Flame on Verdix, Any other good Unix compilers???

Flame on Verdix, Any other good Unix compilers???

[Michael Endrizzi]

I have been using VADS Version 5 Unix self-hosting for 7 months now and
I cannot take it anymore. I am up to my eyeballs in "internal errors"
,"database inconsistency","segmentation error,core dump","out of memory".
4 of us used DEC VMS Ada version 1!!! for 1 and half years and I can
only remember about 4 or 5 errors.  Myself in 7 months has found more
errors then that in a Version 5 compiler!!! I am at the point of
bypassing the recompile facilities and building my own Unix scripts
for recompiling and linking.

So tell me, am I the only one??? Tell me I'm not going insane, I'm
almost regretting moving from C to Ada (now that's desperate). Are
there any other compilers out there who's databases don't go south
on a sneeze. That implement generics in a fashion so that they compile.
That has a runtime system without endless loops. That has library 
management tools that don't core dump and spit out Diana trees.


AAAAAAAHHHHHHHHHHHHHHHHHHH!!!!!!!!!!!!!


				Dreez

=================================================================
=================================================================
               Michael J. Endrizzi
	Secure Computing Technology Corp.
	   1210 W. County Road E #100
	      Arden Hills, Mn. 55112
	        endrizzi@sctc.com
	          (612) 482-7425
	
*Disclaimer: The opinions expressed above are not of my employer
             but of the American people.
=================================================================
=================================================================

Re: Flame on Verdix, Any other good Unix compilers???

[Charles H. Sampson]

In article <1990Aug9.191734.11633@sctc.com> endrizzi@sctc.com (Michael Endrizzi ) writes:
>
>                                                           ... Are
>there any other compilers out there who's databases don't go south
>on a sneeze. That implement generics in a fashion so that they compile.
>That has a runtime system without endless loops. That has library 
>management tools that don't core dump and spit out Diana trees.

     I've been very impressed with the IBM-PC version of the Alsys compiler.
I've never used a Unix version but, knowing the way they keep all incarna-
tions of the compiler in step, I would expect it to be equally solid.

                            Charlie Sampson

was compiler bashing

[Karl A. Nyberg]

If we're going to have to play compiler/vendor bashing here...

In article <1990Aug9.191734.11633@sctc.com> endrizzi@sctc.com (Michael Endrizzi ) writes:
>I have been using VADS Version 5 Unix self-hosting for 7 months now and
>I cannot take it anymore. 

Any self-hosted compiler in particular?  Brand of computer, version of
operating system, more accurate revision level of compiler?  I'll give you a
dollar for it if you want to sell... :-)

>I am up to my eyeballs in "internal errors"
>,"database inconsistency","segmentation error,core dump","out of memory".

I presume you followed the documentation (and Walt Penny's recent e-mail)
and have submitted these as reports.  Of course, if you have enjoyed
masochism for the past seven months, why start complaining now...

>4 of us used DEC VMS Ada version 1!!! for 1 and half years and I can
>only remember about 4 or 5 errors.  Myself in 7 months has found more
>errors then that in a Version 5 compiler!!! 

Have you tried taking the code you're writing now, probably much more
complicated, with more extensive generics, representation specs, etc. and
run it against that version 1 compiler?  There's no telling how old the
compiler you're using is, or whether it is current for the operating system
you're running on.

>I am at the point of
>bypassing the recompile facilities and building my own Unix scripts
>for recompiling and linking.

I did that years ago.  It helps me out when I port to the PC, although I did
see that Meridian's latest summer release has a recompilation facility.  As
I start to gain confidence in the accurate execution of these things, I'll
use them.  Until then, I send in bug reports.

>So tell me, am I the only one??? 

No.  But it's not the only compiler.  I have found this true with numerous
compilers.

>Tell me I'm not going insane, I'm
>almost regretting moving from C to Ada (now that's desperate). 

... or language.  Surprise, surprise - I even have to work around bugs in
the GNU cc.

This is from mail years (literally ago) when I was actually semi-moderating
info-ada and the same sort of thing started up with TeleSoft.

	We here at TeleSoft whole-heartedly agree with your posting that
	asked people NOT to submit compiler questions/results/complaints
	over info-ada [/comp.lang.ada].  Hopefully, people will follow this
	wonderful advice and save both your time and theirs.

	The correct route to take IS directly to the vendor.  Like most
	folks out there, we can be reached by email.  If you've got ARPAnet 
	connections, then mailing to [address deleted for fear it is too old
	- can somebody at TeleSoft fill in?]

