Re: [Gm2] latest fixes to gm2

[Gaius Mulley]

Iztok Kobal <address@hidden> writes:

> Gaius Mulley wrote:
> >Izo writes:
> >>Gaius Mulley wrote:
> >>
> >>>  *  the -Wunbounded-by-reference and -Wverbose-unbounded.
> >>>     Here is the extract from gm2.texi
> >>>
> >>Gaius, I do not understand really the reason for the use of this
> >>option. What situation can it be useful in ?
> >
> >This could (depending upon how code is written) improve performance
> >significantly.
> >
> The interfacing to C (defining the procedure in DEFINITION MODULE
> ["C"]...) does it still by default ?

yes true, the DEFINITION MODULE FOR "C" mechanism converts

   func (foo: ARRAY OF CHAR) ;

into

   func (char *foo)

(and no copy of the contents of the array foo is made)

> Namely, this could also be the point where my code could break since
> both the StonyBrook and GPM (which my M2 code was targeted at the time
> when written) regard the PROCEDURE foo(s : ARRAY OF CHAR) : WHATEVER
> as WHATEVER foo (char* s), without further
> local-to-foo-stack-frame-copy, of course, only when interfacing to
> C.

gm2 should behave the same, it only takes a copy when Modula-2 is
calling Modula-2 and when we have:

   PROCEURE func (foo: ARRAY OF sometype) ;
   BEGIN

   END func ;

no copy is made when we have:

   PROCEURE func (VAR foo: ARRAY OF sometype) ;
   BEGIN

   END func ;

> >>1. What about the .so support ? Is there provided the main body for
> >>_fini and _init ?
> >>
> >
> >still to do.. I can push this as a high priority fix if need be..
> >
> My project is now before the 1st milestone which implies pilot
> implementation of the device and real-world testing. When development
> cycle continues, which I see it somewhere in autumn, I could switch
> from XDS to GM2 if the later is stable and easily produces shared
> objects.

this sounds fine, my target is to finish the ISO libraries in the
summer. This should allow time to implement shared libraries
concurrently and to complete them by the autumn.

Gaius