| 1 |
So, I think I've discovered a subtle bug that should be pretty easy to fix. |
| 2 |
|
| 3 |
I had set EPREFIX to /gsa/ausgsa/projects/g/gnu as this is the path name |
| 4 |
provided to me by the owner of the system. Everything has been emerging fine until |
| 5 |
today I tried emerging python-fchksum-1.7.1. This fails due to a QA notice about |
| 6 |
files being installed outside of the prefix: |
| 7 |
|
| 8 |
* QA Notice: the following files are outside of the prefix: |
| 9 |
* /gsa/ausgsa-p5 |
| 10 |
/gsa/ausgsa-p5/15 |
| 11 |
/gsa/ausgsa-p5/15/gnu |
| 12 |
|
| 13 |
So, I found it odd that it was reporting the prefix as "/gsa/ausgsa-p5" and did some |
| 14 |
checking on the system. In /gsa/ausgsa/projects/g there is a symlink for gnu: |
| 15 |
|
| 16 |
gnu -> ../../../ausgsa-p5/15/gnu |
| 17 |
|
| 18 |
which thus evaluates to "/gsa/ausgsa-p5/15/gnu". So, why is the eprefix being set to |
| 19 |
/gsa/ausgsa-p5 in this particular QA check ? |
| 20 |
|
| 21 |
At any rate, I found an easy workaround. I exported EPREFIX to be the actual _real_ |
| 22 |
path with no symlinks in it, as in: EPREFIX=/gsa/ausgsa-p5/15/gnu and then was able |
| 23 |
to emerge the package successfully. |
| 24 |
|
| 25 |
Is this something prefixed-portage should handle, or the ebuild or should I solve it |
| 26 |
manually via the export trick I did ? |
| 27 |
|
| 28 |
Thanks! |
| 29 |
-- |
| 30 |
gentoo-alt@g.o mailing list |
Archives