diff options
Diffstat (limited to 'eclass/multilib.eclass')
| -rw-r--r-- | eclass/multilib.eclass | 15 |
1 files changed, 5 insertions, 10 deletions
diff --git a/eclass/multilib.eclass b/eclass/multilib.eclass index aad705b..bf59bb4 100644 --- a/eclass/multilib.eclass +++ b/eclass/multilib.eclass @@ -469,17 +469,12 @@ multilib_toolchain_setup() { # I find this pretty confusing. I don't think we want to let tc-getPKG_CONFIG # look up our pkg-config for us once we change CHOST -- that is now going to # pick up the <abi-tuple>-pkg-config executable which could come from crossdev - # or god-knows where... so basically unless we are doing a compile for crossdev, we want + # or god-knows where... unless we are doing a compile for crossdev, we want # to treat all multi-abi pkg-configs the same, /except/ for the two variables - # above... I guess? Who fucking knows, this will probably break something somewhere - # but I am inclined to think that such breakage might indicate a bug in the - # consuming ebuild instead of here. It's a fucking text-file processor, after all, - # like "sed". Who cares if we run our own so long as we give it the right environment - # variables? An alternative idea would be to always rely on tc-get-BUILD_PKG_CONFIG - # in preference to tc-getPKG_CONFIG. Although the names are confusing, I think - # the results of using that in preference to tc-getPKG_CONFIG would actually fix - # a lot of portage problems. A third alternative would be to special-case tc-getPKG_CONFIG - # to return results from the build machine.... + # above. + # + # pkg-config is a text-file processor, after all, like "sed". Who cares if we + # run our own so long as we give it the right environment variables? # # Not sure I'm happy with any of these solutions or the one below. If anyone # has a crystal-clear idea of the full semantic import of this... drop me a line! |