RE: HOWTO: lilypond-1.3.98 on windows32

[Jeff Henrikson]

> > Is it?  Doesn't that tell configure (or make) what the type of processor to
> > optimize for?  If so, then having such a high processor requirement will
> > leave lily useless for a lot of people (personally, I have an i586 at
> > home)...
>
> Trust me.
> :-)

Perhaps a more satisfying answer would be that "regular" optimizations usually 
don't break anything on the old processors, just
make them slower.  And I don't believe there's any real meaning to the cygwin 
"686" string.  It at most controls these optimization
flags.

There are such things as MMX which would actually break things, but these are 
not compiler defaults, are usually coded by hand (or
compiled with a special intel compiler), and are usually backed with redundant 
code in the case that there is no MMX support.  I
don't know if gcc has any optional MMX/3DNOW/random-intel-stupidity code 
generation.  But any sane developer will turn it off
unless there's a good reason for it.


Jeff Henrikson