Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!wuarchive!uunet!mcsun!ukc!warwick!kingpol!cs_b144
From: cs_b144@ux.kingston.ac.uk (Ian Stickland)
Newsgroups: comp.os.os2.programmer
Subject: Re: problems with the IBM 1.3 toolkit and MSC6.0
Message-ID: <1991Jun12.103036.2600@kingston.ac.uk>
Date: 12 Jun 91 10:30:36 GMT
References: <9106110640.AA12807@grape.ecs.clarkson.edu>
Sender: news@kingston.ac.uk (Network News)
Reply-To: cs_b144@ux.kingston.ac.uk (Ian Stickland)
Organization: Kingston Polytechnic
Lines: 42
Originator: cs_b144@gamma
Nntp-Posting-Host: gamma


Ahh!! This reminds me!! When I first got my hands on the IBM 1.2 toolkit
I tried to compile one of the graphics samples using MS C6.0. I can't 
remember if it was graphic1 or graphic2, but it's the one that allows you
to rearrange the shapes into two squares. Anyway it caused the complier to
crash with an internal error, always in the same place. Anyone else had 
this problem??

When I contacted Microsoft about this they didn't want to know and tried 
palming it off as an IBM problem which I don't think it is. As it was only
a sample program though I couldn't be bothered to sit through hours of
Jean Michelle Jarre and then try to explain the problem to someone who 
hadn't even heard of OS/2 let alone the IBM toolkit. 

I must try compiling the sample in the 1.3 toolkit. The result should be 
similar I'd imagine, as the program probably hasn't changed much.

I'd be interested to know if anyone else can duplicate this problem, including
anyone from Microsoft who might be listening (whilst taking a break from the
on-line music).

Re the original problem, if you use the IBM toolkit you will need to delete
the MS C OS/2 include files which get dumped in the INCLUDE directory. These
will conflict with the IBM 1.3 files. Also another source of error msgs is that
Microsoft took it upon themselves to completely change the definition of NULL 
from zero to a void pointer to null which causes code that compiled without
any warnings under previous versions of the compiler to generate pages of the
stuff under C6. This is why the IBM toolkit samples generate so many warnings
(or at least I HOPE that is the reason.) IF you have the patience a liberal 
dose of casting will get rid of the warnings or you can just turn them off.
the samples still work, all except the one that causes the compiler to give up.

As a thought, with all these Microsoft adverts about 'The Hard Way' and the 
'Microsoft way' just remember who invented the hard way... Yep good old
MICROSOFT....   What more needs to be said other than I really hope Borland
come up with the goods  on the 32bit compiler front, and that means FULLY
re-entrant libraries, so then we can ditch MS completely.


                     Another satisfied MS customer...
          
                            Cheers, Ian Stickland.