Neil Bothwick schreef: > On Tue, 06 Dec 2005 20:33:58 +0100, Holly Bostick wrote: > > >> There are reasons that major Portage processes are not supposed to >> be carried out concurrently. It's true that while I'm upgrading >> KDE, I probably could emerge... oh, mutt... at the same time in >> another instance, but checking the entire Portage tree (as in >> emerge -ua** world) is just not recommended, since Portage can't >> really be expected to know which updates are really valid (because >> it's currently performing some of them in another instance, with >> which it can't communicate). > > > The emerge -u invocation wasn't actually emerging anything, so it > won't write to world.
Yes, got that, but it would be scanning the world file to determine what had updates available that had not been performed. If some of the updates were at the same moment being performed in another instance, that is not likely to confuse Portage as to whether an update to konqueror (for example) was available or not (if konqueror was at that moment being upgraded in the other console)? > Emerging KDE only results in one write to world, at the end of the > process. I suppose, if you're only emerging kde-meta (since all the sub-builds are dependencies that will not be added to your world file. If I'm upgrading the split ebuilds, though.... does that remain the case? > I expect there is another cause unless the OP was incredibly unlucky > or impatient. Well, it's less likely, but it does happen. > >> I mean, you could all be quite right, and I wrong, but Istr I got >> this message when I did something similarly unwise, and after the >> emerge actually finished, there was nothing wrong anymore (because >> the left hand knew what the right hand had done). > > > I've had this a few times, and each time it was because the ebuild > for an installed package was no longer available. Also possible. Certainly there's nothing /wrong/ with running emaint --check world. I only wanted to point out that this error could also be a "false positive". Holly -- gentoo-user@gentoo.org mailing list