| Summary: | multilib.eclass:multilib_toolchain_setup: doesnt work properly with more than 2 ABIs (sys-apps/sandbox fails with n32/n64/o32 mips multilib) | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Matt Turner <mattst88> |
| Component: | New packages | Assignee: | Gentoo Toolchain Maintainers <toolchain> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | mips |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | MIPS | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: |
build.log
config-n64.log.gz config-o32.log.gz multilib profile used emerge --info -v multilib_toolchain_setup.patch |
||
Created attachment 265377 [details]
config-n64.log.gz
Created attachment 265379 [details]
config-o32.log.gz
Created attachment 265383 [details]
multilib profile used
This is the multilib profile I'm using when I get the sandbox error.
and your `emerge --info` ? Created attachment 265385 [details]
emerge --info -v
multilib_toolchain_setup seems to really only work two abis, and the order has to be non-default abi followed by default abi. at any rate, i dont see this as a bug in sandbox at all. Created attachment 265389 [details, diff]
multilib_toolchain_setup.patch
try this patch ... seems to do what i had intended originally
(In reply to comment #7) > Created attachment 265389 [details, diff] > multilib_toolchain_setup.patch > > try this patch ... seems to do what i had intended originally This works for me. :) |
Created attachment 265375 [details] build.log sandbox on multilib mips fails to compile. It compiles find for the first ABI, but for the second it appears to just append a new -mabi= flag to the end of the previous ABI's config. So, o32 fails to compile because it tries with gcc -mabi=64 -mabi=32. See attached logs.