Cairo fails to emerge [SOLVED]

[Incomplet]

When I try to emerge cairo (1.10.2-r1 and r2), it fails to compile tossing errors about not being able to use threads and then all the mutex objects fail to build.

The original error said to define CAIRO_NO_MUTEX to 1, so I set up CAIRO_NO_MUTEX="1" in make.conf, and that did nothing. I tried setting it from the command line, and more nothing.

I've tried tinkering with a few other things like the "threads" flag but nothing works.

[ebuild N ] x11-libs/cairo-1.10.2-r2 USE="X glib opengl svg (-aqua) -debug -directfb -doc (-drm) (-gallium) (-openvg) -qt4 -static-libs -xcb"

Environment
Build Log

[PaulBredbury]

Probably not related, but I doubt you want to compile cairo with opengl support.

Last time I tried --enable-gl, gtk+-2 wouldn't compile, and I believe it doesn't work well with nvidia's video driver.

[Incomplet]

Ha yeah I noticed there's a note about opengl support being experimental, so I went back and changed it to directfb.
No, that didn't help it compile any better XD.

[VoidMage]

Well, the first explicit error is:

[Incomplet]

There isn't a clear config.log in var/tmp/portage/x11-libs/cairo-1.10.2-r1/temp.

There are a few files that look like they could be a config log, but their contents are all basically identical to the "is expanded from..." stuff at the beginning of the build log. Is there somewhere else I should be looking?

EDIT: There is something else strange that might be responsible:
checking for native atomic primitives... intel
checking whether atomic ops require a memory barrier... yes

I'm definitely not running intel hardware, so that might be why it fails to find the correct locking functions. I have no idea how to fix that, though.

EDIT2: nevermind. Found a similar old thread, and the intel thing probably has nothing to do with it. However, I still can't find anything else interesting .

EDIT3: Ok.. what would cause it to fail to link to pthreads? I'm not real sure what to start rebuilding x.x. The only things I can think of are glibc and the kernel headers, and I already rebuilt my kernel headers. Do my compiled kernel settings have any effect on that?

[VoidMage]

config.log is (well, in most of the cases, beside recursive configure, obviously) in ${WORKDIR}.

[Incomplet]

Well, if I don't ask stupid questions, I won't learn anything . Also, I'm working in a pure console, no MC or anything, so I'm fairly blind.

However, thanks, I nailed it. For some unknown reason it was spamming my LDFLAGS when it did its conf checking, and gcc returned "unknown option" for all of them, which returned false negatives for a bunch of things.

[HeXiLeD]