Ilya Zakharevich on Thu, 20 May 2004 22:53:02 +0200


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: gp: anal.c more bugs


On Thu, May 20, 2004 at 07:24:52AM -0700, Joe Christy wrote:
> Thanks v. much. I'm much calmer now and my program works well (modulo my 
> bugs). I was tearing my hair out trying to hack Makefile.sh to 
> accomplish an uninformed crude version of your fix.

I'm glad you overcame the problem; however, keep in mind that the
problem was not that the graphic subsystem was included, but that
*only half* of it was included.  Obviously, this lead to linking
problems.

> >*  an extra libparigp.so containing highlvl, graphical interfaces, and 
> >possibly other useful stuff from GP (all headers would need to be 
> >installed then).

I do not know any argument against such a separation...

> >If a program intends to be linked with -lparigp, it will provide a Makefile
> >(or 'configure') to check the pari.cfg file and find the proper -I/-L/-l
> >flags  [ mostly, the same ones used to compile gp ]

Where is it going to find this file?  I propose a small shell script
pari_config which will emit the contents of configuration; something like

  echo <<'EOE'
  foo="bar"
  parigp_l="-LXXX -lyyy"
  EOE

And put it on PATH alongside the gp executable.

Hope this helps,
Ilya