1 |
On Mon, 2009-08-03 at 20:48 +0200, Fabian Groffen wrote: |
2 |
> On 31-07-2009 15:03:25 +0100, Alan Hourihane wrote: |
3 |
> > Can we finally get rid of the horrible GCC kludge for the symlink issue |
4 |
> > in the eclass ? |
5 |
> > |
6 |
> > Or at the very least provide a flag in which I can turn it off without |
7 |
> > having the need to go poking eclass everytime I've re-synced ? |
8 |
> |
9 |
> how about this compromise? |
10 |
> |
11 |
> @@ -19,7 +19,7 @@ |
12 |
> toolchain_pkg_setup() { |
13 |
> # yuck, but how else to do it portable? |
14 |
> local realEPREFIX=$(python -c 'import os; print os.path.realpath("'"${EPREFIX}"'")') |
15 |
> - if [[ ${EPREFIX} != ${realEPREFIX} ]] ; then |
16 |
> + if [[ -z ${I_KNOW_MY_GCC_WORKS_FINE_WITH_SYMLINKS} && ${EPREFIX} != ${realEPREFIX} ]] ; then |
17 |
> ewarn "Your \${EPREFIX} contains one or more symlinks. GCC has a" |
18 |
> ewarn "bug which prevents it from working properly when there are" |
19 |
> ewarn "symlinks in your \${EPREFIX}." |
20 |
> |
21 |
|
22 |
That'd be great, but remember that the realEPREFIX check is in more than |
23 |
one place. |
24 |
|
25 |
Thanks, |
26 |
|
27 |
Alan. |