At 4:49 AM +0200 on 6/2/99, M. Uli Kusterer wrote:
>Anthony,
>
> I'm working on getting HyperInterp to compile under CodeWarrior. First thing:
>
>OTVar.h line 342:
> Make it 255UL to indicate it's also an unsigned long, or typecast
>the number to an unsigned long. Else I get an error that ( long, int )
>doesn't match (type T, type T).
Grr... maybe. Hmmm... implicit type conversions vs. templates. Interesting
match.
I'll change it.
>Same applies to the operator overloads in
>OTVar.h, which should e.g. do (*this)+=1L; for the increment operator to
>force the appropriate operator to be used. Else CW can't decide between
>long and double for some reason.
Well, then CodeWarrior is just conducting a war on standard C++.
>
>OTVar.h line 227:
> Make the conditionals that check for _MRC__ < 0x0410 also check for
>!__MWERKS__ since somehow metrowerks ends up inside that conditional even
>though it only compiles the other.
How about:
#if defined(__MRC__) && __MRC__ < 0x0410
>
>With these changes I got it to compile up to TokenizerTest,
Could you send me a copy of all the files you changed? That way I can
compare them and patch my copies.
>which gave me
>an error because it needs a file called "HandleAppender.h" could you send
>me that and any other files it depends on?
Oops. I'll send that out.
>A look at the makefile shows
>that you probably wrote it yourself, as it's in a folder named "AnthonyLib".
Good guess :)
>
> BTW, here's a list of what should be added to HyperInterp and how in my
>opinion:
>
> objects
> properties
> message handler calls
> handlers (i.e. on/end and function/end)
Agreed.
>
>It should pass these to external handling functions into which other
>developers can hook in, probably to a C++ instance that is passed in on
>startup. What I have in mind is something along the lines of what Velocity
>had in this regard.
I'll have to look at it.
