issues compiling octave-2.1.69 on FreeBSD 5.4

[Joan Picanyol i Puig]

* John W. Eaton <jwe at bevo.che.wisc.edu> [20050422 22:05]:
> On 22-Apr-2005, Sebastian Schulze Struchtrup <seb at struchtrup.com> wrote:
> 
> | > On 22-Apr-2005, Sebastian Schulze Struchtrup <seb at struchtrup.com> wrote:
> | >
> | > | I don't know if it's relevant and I don't know why, but octave compiles
> | > | and runs without problems on a recent 6.0-CURRENT.
> | > | Tested a few days ago with the latest port (just removed the BROKEN
> | > line)
> | >
> | > Does that system have the extern "C" wrapper around the contents of
> | > the system fnmatch.h and glob.h files?
> | 
> | It has, but it's the same as on 5.4
> | 
> | All declarations are wrapped by __BEGIN_DECLS/__END_DECLS which will
> | expand to extern "C" {  } in case of C++ or to nothing in case of C
> | 
> | I also have a 5.3 box available, I can test the build there during the
> | weekend.
> | I assume it should be the same issue on 5.3 and 5.4.
> 
> OK, then I am out of ideas.  Someone who has access to a system where
> there is a problem will have to debug it.

I've started a binary search to find out when octave broke, but I need
some help. Apparently the breakage happened somewhere between 2.1.63 and
2.1.69. Unfortunately I haven't found the tarballs on-line, and I'm a
little bit lost regarding octave's CVS branches. I've cheked out
ss-2-1-67 and will report that, but the process is tedious and long, so
any guidance will be most welcome (particular dates to try, or the old
tarballs).

tks
-- 
pica