[email protected] wrote:
Hey Bart,
On Wed, Nov 25, 2009 at 02:26:19PM -0800, Bart Smaalders wrote:
Some fixes to facilitate faster/simpler publication to file://
repos:
http://cr.opensolaris.org/~barts/faster-publication/
This looks good. I just had a few questions:
depot.py:
- This has been coded so that add-content doesn't work if the depot is
in read-only mode. The writable-root option allows the catalog to
be upgraded v0 -> v1 into an alternate directory if readonly and
writable-root are set. Would make sense to allow
rebuild/add-content on read-only if the rebuild were to occur into
the writable root?
This seems quite confusing... are we sure we want to go here?
repository.py:
- lines 660 - 664: If add_package fails because of a duplicate entry,
is the catalog state kept consistent? With the old catalog, this
might have resulted in a partial entry, but I can't remember for
sure.
- lines 744/749: Should we be setting SA_NOCLDSTOP so that job
control operations don't cause a SIGCHLD to be sent while we're
waiting for the process to exit? Or is something else sending us
signals and causing the EINTR?
Thanks,
-j
For some reason, we're setting a sig child handler in the depot code...
I can see if I can avoid doing so if we're going to be synchronous.
- Bart
--
Bart Smaalders Solaris Kernel Performance
[email protected] http://blogs.sun.com/barts
"You will contribute more with mercurial than with thunderbird."
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
