[comp.lang.c++] Zortech C++ for Mac still shipping with incomplete MPW

nagle@well.sf.ca.us (John Nagle) (05/24/91)

     Zortech's corrected release V2.1R1 reached me today.  It can be
distinguished by a date of 5/20/91 on the document "Installation";
disk markings are identical to the previous defective version.

     The MPW libraries missing from the original release have now been
supplied.  (Confusingly, they're on the compiler disks, not the MPW disks.)
The C++ libraries seem to be comparable to Apple's, although "iostream.h"
is missing from the Zortech release.

     At least the following parts of MPW are still missing:

	- The Installer
	- Cmarker
	- Projector
	- MacsBug
	- SADE
	- ResEdit

     The compiler seems to be in good shape (although I have one program
that crashes the compiler with an assertion failure).  All other correct
programs compiled and ran properly; the only changes required were
where Zortech detected some error Apple's CFRONT didn't.

     This may be a good product once they get their act together.  But
I'd wait a few weeks before ordering.  

					John Nagle

nagle@well.sf.ca.us (John Nagle) (05/25/91)

I previously wrote:

>     At least the following parts of MPW are still missing:

>	- The Installer
>	- Cmarker
>	- Projector
>	- MacsBug
>	- SADE
>	- ResEdit

     This needs to be corrected.  Projector is part of the MPW shell, and
is present.  Apple (d/b/a APDA) considers MacsBug, SADE, and ResEdit
to now be separate products from MPW, even though the MPW manuals
indicate that they are part of MPW.  APDA is only shipping MPW 3.1,
because of some system 7 compatability problems, so they aren't sure
if CMarker is part of MPW.  The Installer for MPW is definitely still
missing, but it's not essential to installation.

     Zortech says that their file "streams.hpp" can be used in place of
the missing "iostream.h".  

     On the compiler itself, the code looks good, but not great.  All
the heavy global optimizations documented in the manual seem to be
working.  "inline", though, is apparently ineffective in this release.
Code is always generated for the 68000/no coprocessor case; turning on
68020 or math coprocessor code generation with the ZTC Build File 
maker produces a warning note that these aren't implemented yet.

     So far, everything compiled has run without error, including some
large class libraries.  There are rough edges, but Zortech has done the
hard parts right.  I look forward to the next release.  Zortech has a good
track record on fixing bugs quickly.

					John Nagle