[Bug c/30477] Integer Overflow detection code optimised away, -fwrapv broken

["tg at mirbsd dot de" <gcc-bugzilla@gcc.gnu.org>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1178 bytes --]



------- Comment #15 from tg at mirbsd dot de  2007-01-22 23:54 -------
Subject: Re:  Integer Overflow detection code optimised away,
 -fwrapv broken

pinskia at gcc dot gnu dot org dixit:

>fold-const.c changed a lot, etc.

>Actually there are two different code, one I wrote which is does
>folding of a-10 > 0 into a>10 and other code which folds a-10>a into true,
>I wrote the first one.

I found the second one in CVSweb, and it's not the cause for
this unsafe "optimisation". I even changed fold-const.c to
have some wrapper around fold() which debug_tree()s me the
input and output, and the '100' stays in (at -O1, which does
exhibit the faulty behaviour already):

[…] arg 1 <integer_cst 0xa7d85b7c constant 100>> arg 1 <parm_decl 0xa7d1cec4
a>>

Now I don't know any gcc internals, but I suppose this isn't
done in fold-const.c… thanks to fprintf, my beloved debugger ;)

bye,
//mirabile
--
  "Using Lynx is like wearing a really good pair of shades: cuts out
   the glare and harmful UV (ultra-vanity), and you feel so-o-o COOL."
                                         -- Henry Nelson, March 1999


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30477