re: pcc and build system

[matthew green <mrg%eterna.com.au@localhost>]

   On Tue, Jul 08, 2008 at 11:43:44AM +1000, Gregory McGarry wrote:
   > I agree.  Most occurrences of HAVE_GCC are actually testing for use of 
   > compiler warning flags, like -Wno-pointer-sign.  Maybe there is a better 
   > mechanism we can use to solve this?
   
   A good starting point would be fixing the code that chokes on
   -Wpointer-sign :-)
   
   I just arbitrarily checked sbin/newfs_sysvbfs and it doesn't appear
   there's any reason to keep the -Wno-pointer-sign; it compiles cleanly
   without it using both -fsigned-char and -funsigned-char.
   
   Maybe after 5.0 is branched someone (probably me) should go through
   that list and clean out the special cases that are no longer required.


that would be cool.  doc/HACKS should document these as well as grep -r.


.mrg.