Re: [Patch] Fixing two bugs in sbitmap functions.

[Daniel Berlin <dberlin at dberlin dot org>]

On Dec 23, 2003, at 8:36 AM, Roger Sayle wrote:

> On Tue, 23 Dec 2003, Mostafa Hagog wrote:
> > 2003-12-12  Mostafa Hagog <mustafa@il.ibm.com>
> >
> >         * sbitmap.c (sbitmap_union_of_diff_cg, sbitmap_a_and_b_cg,
> >           sbitmap_a_xor_b_cg): accumulate "changed" properly.
> >           (sbitmap_not): zero all bits past n_bit.
>
> This is OK for mainline.
>
> Remember to capitalize the first letters of the sentences in the
> ChangeLog entry "Zero all bits past n_bit", each line begins with
> a tab character so the second and subsequent lines align under the
> asterisk, and in future please report the full target triple on
> which you bootstrapped and regression tested.
>
>
> However, the changes themselves are obviously correct bug-fixes.

Yup. I'm surprised it took so long to spot them, since the "did it 
change" versions are used in LCM, bt-load, at -O*, and GCSE (VBE) at 
-Os.

This probably increases our time slightly in these passes now that we 
iterate the proper number of times again.
--Dan