[ a rather harsh-toned argument from me deleted ]

   No I am not an idiot contrary to popular belief; when I say it does not
   compile, I mean that the resulting binary core-dumps even after I fix
   all of the for loop index variables. I know that this is a compiler bug, 
   because the code compiles and runs with 2.6.2 after my fixes...

   Have you compiled a working version on groff on the sparc with 2.7.0?

   christos

I want to apologize for my tone, it was not clear that you knew about
the for-scope change to me.  I have seen several complaints on this
very topic and I was getting tired of explining it.  Peace?  I don't
state GCC 2.7.0 is bug-free, no more than any other piece of
non-trivial software.  However, groff could still be at fault, even
though 2.6.2-compiled versions worked.  Sometimes application bugs
doesn't harm until codegen strategies are
changed... (e.g. destructions of temporaries, used before initialized
vars etc.)  But I do prefer to believe it is a GCC bug, in cases like
this, just like you do.

Anyway, I really want to apologize for my tone.

Niklas

PS No I haven't ever compiled groff on a sparc...  however several of
   the G++ maintainers do work on sparcs, and they do test groff
   before a release.  Are you using special flags?

Niklas Hallqvist	Phone: +46-(0)31-40 75 00
Applitron Datasystem	Fax:   +46-(0)31-83 39 50
Molndalsvagen 95	Email: niklas@appli.se
S-412 63  GOTEBORG	WWW:   Here
Sweden