TEAM-ADA Archives

Team Ada: Ada Programming Language Advocacy


Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Condense Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
"Team Ada: Ada Advocacy Issues (83 & 95)" <[log in to unmask]>
Mike Bates 5598 Computer_Systems <[log in to unmask]>
Wed, 10 Nov 1999 04:51:59 -0600
Message from [log in to unmask] of "Tue, 09 Nov 99 21:46:12 -0500." <[log in to unmask]>
text/plain (52 lines)
On Tue, 9 Nov 1999 21:46:12 EST  [log in to unmask] wrote:
> Has anyone else been experiencing problems with the Windows NT host of Green
> Hills AdaMULTI?  Target processor is PowerPC (Radstone PPC-2A) card running
> VxWorks 5.3.1.

We are using AdaMULTI 1.8.9.B1 with an Intel Pentium target running
VxWorks 5.3.1.

> We've been plagued with a rash of little problems with the compiler (can't
> handle large files; can't handle too many separates; wierd code generation;
> exceptionally large executable sizes).

We have not seen any of those problems, perhaps because our work is
new development, and our architecture avoids large files and subunits,
but we have seen other little bugs.  The Booch Components for Ada 95
would not compile under 1.8.8, but everything seems to be fine now.
There were also problems with exception handling (especially with some
of the new exception features in Ada 95) and bit arrays.

To Green Hills' credit, they have been very responsive to our problem
reports, giving us prerelease patches as needed to keep us up and
running.  For our part, we've done our best to isolate problems before
reporting them, to help them get to the problem faster.

> Have we progressed in the new features of Ada95 where we are starting to
> exceed the capabilities of the compiler vendors?

I have been surprised at some of the problems.  I would have expected
a validation suite to catch most of them.  Some of the problems have
been in their back end, some appear to have been in the Intermetrics
front-end, and in one case what we thought was a Green Hills bug was
the GNAT compiler being more permissive than the reference manual.

> I'm not picking on Green Hills because I think they have a poor product.
> We're using it at work on multiple projects and everyone is a little
> "gun-shy" when they hear about problems.  The support folks from Green Hills
> have been great, and I'm sure our legacy code is probably taxing the compiler
> to it's limits.

It has taken us a while to get accustomed to both AdaMULTI and VxWorks
-- embedded development is a new experience for most of us -- but the
Green Hills product works well with the VxWorks target and (hooray)
has a command line interface, too.  And I'll agree that their support
folks have been very helpful.

Mike Bates                                 FlightSafety Simulation
Staff Engineer          [log in to unmask]  2700 North Hemlock Circle
Computer Systems Group  918-259-5598       Broken Arrow, Oklahoma 74012