tv@pobox.com said:
> More to import a release version than an experimental snapshot.  There
> are cpp bugs in 2.96, but not major ones; however, I may be inclined
> to pull in a 2.96 snapshot instead of 2.95.2 if there's a call from
> affected parties (Matt Thomas?  Richard Earnshaw?) 

Pulling in 2.96 would be, IMO, high risk and we would likely end up with a 
lot of local modifications to work around bugs which are bound to get 
fixed in fairly short order in the main tree.  I'd much rather see some 
effort going into making the netbsd source tree build with an out-of-the 
box gcc (rather than requiring the customized in-tree one -- the 
__KPRINTF_ATTRIBUTE__ is an excellent example).  Then I could do trial 
builds of the netbsd world with any gcc snapshot to test for regressions 
in the compiler.  If we are pro-active in making the standard releases 
work well on NetBSD, then we stand much more chance of making failures 
that matter to us matter to those doing GCC releases.

R.

PS I'd also like to see us reduce the number of in-tree mods that aren't 
propagated back into the master tree (for example, gcc-2.96 doesn't have a 
i386-netbsdelf port yet).