http://bugs.winehq.org/show_bug.cgi?id=22890
--- Comment #11 from Richard shiningarcanine@gmail.com 2010-05-30 17:16:01 --- (In reply to comment #10)
Richard, I cannot reproduce that, I tested this extensively, strip-flags (which is what's run in the 1.1.44 ebuild when you have the custom-cflags USE flag disabled) removes -fomit-frame-pointer for me on Portage 2.2_rc63.
Review the strip-flags function and the setup-allowed-flags functions in /usr/portage/eclass/flag-o-matic.eclass
If you have -fomit-frame-pointer in your system CFLAGS, try running "ebuild $(equery which app-emulation/wine) compile" and run grep "-fomit-frame-pointer" on the build.log that is generated in a subdirectory of /var/tmp/portage. My system is currently running sys-apps/portage-2.1.8.3 and at the moment, sys-apps/portage-2.2_rc67, which you are running, is masked in the portage tree, so we are running different versions of portage, which I would suspect is the reason for the discrepancies in our observations should you not see grep "-fomit-frame-pointer" return anything from the build.log generated by the WINE ebuild on your system.
By the way, regardless of what is causing the discrepancies in our observations, I recompiled WINE with frame pointers by specifying CFLAGS, CXXFLAGS and CPPFLAGS for the build and confirmed that this issue occurs in the presence of frame pointers, so both of us are in agreement that this is an issue in WINE. It might be a good idea to take a discussion regarding whether or not the ebuild strips the flags over to the bug report I filed at the Gentoo bug tracker, because that is not a WINE issue.