| 1 |
On Thursday 17 December 2009 04:38:36 Markus Duft wrote: |
| 2 |
> I recently did some ia64-hpux hacking in prefix, and i stumbled over |
| 3 |
> alittle problem: |
| 4 |
> |
| 5 |
> the native linker collects -L options on the command line, and python |
| 6 |
> uses -L. while building. this results in . beeing in rpath. |
| 7 |
> |
| 8 |
> now when $PWD == $S, there is libpython-*.so, and executing the already |
| 9 |
> installed python finds (through rpath ".") this local libpython. this |
| 10 |
> (although binary exactly the same) refuses to load installed extensions |
| 11 |
> (Fatal Python Error: ....). |
| 12 |
> |
| 13 |
> The problem was caused by the filter-bash-environment.py call in |
| 14 |
> ebuild.sh, _only_ during the "test" phase seemingly. |
| 15 |
> |
| 16 |
> the attached patch fixes the problem for me - what do you think? I guess |
| 17 |
> there are better ways to tackle the problem, but i needed to get this to |
| 18 |
> work ;) |
| 19 |
|
| 20 |
this ignores the real problem. there should never be an installed binary |
| 21 |
whose rpath includes $PWD. this is a terrible security issues waiting to bite |
| 22 |
you in the ass, or cause random failures depending on what the user's cwd is |
| 23 |
at any point in time. |
| 24 |
|
| 25 |
the second change (adding the || exit) is a good idea in theory, but the |
| 26 |
proposed change is incorrect. use the assert helper so that the exit values |
| 27 |
of all commands in the pipeline are checked. |
| 28 |
-mike |
Archives