On 11/19/10 01:07 PM, Danek Duvall wrote:
Peter Dennis wrote:
Okay - and that was the answer I was looking for - essentially this means
it is something that pkg(5) can handle (by setting a timestamp attribute
in the manifest) but it is something that has not been implemented in the
manifests - and so the question then translates should it be ?
In general, there are very few files that require a particular timestamp.
The primary use case is for .py files, because the corresponding .pyc files
have the .py file timestamps baked into them, and will be recompiled if the
timestamps don't match. It's a bad scheme, IMHO, but it's what we have to
deal with right now.
Like johansen said, if there are other use-cases, then we can always put
the timestamp attribute on, but I'm not aware of any at the moment.
It would seem to me that in the absence of a timestamp on the individual
file actions, the timestamp on the containing package FMRI would be
better than using the installed time. Without setting timestamps
consistently across installations, tools that compare timestamps to
determine what to do don't work at intended. For example, make(1). On
a single installed environment, this generally isn't a problem. When
you want to distribute work across machines and/or use zones, the
inconsistent timestamps can be a problem.
-Norm
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
