Ilya Zakharevich on Fri, 09 Apr 2004 10:37:21 +0200


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

Re: gp: anal.c more bugs


On Thu, Apr 08, 2004 at 11:50:50PM +0200, Bill Allombert wrote:
> > > gcc main.c -lpari
> > 
> > This is a laudable goal; unfortunately, it has no chance to work
> > portably.  At least you need -lm.

> Let say gcc main.c -lpari -lm then.
> -lm is independent of the Configure options.

Is -lm always present in ANSI C?  I remember that it is not always
present (at least if you include pre-ANSI era).

> > > collect2: ld returned 1 exit status
> > 
> > *These* are because one plotting module is omited.  Adding it and
> > -lX11 should fix things.  AFAIU, on most architectures one can compile
> > libpari.dll so that -lX11 is not needed on the main application.

> But what if PARI was configured with an alternative graphic engine
> (fltk,qt,etc...) ? The program will be linked with that library, and
> will not work with libpari.so.x build with a different graphic engine.

Why?  If it links, it works.  If the .dll file is linked "correctly",
you do not need any additional -l on the executable: the loader knows
that if you load pari.dll, you need to load X11.dll (or whatever is
needed).

> Anyway we need to remove plotport.o from libpari until the issue above
> is sorted out.

I agree on one count only: it makes no sense to include plotport.o and
not include plotARCH.o.

Hope this helps,
Ilya