Summary: | dev-python/numpy-1.3.0-r1: sandbox violation | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | John Armitage <johnparmitage> |
Component: | Current packages | Assignee: | Gentoo Science Related Packages <sci> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | ansla80, dabbott, Dan.Johansson, gentoo.cart9, jw5801, pacho, prefix, python, qa, SebastianLuther, slyfox, steven, swapon, toto |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 287575 | ||
Attachments: |
access violation summary
Fix sandbox violations |
Description
John Armitage
2009-10-10 20:00:06 UTC
Created attachment 206700 [details]
access violation summary
Created attachment 206716 [details, diff]
Fix sandbox violations
Revert ${D} => ${ED}
This problem seems indeed to be due to the ${ED} bit in src_install and was introduced by commit #1.5. However, ED is part of prefix as far as I know so we shouldn't just change it back. I've cc'd the relevant person for advice on how to fix this since I am not sure where ED is supposed to come from. Thanks, Markus I have the same problem. What shoul I do? Thank you very much. I screwed up the process indeed, a missing line like this was missing: [[ -z ${ED} ]] && local ED=${D} this looks somewhat useless, but in Prefix ED is already defined by Portage to the offsetted variant of ${D}. I'm sorry for the trouble caused. Just remember that you can always revert stuff and file a bug against the Prefix team, as it's more important that normal Gentoo works at this stage. If this bug causes a problem for stabilisation, how about dumping the prefix bits in a -r2 and reverting -r1 to its original state? (In reply to comment #5) > If this bug causes a problem for stabilisation, how about dumping the prefix > bits in a -r2 and reverting -r1 to its original state? Just fix it in 1.3.0-r1. Also please consider discussing future changes outside of KEYWORDS variables with packages' maintainers before committing them :) . (In reply to comment #5) > If this bug causes a problem for stabilisation, how about dumping the prefix > bits in a -r2 and reverting -r1 to its original state? > 1.3.0-r1 is already stable on quite a few arches which is why this is sort of urgent ;) Since I unfortunately don't know enough about prefix I suggest that you either fix the stable 1.3.0-r1 or revert the changes for now. I presume it was an oversight that you added the prefix stuff to a stable version since these types of major changes should definitely go into ~ first I think. Thanks a lot, Markus I fixed it immediately, just didn't close the bug because I wanted you to verify. It's fixed. (In reply to comment #8) > I fixed it immediately, just didn't close the bug because I wanted you to > verify. > Thanks a lot for the quick response and it's fixed now. cheers, Markus *** Bug 288573 has been marked as a duplicate of this bug. *** |