🚀 go-pugleaf

I am following up on my original posting.

In article <ISHIKAWA.92Oct21170526@ds5200.personal-media.co.jp> ishikawa@ds5200.personal-media.co.jp (Chiaki Ishikawa) writes:

   Hi.

   We have installed GCC 2.2.2 on our Sun3, SunOS 4.0.3 machine.

   When I tried to compile EMACS using gcc, I found out that fixincludes
   does NOT handle the brain-damaged Sun3 headers (non-ANSI handling of
   arguments in _IO, etc.) anymore. It did handle these problems well in
   2.2.1.

   Does anyone have fix for these problems???

It turns out it is a buggy SED(!) that came with SUN Microsystems JLE
(Japanese Language Environment) which was wreaking havoc :-(.

Someone found a temporay replacement of SED (I am not sure if he used
gnu-emacs SED or not), and now fixincludes seems to work.

I checked the output of, say, Gnu-Emacs sysdeps.c with

gcc -E sysdep.c | grep -v '^#' | grep -v '^$' | grep -v '^ $' > /tmp/g.c
cc -E sysdep.c | grep -v '^#' | grep -v '^$' | grep -v '^ $' > /tmp/o.c
diff -wc /tmp/g.c /tmp/o.c

And found that now it produces identical output (module some
whitespace differences).

I have been telling SUN people to install Japanese modification stuff
on their computers in U.S.A. so that bugs in the Japanese localization
will be caught earlier. Tsk, tsk...

Thank you again for your attention, and sorry that it was non-GNU
software's bug.



Regards,