compile and install bugs on irix

[Jeffrey A Law]

  In message < orzp1r1h8o.fsf@saci.lsd.dcc.unicamp.br >you write:
  > > I don't know.  Presumably libz is needed for something on some systems.  
  > > Otherwise the test wouldn't be there to start with.
  > 
  > > Simply deleting it seems wrong.
  > 
  > Nope, we don't need libz for our builds.  Only install-info is
  > zlib-aware, in order to be able to read compressed info and dir files,
  > but we don't use the just-built install-info in gcc/Makefile and, even 
  > if we did, we're not creating compressed info files, and the Makefile
  > rule checks for the existence of $(infodir)/dir, not
  > $(infodir)/dir.gz, so we're quite safe.
You should talk to Gerald since he explicitly re-enabled install-info
presumbly so that it could be used somewhere in our tree.


  > Moreover, we could stop building install-info, since we don't (appear
  > to) use it, or start to use it instead of wasting CPU cycles on a
  > useless build.
As you know, the plan is to totally zap the whole texinfo tree after
gcc-2.95.  Until then I want to make the absolute minimal changes necessary
to this stuff.

What I do not understand (I did not follow the whole thread) is why other
folks are not having this problem.  What is different about the particular
system that's choking on libz right now?

jeff