This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [patch] fix tfmode for i386-darwin



On May 25, 2007, at 3:22 PM, Geoffrey Keating wrote:



On 24/05/2007, at 11:29 PM, Eric Christopher wrote:


So, when Uros added the tfmode support he forgot darwin :)

This does that, the problem is that darwin doesn't really support strong aliases. There's a workaround of doing this via some assembly that'll work unless we do some odd rearrangements at link or load time. That said I'm a little leery of checking this in for that reason.

Geoff, what do you think?

Mach-O does not, has not, and will not, support aliases in any form. It can and will do odd rearrangements at link time (for example, when using 'order files' or 'dead code stripping'). You'd need to write little stub routines, along the lines of 'int fooTI (int a, int b) { return fooElse(a, b); }'.

Yeah, that was the other idea, I wasn't thinking of libgcc.a here and was just thinking of dynamic linking where we're not rearranging things. I'll write the stubs up.


thanks.

-eric


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]