Re: Weird error in ECLiPSe 5.5

From: Joachim Schimpf <j.schimpf_at_icparc.ic.ac.uk>
Date: Thu 30 Jan 2003 04:28:13 PM GMT
Message-ID: <3E39529D.22E1AF74@icparc.ic.ac.uk>
Ulrich Scholz wrote:
> 
> Secondly, I did not isolate the bug/behaviour. It is deeply hidden in my code
> and it would take some time to peel it out - as it did hours to bould down the
> consecutive constant assignments bug into a couple of lines,

It is often not necessary to cut down the code. It is no problem
when you send us a large program and data, as long as it allows
us to reproduce the problem in a reasonable time frame.


> and then you said
> "Oh, we already know this bug."  Maybe you know this too:
> 
> The following lines of code:
> 
>           writeln(["X0", PDDL_Variable]),
>           The_Type2 = type with [ id: ID, objects: Objects,
>                                   name: TypeName ],true,
>           printf("xxx", []),
>           writeln(["X1", PDDL_Variable]),
> 
> is traced like this
> 
> S (18305) 11 CALL  writeln(["X0", V_5643])   %> creep
> [X0, V]
> S (18305) 11 EXIT  writeln(["X0", V_5643])   %> creep
> S (18306) 11 CALL  printf(type(t_1, [], "t_1", [object, t__3], [], [], [t_1, t_1x, ...]), [])   %> creep
> S (18306) 11 FAIL  printf(..., ...)   %>
> 
> It also fails without trace.  BTW, the type(...) in the printf is the one
> bound to The_Type2


The silent failure of printf was recently fixed (version > 5.5#52).

The passing of the wrong argument to printf looks like a compiler
bug similar to the one you experienced earlier. I cannot say
more from just looking at a code fragment.


-- 
 Joachim Schimpf              /             phone: +44 20 7594 8187
 IC-Parc                     /      mailto:J.Schimpf@imperial.ac.uk
 Imperial College London    /    http://www.icparc.ic.ac.uk/eclipse
Received on Thu Jan 30 16:28:28 2003

This archive was generated by hypermail 2.1.8 : Wed 16 Nov 2005 06:08:19 PM GMT GMT