ar script and bison errors

Peter Naulls peter at
Sat Feb 23 10:55:53 PST 2002

In message <Pine.LNX.4.44.0202231827190.1020-200000 at>
          Theo Markettos <atm26 at> wrote:

> Lots of packages out there seem to call ar directly.  While libfile will
> do the job, it's not a direct alias as the options are different.  Has
> anyone written a script to convert ar to libfile options?  I've written
> the attached Basic program, which will do the job using Brandy, but
> something more portable (in shell or perl) is probably better -
> unfortunately I don't know enough of either of those to do it properly.
> I've attached the above in case it's of use to anyone, or someone has
> something better.

Ah, good.  There's also the matter of ranlib of course.  Probably an
empty program will be fine.   I can redo this in Bash easily enough.

On the same note, I've got plans to do a GCCSDK versin of xmkmf, which
will alleviate issues building some X programs.

> Also, building release_2_95 (as of 18:15 today), I get:
> make -C objc
> make[3]: Entering directory `/home/atm26/extra/gccsdk/gcc/gcc/objc'
> bison objc-parse.y -o objc-parse.c
> objc-parse.y contains 72 shift/reduce conflicts.
> expected 66 shift/reduce conflicts
> make[3]: *** [objc-parse.c] Error 1
> Now I know no yacc at all, so wouldn't know whether this is spurious or
> not.  I don't actually care about building the objective C compiler anyway
> - is there an easy way to exclude it?

I know enough to know that's not a good error to have, but not much more
than that.


 Peter Naulls - peter at
 RISC OS Projects Initiative  -
 Java for RISC OS and ARM     -
 Debian Linux on RiscPCs      -

More information about the gcc mailing list