LWN: Comments on "Gentoo Weekly Newsletter 20 September 2004" https://lwn.net/Articles/102824/ This is a special feed containing comments posted to the individual LWN article titled "Gentoo Weekly Newsletter 20 September 2004". en-us Fri, 03 Oct 2025 11:27:32 +0000 Fri, 03 Oct 2025 11:27:32 +0000 https://www.rssboard.org/rss-specification lwn@lwn.net emerge world broken? https://lwn.net/Articles/102880/ https://lwn.net/Articles/102880/ Duncan I'm not having the problem, altho I switched to the 2.0.51 series back <br> with about pre20, and updated from there to pre23, before updating to rc1. <br> Also, I'm on AMD64 altho that really shouldn't affect portage, and run an <br> all ~amd64 system, save for a few things I upgrade early that are still -* <br> masked. (Such as ..51 was when I upgraded, but I read the dev list and <br> they said it was about to go stable and they needed testers then, so I <br> tested. It also incorporated an etc-update fix I bugged and posted the <br> fix for, so tho I had installed the fix manually for ..50, I wanted to <br> upgrade to one including it. Also, the xorg 6.7.99 betas for 6.8.0, and <br> occasionally I'll keep a previously ~amd64 testing ebuild, like glibc or <br> binutils, when it gets arch removed due to some bug I don't see, such as <br> the latest glibc withdrawal due to kernel 2.4 issues, when 2.4 has been <br> depreciated on amd64 for some time.) Anyway, with all that, it's possible <br> I'm not seeing it because I had an earlier version installed or because I <br> normally run ~amd64 or some other similar difference. <br> <br> I'd suggest verifying whether your problem has been bugged or not, <br> however, and bugging it if not. That's the sort of info they need right <br> now. Even if it happens to be something strange with your system, getting <br> that sort of strange problem bugged gives them another datapoint if <br> someone else ends up with it as well, if /nothing/ else. However, /do/ be <br> sure and point out anything you are running like the few remasked things I <br> run on occasion, as that /could/ be why. It's still a datapoint and <br> therefore valuable info to have at this point, tho, so don't let that stop <br> you bugging it. <br> <br> FWIW, yes, I've had a number of ebuilds bail, and yes, I do use the ebuild <br> command to pick up again, so yes, I've tested it, and yes, it does work, <br> here. Same with the downgrade, tho I seldom let it do a downgrade the <br> first time it wants to. I check out /why/ it wants to downgrade, first, <br> why the keyword was removed, and /then/ decide if I want to let it do so <br> or put the ebuild in my overlay with ~arch reinserted so it doesn't <br> downgrade. Either way, it does seem to work, here, as I just downgraded <br> from gcc-3.4.2 back to 3.4.1, when it was masked, as I /was/ having issues <br> with 3.4.2 myself, and the downgrade worked then. <br> <br> (BTW, yes, I /know/ this is a bit wordy, but.. &lt;g&gt;) <br> <br> Duncan <br> Tue, 21 Sep 2004 08:44:21 +0000 emerge world broken? https://lwn.net/Articles/102841/ https://lwn.net/Articles/102841/ mikachu forgot to mention, this is about portage-2.0.51_rc1<br> Mon, 20 Sep 2004 21:09:15 +0000 emerge world broken? https://lwn.net/Articles/102840/ https://lwn.net/Articles/102840/ mikachu It seems that emerge world stops working when the ebuild in /usr/portage for the specific version of a package you have installed is removed in an rsync. It just spits out some aux_get errors and dies. Seems something similar happens for packages installed with ebuild foo.ebuild qmerge (which is nice when the make dies in a large file, you fix it and run make, then ebuild foo.ebuild install qmerge)<br> Mon, 20 Sep 2004 21:08:21 +0000