Question about mem_set_list in propagate_block_info

[Toon Moene]

Diego Novillo wrote:

> On Wed, 2003-06-04 at 14:46, Toon Moene wrote:
> 
>>Richard Henderson wrote:
>>
>>>On Wed, Jun 04, 2003 at 12:25:05AM -0400, Diego Novillo wrote:
>>>
>>>>Wait, that doesn't follow.  You can have different alias set numbers
>>>>that are conflicting.
>>>
>>>Oh, yes, so you can.
>>>
>>>I still bet his source is wrong.
>>
>>Huh ?  What am I missing ?  What's the use of alias sets, then ?
>>
> Alias sets are organized in a tree hierarchy (see the comments at the
> start of alias.c).

:-)  This is soooo funny.  Of course I wouldn't have thought of looking 
into alias.c for an explanation, as to me that file was home of John 
Carr's implementation of alias analysis, which I defended in vain on the 
old gcc mailing list (before the egcs fork).

I'm thinking of a line from Tracy Kidder's "The soul of a new machine": 
Hire graduates fresh from school - they don't know yet what's impossible.

Cheers,

-- 
Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://gcc-g95.sourceforge.net/ (under construction)