	Of course, the telephone has worked for many satisfied customers. 
	Anyone with a product should have the vendor's phone number
	somewhere in the documentation.  Ours is (619)457-2700.

	Once again, thanks for encouraging the readers of info-ada to make 
	proper use of the channels that vendors have built to care for their
	user's in the most effective way possible. 

-- Karl --

Karl A. Nyberg				karl@grebyn.com
Post Office Box 497			Grebyn Corporation
Vienna, VA 22183-0497 USA		+1-703-281-2194

Disclaimer: Meridian, TeleSoft and Verdix (among others) buy a lot of my
products and services at market rates.  Have you gotten YOUR own copy of the
Annotated Ada Reference Manual from YOUR compiler vendor yet? :-)

Re: was compiler bashing

[Michael Endrizzi]

karl@grebyn.com (Karl A. Nyberg) writes:

>If we're going to have to play compiler/vendor bashing here...

>In article <1990Aug9.191734.11633@sctc.com> endrizzi@sctc.com (Michael Endrizzi ) writes:
>>I have been using VADS Version 5 Unix self-hosting for 7 months now and
>>I cannot take it anymore. 

>Any self-hosted compiler in particular?  Brand of computer, version of
>operating system, more accurate revision level of compiler?  I'll give you a
>dollar for it if you want to sell... :-)

Version 5 Verdix, Sun Unix 4.0.3 running on Sun 3/60. Reports about
Version 6 are not very encouraging.

>masochism for the past seven months, why start complaining now...

Because as it turns out I am not the only one. My request for information
was answered by others that are experiencing the same level of difficulties.
And it's not just Verdix, several others were mentioned (I won't name 
them because I did not use them). 

>No.  But it's not the only compiler.  I have found this true with numerous
>compilers.

Alsys just called me. They were very diplomatic about commenting on other
compiler vendors simply stating that "our compiler will speak for itself".
We'll see.

>... or language.  Surprise, surprise - I even have to work around bugs in
>the GNU cc.

We use GNU cc extensively and have very very few problems. It's an apples
and oranges game because Ada is suck a complex language.

>This is from mail years (literally ago) when I was actually semi-moderating
>info-ada and the same sort of thing started up with TeleSoft.

>	We here at TeleSoft whole-heartedly agree with your posting that
>	asked people NOT to submit compiler questions/results/complaints
>	over info-ada [/comp.lang.ada].  Hopefully, people will follow this
>	wonderful advice and save both your time and theirs.

>  [ stuff is deleted ]

This is BS. The net is a public bulletin board and ANY manufacture
should be PROUD to have their product reviewed in a public forum.
Other forums (rec.audio,rec.video,rec.cars(??),comp.database) publicly bash and praise 
products, why can't a piece of software???  Only manufacturers
who are trying protect their backsides would make such a statement
as above. Reminds me of GM trying to fending off the Japanese by
hiding their dirty laundry underneath marketing blitz's and
voluntary restraints. In the end, the truth stands, and quality
products are rewarded by the marketplace.

I do submit bug reports. If any of the problems I am experiencing
are my fault, PLEASE let me know so I can fix it. Maybe others
went through the same problems and resolved it. By posting complaints,
Verdix is not the only one with it's neck on the line. What if the
problems I am complaining about are my fault?? Makes me look pretty
stupid, too. However, in this case others have stated that they too
are having similar problems. 

				Dreez

=================================================================
=================================================================
               Michael J. Endrizzi
	Secure Computing Technology Corp.
	   1210 W. County Road E #100
	      Arden Hills, Mn. 55112
	        endrizzi@sctc.com
	          (612) 482-7425
	
*Disclaimer: The opinions expressed above are not of my employer
             but of the American people.
=================================================================
=================================================================

Re: Flame on Verdix, Any other good Unix compilers???

[Lennart AtternaesTT/UO]

endrizzi@sctc.com (Michael Endrizzi ) writes:


>I have been using VADS Version 5 Unix self-hosting for 7 months now and
>I cannot take it anymore. I am up to my eyeballs in "internal errors"
>,"database inconsistency","segmentation error,core dump","out of memory".
>4 of us used DEC VMS Ada version 1!!! for 1 and half years and I can
>only remember about 4 or 5 errors.  Myself in 7 months has found more
>errors then that in a Version 5 compiler!!! I am at the point of
>bypassing the recompile facilities and building my own Unix scripts
>for recompiling and linking.

