MessageTrans problems

[Jonathan Coxhead]

On 30 Oct 2001, at 11:58, Tony van der Hoff wrote:

> On 29 Oct 2001, in message <3BDD7833.478.D665C1 at localhost>, "Jonathan
> Coxhead" <jonathan at doves.demon.co.uk> wrote:
> 
> Carl Heterington wrote:
> 
> [snip]
> 
> > 
> > > PRM 3-755 Messagetrans_GSLookup returns a pointer to the result string in r2
> > > messagetrans_gs_lookup() doesn't return this.
> > 
> >    That one certainly looks like a bug ... I wonder why the exit condition
> > for  GSLookup is different from the one for Lookup?
> > 
> 
> I guess you just had a bad day ;-)

   Indeed!

> I'll make it behave like _Lookup. This introduces a compatibility issue, in that
> currently the non-x version of _GSLookup returns R3; it'll return R2 in future.
> As it's a bug-fix anyway, this shouldn't matter.

   Good!

   It also introduces an extra parameter to messagetrane_gslookup(), of course 
(the address of the extra output parameter), so code that uses it will fail to 
compile, and have to be fixed. This is exactly what we want. (If it was a 
"quiet change", that caused no compile errors and silently did the wrong thing, 
we could not accept the change.)

   So that's all OK.

        /|
 o o o (_|/
        /|
       (_/