This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: CFG fixup after early inlining
> Yes, it is not supposed to happen so early. fixup_cfg does two things -
> it clears EH after setting after_inlining to tru and it adds edtges for
> non-local gotos. It looks like proper way is to break out this code and
> move it to inliner (I must admit I am not sure at all why I placed it in
> fixup cfg at first place, perhaps I just didn't wanted to introduce
> another pass over GIMPLE and early inlining didn't exist at that time).
I wrote this non-local gotos fixup code. :-)
> I can prepare patch this evening and send you it for testing (and try to
> reproduce problem on my setup too)
Fine with me. Thanks for stepping in, I'm not familiar enough with the new
machinery yet to write something well thought out here.
> Sorry for the problem,
No problem, the mainline FSF Ada compiler is in good shape, we only detected
the bug because we use slightly different settings locally.
--
Eric Botcazou