On Sun, 20 Jan 2008, Kenneth Zadeck wrote:
> If you
> build your compiler with --enable-checking=df,yes (basically add "df"
> onto what ever you normally set for --enable-checking) it will check that
> there have been no unauthorized changes to any instructions after
> every rtl pass.
For the reco
However,making this simpler works:
rtx tmp_reg_rtx = copy_to_mode_reg (QImode,gen_rtx_MEM (QImode, addr1));
emit_move_insn (addr1, gen_rtx_PLUS (Pmode, addr1, const1_rtx));
aka
Rx=[Ry]
Ry=Ry+1
For now I have gone back to the second case, though the code is not
quite as good.
When your patter
Alas, enable-checking produced no different result or additional
warnings or errors (though it might help me in the future!)
I have a work around but don't fully understand why a define_expand should
have caused segmentation fault.
I believe the issue might be that gcse does not expect to see
It is most likely that this bug is actually caused by one of the
modifications that you have made to the backend.
In 4.3, the rtl backend is very picky about the kinds of modifications
that can be made to insns and in particular the api's that are used to
modify, insert or delete insns or even w
