bug Ada Verdix SUN : segmentation fault at compil time

bug Ada Verdix SUN : segmentation fault at compil time

[compte Simulation de Contexte Adaptee Localement a l'Application]

We have a non easy reproductible bug with Ada Verdix on SUN :

This bug can be :

- segmentation fault at compil time ,

- this segmentation fault may disapear by a small change in the code 
but in that case it becomes a "CONSTRAINT ERROR" at execution !

Somebody in the world has an idee ??

Re: bug Ada Verdix SUN : segmentation fault at compil time

[Esther Lumsdon]

scala@stna.dgac.fr (compte Simulation de Contexte Adaptee Localement a l'Application) writes:

>We have a non easy reproductible bug with Ada Verdix on SUN :

>This bug can be :

>- segmentation fault at compil time ,

>- this segmentation fault may disapear by a small change in the code 
>but in that case it becomes a "CONSTRAINT ERROR" at execution !

>Somebody in the world has an idee ??

Since you've indicated above that this problem is not easy to diagnose, you 
should do one of two things:

1)  report this to Olivier Roubine, Verdix' representative in Boulogne, France.
    All of the French-speaking people who work for/with Verdix are in
    Europe.
2)  email a sufficient English description/testcase to report@verdix.com, if 
    you don't have a current support contract.  Is there a record declaration
    with a default component initialization?  Is there a generic instantiation?
    What version of SunAda are you using?  If you compile the unit with 
    "ada -ev"  does the error message get attached to a particular
    source line?
This sounds too complex to diagnose from your brief description.
-- 
-- Esther Lumsdon, not speaking for Verdix.   esther@verdix.com
"It's time to cut bait and talk turkey.  It takes 2 snakes to cross a
puddle. You have to bale hay while the tractor is warm."
  ---- either H. Ross Perot or Dave Barry

bug Ada Verdix SUN : segmentation fault at compil time

[SAHARBAUGH]

I tried to send this direct to the poster but apparently

a computer in France is taking Easter Sunday off.

I was once plaqued by "segmentation fault" on a Verdix compiler
on a Unix machine.  After much work I traced it down to the fact
that the code generation code assumed that a math co-processor was
present in the machine.  My machine did not have a math co-processor
so I got segmentaion errors.  I recall that I used a case statement
to isolate the problem.  When the code generator genertaed the code
for computing the address for the case statement branching it asssumed
the presence of the math co-processor.
That was 6 years ago and not on a Sun computer.  Let me know what
you find.
sam harbaugh  SAHARBAUGH@ROO.FIT.EDU