Andy Stubbs on Wed, 20 May 1998 15:19:21 +0200

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

shared libraries: namespace problem / an unmoveable object meets an irresistable force / Java bindings for PARI


I've been experiencing a rather intractible problem that I hope someone
can explain. I'm surprised nobody has had this problem before (unless I'm
being really dense): when both and are in memory as
shared libraries, _any_ reference to pari's err() function (including
references made internally by is passed to libc's err() 
function... so the kind of error caused by flisexpr("2+[2]") typically
results in: 

myapp: Scalar: No such file or directory
bash:~$ echo $?

This is not exactly desirable behaviour... I tried to solve it using
dlopen(), but still got nowhere. I have a work-around; basically I patched
PARI so that every instance of err() was replaced by pari_err(), but this
is _far_ from being a good solution. If someone who knows something about
shared libraries (specifically on Linux 2.0.30, but _any_ hints would be a
bonus) can please tell me how to make sure that pari's err() function is
called instead of libc's then I'd be extremely grateful.

How did I bump into this problem? Well, I notice from the UserGuide that
there seem to be no Java bindings for PARI... Seeing a gap in the market I
decided to corner a niche and write my own ;) The current status is 'too
soon to tell', but it contains a few cute features and is quite usable. 
Most of the low-level interface exists (I always love the low level
stuff), but the JNI layer mapping Java functions to pari routines is very
spartan at the moment, and, due to my rather ad-hoc approach, the top
layer of Java Objects is also likely to change radically. Here's a little
taster of a fragment of code, if anyone's interested: 

  PariVariable x; /* x is a PARI variable */
  PariGen y;      /* y is a generic PARI GEN object */

  x = new PariVariable( "t" ); /* just to confuse everyone */
  /* actually, x.polx() below can be shortened to just x */
  y = pMath.gsqr( pMath.gadd( x.polx(), x.polun() ) );

  /* t^2 + 2*t + 1 + t = t^2 + 3*t + 1 */
  System.out.println( y + " + " + x + " = " + pMath.gadd(x,y) );

  /* d(t^2 + 2*t + 1)/dt = 2*t + 2 */
  System.out.println( "d(" +y + ")/d" + x + " = " + pMath.deriv(y,x) );

  y = null; /* space on PARI stack will also be cleared */

But, this aside, I hope someone can shed light on my problem with shared



Andy Stubbs                              +44 151 794 4045