Weird error using SunAda

Weird error using SunAda

[M. Alvarez-Campana]

Hi,

Maybe this is not the adequate newsgroup, but I hope someone
has had a similar problem. I'm using SunAda to compile a
program that had been successfully compiled before on VMS.
When I run the program I get a 'Bus error (core dumped)' message
and the program aborts. Using the debugger I get:

a.db example
>r
example 
"Bus error" [10]
--> Hardware alignment error
 stopped 1 instruction after "/tmp/topnet-sublib/xs_kernel.a":130 in deposit'3
 130              if TM /= 0.0 then


Any help would be very appreciated.

Manuel
--
.-------------------------------------------------------------------------.
| Manuel Alvarez-Campana Fernandez-Corredor   | Phone: +34 1 3367328      |
| Dpto. Ingenieria de Sistemas Telematicos    |     or +34 1 5495700 x366 |
| E.T.S. Ingenieros de Telecomunicacion       | Fax: +34 1 5432077        |
| Ciudad Universitaria S/N                    | E-mail: mac@dit.upm.es    |
| 28040 - MADRID                              |                           |
`---------------------------------------------^---------------------------'

Re: Weird error using SunAda

[David Tannen]

Manuel,

Yes you have found the right group to ask your question.  Sorry I do
not have answer.

But I have a statement to the other members of this group:

	Please STOP all the _flamewars_.  When someone has to ask
	if this is the right group to ask a technical Ada ?, we have
	a problem.  Let us all agree to ignore flamebait and then it 
	will go away.

Thank you,
---
David Tannen
tannend@source.asset.com
tannen@tigger.geg.mot.com
----------------------------------------------------------------------
-- "Dependance on wizardry to mitigate the fundamental limitations
--  of software is called 'hacking'."  Grady Booch.
--
-- Writing a MS-Windows application often requires hacking.
----------------------------------------------------------------------

Re: Weird error using SunAda

[agate!howland.reston.ans.net!usc!cs.utexas.edu!tamsun.tamu.edu!news.utdal]

In article <MAC.93Jul9153020@caribe.dit.upm.es> mac@dit.upm.es (M. Alvarez-Camp
ana) writes:
>Maybe this is not the adequate newsgroup, but I hope someone
>has had a similar problem. I'm using SunAda to compile a
>program that had been successfully compiled before on VMS.
>When I run the program I get a 'Bus error (core dumped)' message
>and the program aborts. Using the debugger I get:
>
>a.db example
>>r
>example 
>"Bus error" [10]
>--> Hardware alignment error
> stopped 1 instruction after "/tmp/topnet-sublib/xs_kernel.a":130 in deposit'3
> 130              if TM /= 0.0 then

It appears that TM is not stored on a word boundary. In visual mode, 
disassemble the code and see if the SIGBUS comes on the load instruction.
If it does, do the following on the a.db command line 'tm:a' - that should
print the address of tm. If it is not a word-boundary, and you did not force
it to be there w/ a rep-spec, call your SunAda (Verdix) rep.



-- 
Boris Pelakh		Ada Project Leader          pelakh@convex.com
		     Convex Computer Corporation
"If winning isn't important, why keep score ?"	-- Lt. Worf, Star Trek TNG.
			

Re: Weird error using SunAda

[Robert I. Eachus]

In article <MAC.93Jul9153020@caribe.dit.upm.es> mac@dit.upm.es (M. Alvarez-Camp
ana) writes:

   When I run the program I get a 'Bus error (core dumped)' message
   and the program aborts. Using the debugger I get:

   a.db example
   >r
   example 
   "Bus error" [10]
   --> Hardware alignment error
    stopped 1 instruction after "/tmp/topnet-sublib/xs_kernel.a":130 in deposit
'3
    130              if TM /= 0.0 then

   Normally I would say it's time to exercise your compiler support
contract, but there may be enough information here to diagnose the
problem...

   The SPARC architecture has fairly severe memory alignment
requirements.  In particular requiring 64-bit values to be doubleword
(64-bit/8-byte) aligned in memory.  Thirty-two bit values must be word
aligned, 16-bit values half-word aligned.  The VAX on the other hand,
is willing to reference any object at any byte boundary.

    The SunAda compiler does allow you to "shoot yourself in the foot"
by misaligning floating point values in records.  Usually you have to
ask for trouble with alignment or record representation clauses, but
it might be possible to cause the problem by nesting records.  In any
case, a priority 10 trap is mem_address_not_aligned, and I suspect
that the bad reference is in line 131.

--

					Robert I. Eachus

with Standard_Disclaimer;
use  Standard_Disclaimer;
function Message (Text: in Clever_Ideas) return Better_Ideas is...