" just wastes time 
> evaluating ports that won't get upgraded.

makes sense

> Just having a newer version of Xcode should not trigger a port to show 
> up in "port outdated", though once we start storing information about 
> Xcode version and such with each installed port, there could be an 
> optional more-pedantic "outdated" which could list this.

perhaps a "port rebuild outdated", where "outdated" in this case refers 
to the environment rather than the port version?  I'm assuming "port 
upgrade outdated" would eliminate the issue in the cases where the port 
*does* have a newer version.

> If you 
> want to go back to the old version, you can deactivate the new version 
> and (re)activate the old one.

Ahhh, cool.

> You already have that, for individual ports. To rebuild port foo even if 
> port doesn't think it needs to be rebuilt:
> 
> sudo port -ncuf upgrade foo

Also cool.  Though this is going to fetch the latest version, which may 
not be what you want.  Perhaps if you specified the existing version as 
the "upgrade"?

> installed and activated dovecot @1.0.13_0+darwin_8 and deactivated (but 
> left installed) dovecot @1.0.0_0+darwin_8.

> Then you asked for dovecot to be installed, and because the 
> set of variants MacPorts decided to use (only +darwin_9) differed from 
> what was already installed, it built it again, and installed (but 
> couldn't activate because of the other version already active)

I'm confused why it could activate the new version the first time and 
not the second?

> If you mean something that would rebuild all ports in order, then we 
> don't have it, and I don't think it would be "basic"; I think it would 
> be a non-trivial script to write. For me at least.

The ncuf thing is what I meant I suppose, though doing it recursively 
would be nice.  Isn't the basic code there already for the "outdated" 
operation?  There may not really be anything that needs doing, saving 
fixing the bugs I'm running into...

> I haven't had a chance to test this and my battery is about to die so 
> I'm going to send this mail now. :)

;-)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3263 bytes
Desc: S/MIME Cryptographic Signature
Url : http://lists.macosforge.org/pipermail/macports-users/attachments/20080613/0ea3b5f2/attachment.bin