--- Additional Comments From para at cfl dot rr dot com 2005-07-05 14:40
---
The analysis is slightly flawed. For example:
UINT32 r0045025C = opt_and(ic, r004501D4);// N = and (_, M) :00022108
UINT32 r00450994 = opt_not(r0045025C);// b = not (N) :fffddef7
--- Additional Comments From rsandifo at gcc dot gnu dot org 2005-07-05
14:51 ---
The analysis is slightly flawed. For example:
UINT32 r0045025C = opt_and(ic, r004501D4); // N = and (_, M) :00022108
UINT32 r00450994 = opt_not(r0045025C); // b = not (N)
--- Additional Comments From rsandifo at gcc dot gnu dot org 2005-07-04
13:49 ---
Are you sure this reduced testcase is an accurate reflection of the
real one? The transformation performed by gcc appears to be correct.
For example, r00450DB4 evaluates to 0, so anything ANDed with it
--- Additional Comments From rsandifo at gcc dot gnu dot org 2005-07-04
13:52 ---
Oops, the following line was bogus:
UINT32 r00450994 = opt_not(r0045025C);// b = not (N) :fffddef7
the value should be . It doesn't affect the analysis though.
--
--- Additional Comments From pinskia at gcc dot gnu dot org 2005-06-08
23:14 ---
Confirmed, but this looks like only a bug in 3.4.x and before. It might be a
latent bug on the mainline
still I don't know for sure.
--
What|Removed |Added