IE6 and IE7 (and probably many other Windows applications) are broken on Gentoo when using glibc 2.11 or later. A longjmp check has been added to glibc that causes them to crash with "longjmp causes uninitialized stack frame". This check is only enabled when _FORTIFY_SOURCE > 0. Normally the default is 0 but Gentoo overrides this to 2 in its GCC patches. "NOTE: In Gentoo, @option{-D_FORTIFY_SOURCE=2} is set by default, and is activated when @option{-O} is set to 2 or higher. This enables additional compile-time and run-time checks for several libc functions. To disable, specify either @option{-U_FORTIFY_SOURCE} or @option{-D_FORTIFY_SOURCE=0}." Building Wine with -D_FORTIFY_SOURCE=0 avoids the problem. It's been suggested in the Wine bug report that the underlying problem should be fixed but it's also been suggested that something that causes crashes on "false positives" should be opt-in. I'd personally prefer to have _FORTIFY_SOURCE set to 2 by default as we are doing already. Hopefully they'll fix the underlying problem anyway.
I'd say upstream resolution is incorrect, but...
I can confirm that the resolution does work. I had the same error when trying to launch the FirstClass client, so I rebuilt wine with -D_FORTIFY_SOURCE=0 and now it will launch.
This has been worked around upstream, at least for now.
Gentoo isnt the only distro to enable fortify by default, and the upstream report seemed short on details as to why the fortify code was doing something wrong. either way, i see this being something they need to address.
I wonder if wine devs will even try to discuss it with glibc upstream. If they do, the answer may be interesting.
