Re: [Gm2] gm2 segfault (checked with GPM and XDS)

[Gaius Mulley]

Andreas Fischlin <address@hidden> writes:

> That's a correct diagnosis and the other compilers err, if they
> tolerate such a constant declaration.
> 
> Andreas
> 
> John O Goyo wrote:
> > Gaius Mulley wrote:
> >> Dmitry Shatrov <address@hidden> writes:
> > [...]
> >>>>> MODULE test;
> >>>>>
> >>>>> VAR A : CARDINAL;
> >>>>>
> >>>>> CONST B = A;
> >>>>>
> >>>>> END test.
> > [...]
> >>> I have a program for JPI Modula-2 in the front of me, and it looks like
> >>> in JPI such assignment is OK. Is it a deviation from PIM and/or ISO?
> >>>
> >>> Regards,
> >>> Dmitry
> > [...]
> >
> > I tried compiling Dmitry's example with the GPM and XDS compilers
> > (both ISO).
> >
> > Here is GPM's response:
> > =================================================
> >     5   CONST B = A;
> >  ****   ..........^ SemanticError #  208
> >  **** 208 Identifier is not a constant ****
> > =================================================
> >
> >
> > Here is XDS's response:
> > =================================================
> > XDS Modula-2 v2.40 [x86, v1.50] - build 10.05.2005
> >
> > Compiling "test.mod"
> >
> > * [test.mod 5.11 E087]
> > * expression should be constant
> > CONST B = $A;
> > errors  1, no warnings, lines    7, time  0.05
> > =================================================
> >
> > I suspect that we are seeing a JPI extension.
> > (I have the JPI manuals at home and will check over
> > the weekend.)
> >

Hi,

thanks to all for testing this code on other platforms,

regards,
Gaius