RFH stands for ‘Request for help’;

 

I discussed a while ago with Daniel Jacobowitz

the possibility of adding some basic tests for pascal language support.

 

  My main problem is that I never used anything like expect of autoconf before.

 

  The first step is to add a new directory called gdb.pascal in gdb/testsuite directory.

But I am already lost here:

  I saw in the sources that I should add reference to this directory in

configure.ac but if I run autoconf after the configure script is changed a lot.

I read in the Contribute file that we do not need to commit the regenerated configure file,

but if we want to check the effect before committing, the fact that

the regenerated configure is different from the one automatically

regenerated on the CVS repository might lead to unnoticed problems.

 

  The second point is that there are mainly two free pascal compiler:

GNU pascal, (gpc executable)

and

Free Pascal, (fpc executable)

  The pascal testsuite will probably need to be handled

a little bit differently depending on which pascal compiler is used

(the command line options are completely different for the two compilers).

 

  Furthermore, none of these two compilers are included by default

on all platform, as can be expected for GCC.

  This means that we need to find a way to check for the presence of

a valid pascal compiler. There might be some equivalent code

for fortran or ada compilers, but these seem to be

special switches of gcc whereas both gpc and fpc.

 

  So the main question is where should I include code

to test for the presence of a valid pascal compiler

and how to only run the tests in gdb.pascal

if a valid compiler was found.

 

Pierre Muller

Maintainer of GDB pascal language support.

 

PS, specifically for gpc mailinglist:

is there a unified and version independent symbol name for

the main procedure in the main program?

Using ‘break main’

stops at some  <implicit code>

The main procedure seems to be called

_p__M0_main_program

but this is for gpc 20060325 based on gcc 3.4.6 for Ubuntu

Is this symbol still valid?

Is it supposed to stay unchanged in future releases?

To be able to break at the first line of the main procedure

seems like a minimal requirement that is not met for GPC

and GDB right now…