Summary: | sys-libs/db-4.1.25_p1-r4 - emerge fails to unpack tarball when emerging system, tar complains that "Directory renamed before its status could be extracted" | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Sam Gardner-Dell <sam> |
Component: | [OLD] Core system | Assignee: | Paul de Vrieze (RETIRED) <pauldv> |
Status: | RESOLVED INVALID | ||
Severity: | critical | CC: | base-system |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | x86 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Sam Gardner-Dell
2005-01-09 03:56:24 UTC
I've never seen problems with it before, but looking at your monstrous compiler flags, there is probably one (or more) of them causing havoc with tar. Please check info gcc on each an every of the flags. For example -ffast-math is not suitable as a global flag, and should only be enabled by the author on one package. Also keep in mind that the gcc compiler is not bug-free. Especially sse2 support is still broken. So, recompile tar (and possibly other programs) and try again I'm also having problems with db-4.1.25 during install of 2004.3 stage-1. scripts/bootstrap.sh seems unable to download it from anywhere. I've run emerge sync a couple of times to see if it's fixed in CVS, but at this point, even attempting to download it from sleepycat fails. I just "hack" fixed my ebuild to get this one to work. Summary: "sleepycat" change the location of the file download. Ebuild has not been updated. See: http://forums.gentoo.org/viewtopic.php?p=2033440#2033440 Who's the ebuild maintainer for this? I maintain it, but your problems are not caused by moved tarballs. I'll change the location in the ebuilds. This kind of behaviour from upstream is why gentoo mirrors packages. That makes sure that people can download things. If things can not be downloaded the error is different though. This seems to be an issue with tar (Sam, do you by any chance run a different tar from app-arch/tar)? I have now bootstapped again from scratch using a more mild set of optimisations - actually not that much milder, went back to my old trusted set (removed malign-double and ffast-math which were recommended on one particular forum thread about Pentium 4M flags). Sorry to inconvenience you Paul - and well spotted, I would have expected compile time errors rather than odd behaviour from tar. Still we learn something new every day. Changing status to Resolved, Invalid |