Re: GNAT 3.11 worth installing

[dennison@telepath.com]

In article <36DFA064.B610D4D1@frqnet.de>,
  Andreas Winckler <andreas.winckler@frqnet.de> wrote:
>
> dennison@telepath.com wrote:
>
> > I hope you are looking into why this is happening, rather than just
> > throwing up your hands and blaming the compiler. Odds are there's
> > something non-kosher in the code. It could even be a timing problem
> > that the new compiler just causes to happen sooner!
>
> Yes, we tried to find out what's going on. But because developing for
> SUN with GNAT ist not our major emphasis we decided that further
> investigations are not worth the efforts. In this case using the old
> compiler is the better way ..... says my boss. ;-)

It sounds like your program is doing extensive tasking and interprocessor
communication. Given that I'd put good money on the problem being a timing
problem (aka: race condition) that exists on *all* your environments. It just
shows up most regularly on the SunOS target with the new gnat because the
timing environment with that setup happens to be just right.

Given that, it would definitely be worth your while to install the new gnat
and debug it there. The bug won't go away just becuase you wish it to.
Believe me, I've tried wishing. :-)

T.E.D.

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own