TTY debugger and the portray mechanism.

Hi folks, 

2 bugs here. The first is that in 5.5 and later the portray of hash tables
has not been updated to cope with the fact that the implementing struct has
two extra arguments. 

--------

ECLiPSe Constraint Logic Programming System [kernel]
Copyright Imperial College London and ICL
Certain libraries copyright Parc Technologies Ltd
GMP library copyright Free Software Foundation
Version 5.4 #41, Tue Jul 16 00:14 2002
[eclipse 1]: lib(hash).
hash.eco   loaded traceable 0 bytes in 0.01 seconds

Yes (0.01s cpu)
[eclipse 2]: hash_create(H).

H = hash(4, 0, [])
Yes (0.00s cpu)
[eclipse 3]: 
bye
testserver1: /usr/local/eclipse/5.5/bin/i386_linux/eclipse
ECLiPSe Constraint Logic Programming System [kernel]
Copyright Imperial College London and ICL
Certain libraries copyright Parc Technologies Ltd
GMP library copyright Free Software Foundation
Version 5.5 #63, Tue Feb 11 00:22 2003
[eclipse 1]: lib(hash).
hash.eco   loaded traceable 0 bytes in 0.02 seconds

Yes (0.02s cpu)
[eclipse 2]: hash_create(H).

H = hash_table(4, 0, []([], [], [], []), [], [])
Yes (0.00s cpu)
[eclipse 3]: 

--------

I've run into the same kind of problem when I've defined my own portrays.
You add arguments to the implementing structure, but forget to update the
arity in the :- export(portray ... line. It would be nice if at compile time
this line could look up the arity of the struct just given the functor.  

The second bug is that it is unpredictable in the TTY debugger whether
compound terms get displayed using their portray predicate definitions or
not. This seems to be an interaction between the TTY debugger, the module
system and the portray mechanism. 
Here's an example.

--------

Suppose I have two modules, test_module defined as:

%%%%%%%%%%
:- module(test_module).

:- export test_length/2.

test_length(List, Length):-
        length(List, Length).

%%%%%%%%%%%%

And secondly test_module_b, defined as:

%%%%%%%%%%%

:- module(test_module_b).

:- local struct(test_struct_impl(
                                    arg1,
                                    arg2, 
                                    arg3
                                )
               ).

:- export(portray(test_struct_impl/3, test_portray/2, [])).

:- export(test_portray/2).

test_portray(test_struct_impl with [arg1:X], test(X)).

:- export(test_create/4).

test_create(X,Y,Z, test_struct_impl with [arg1:X, arg2:Y, arg3:Z]).
        
%%%%%%%%%%%%%

Then you would expect that whenever a test_struct_impl/3 crops up in a
trace, we only see test/1. But what actually happens (on 5.6 prerelease) is
as follows:

testserver1: /usr/local/eclipse/5.6/bin/i386_linux/eclipse
ECLiPSe Constraint Logic Programming System [kernel]
Copyright Imperial College London and ICL
Certain libraries copyright Parc Technologies Ltd
GMP library copyright Free Software Foundation
Version 5.6 #28, Mon Apr 21 00:13 2003
[eclipse 1]: use_module(test_module).
test_module.ecl compiled traceable 124 bytes in 0.01 seconds

Yes (0.01s cpu)
[eclipse 2]: use_module(test_module_b).
test_module_b.ecl compiled traceable 196 bytes in 0.01 seconds

Yes (0.01s cpu)
[eclipse 3]: test_create(1, 2, 3, T).

T = test(1)
Yes (0.00s cpu)
[eclipse 4]: trace.
Debugger switched on - creep mode
[eclipse 5]: test_create(1, 2, 3, T).
  (1) 1 CALL  test_create(1, 2, 3, T)   %> skip
  (1) 1 EXIT  test_create(1, 2, 3, test(1))   %> creep

T = test(1)
Yes (0.00s cpu)
[eclipse 6]: test_create(1, 2, 3, T), test_length([T], X).
  (1) 1 CALL  test_create(1, 2, 3, T)   %> creep
  (1) 1 EXIT  test_create(1, 2, 3, test(1))   %> creep
  (2) 1 CALL  test_length([test(1)], X)   %> creep
  (3) 2 CALL  length([test_struct_impl(1, 2, 3)], X)   %> skip
  (3) 2 EXIT  length([test_struct_impl(1, 2, 3)], 1)   %> creep
  (2) 1 EXIT  test_length([test(1)], 1)   %> creep

T = test(1)
X = 1
Yes (0.00s cpu)
[eclipse 7]: 
            
-----------

notice how when tracing the length goal the portray does not get applied.
But if you just do a normal length call, it works: 

----------

[eclipse 7]: test_create(1, 2, 3, T), length([T], X).
  (1) 1 CALL  test_create(1, 2, 3, T)   %> creep
  (1) 1 EXIT  test_create(1, 2, 3, test(1))   %> skip
  (2) 1 CALL  length([test(1)], X)   %> skip
  (2) 1 EXIT  length([test(1)], 1)   %> creep

T = test(1)
X = 1
Yes (0.00s cpu)
[eclipse 8]:          

----------

Maybe this is because the portray exported by test_module_b is not visible
in test_module. But this is to be expected, since test_module is often a
low-level utility, which wouldn't be expected to know about the
peculiarities of some application-specific data module (test_module_b).

cheers, 

josh

Josh Singer
Software Developer, Parc Technologies Limited
Tel. +44 (0)207 261 4067
josh.singer@parc-technologies.com
http://www.parc-technologies.com

This e-mail message is for the sole use of the intended recipient(s)
-its contents are the property of Parc Technologies Limited (or its
licensors) and are confidential. Please do not copy, review, use
(except for the intended purposes), disclose or distribute the e-mail
or its contents (or allow anyone else to do so) without our prior
permission. Parc Technologies Limited does not guarantee that this
e-mail has not been intercepted and amended nor that it is
virus-free. You should carry out your own virus checks before opening
any attachment.  Any opinions expressed in this e-mail message are
those of the author and not necessarily Parc Technologies Limited.