>So tell me, am I the only one??? Tell me I'm not going insane, I'm
>almost regretting moving from C to Ada (now that's desperate). Are
>there any other compilers out there who's databases don't go south
>on a sneeze. That implement generics in a fashion so that they compile.
>That has a runtime system without endless loops. That has library 
>management tools that don't core dump and spit out Diana trees.


I have used VADS version 5.5 and 5.7 for three years and I have exactly the
same experience. VADS can't handle nested generic, the library manager is a
disaster with internal errors as a consequence and using floating point
support (MC68881) needs a workaround. I don't know how many hours I have
spent the last three years to find workarounds.

A half a year ago I bought version 6.0g for Sun4 and naive enough I thougt
that the new version would be improved, however, I found a bug that makes it
impossible for us to use it, so we have to wait for the next release. My
opinion is that VADS is not worth to be even compared to the VAX VMS Ada.

Now I have had it and I would appreciate if someone could give me some
information about other compilers for Sun3 and Sun4.

--
Lennart Atternaes

Ericsson Telecom AB
Dep. KK/ETX/TT/UO	   Telephone		 Fax
S-126 25 STOCKHOLM	   Nat	  08-7197564	 Nat	08-7196000, 7080
SWEDEN			   Int +46 8 7197564	 Int +46 8 7196000, 7080

E-mail: ytlat@kk.ericsson.se

Re: Flame on Verdix, Any other good Unix compilers???

[Ted Holden]

In article <1990Aug9.191734.11633@sctc.com> endrizzi@sctc.com (Michael Endrizzi ) writes:
>
>I have been using VADS Version 5 Unix self-hosting for 7 months now and
>I cannot take it anymore. I am up to my eyeballs in "internal errors"
>.........
>AAAAAAAHHHHHHHHHHHHHHHHHHH!!!!!!!!!!!!!

Rumor has always had it that Verdix' Ada compiler was the very best,
i.e. that everybody else has been experiencing more grief than you have.
My own experiences with Ada gives me no reason to doubt this, nor do
other experiences which I read about.

Ted Holden
HTE

Re: was compiler bashing

[Bob Kitzberger @sation]

In article <21105@grebyn.com>, karl@grebyn.com (Karl A. Nyberg) writes:
> 
> This is from mail years (literally ago) when I was actually semi-moderating
> info-ada and the same sort of thing started up with TeleSoft.
> 
> [...]
> 	The correct route to take IS directly to the vendor.  Like most
> 	folks out there, we can be reached by email.  If you've got ARPAnet 
> 	connections, then mailing to [address deleted for fear it is too old
> 	- can somebody at TeleSoft fill in?]

The correct email address is support@telesoft.com.  Phone 619/457-2700.

	.Bob.
-- 
Bob Kitzberger               Internet : rlk@telesoft.com
TeleSoft                     uucp     : ...!ucsd.ucsd.edu!telesoft!rlk
5959 Cornerstone Court West
San Diego, CA   92121-9891   "There's too much caffeine in your bloodstream..."
(619) 457-2700 x163                                              -- The Smiths
------------------------------------------------------------------------------

Re: Flame on Verdix, Any other good Unix compilers???

[James THIELE]

In article <2052@cod.NOSC.MIL> sampson@cod.nosc.mil.UUCP (Charles H. Sampson) writes:
|     I've been very impressed with the IBM-PC version of the Alsys compiler.
|I've never used a Unix version but, knowing the way they keep all incarna-
|tions of the compiler in step, I would expect it to be equally solid.
|
|                            Charlie Sampson

I remember once using the Alsys hosted on Apollo and getting the error
message "Uncaught Ada exception at xxxx".  I eventually figured out
that this was the compiler's clever and informative way of telling me
I had an undeclared identifier.

James Thiele -- microsoft!jamesth

Re: was compiler bashing

[Vladimir G. Ivanovic]

You know, at a certain point one has to decide, "Do I prefer to be right, or
do I prefer to be effective?"  I think that what Karl is saying (pardon me if
I put words into his mouth), is if you want to be effective, 

	G O    D I R E C T L Y    T O    T H E    V E N D O R

What part of the previous line don't you understand?

-- Vladimir

Re: was compiler bashing

[Michael Endrizzi]

vladimir@prosper (Vladimir G. Ivanovic) writes:

>	G O    D I R E C T L Y    T O    T H E    V E N D O R

>What part of the previous line don't you understand?

I have gone to the vendor. So has MANY!!! other people that have
written me from all over the world. Verdix Users are on Release 6
and it is not getting any better.  Our projects/jobs are dependant
on tools that DON"T WORK. What do you suggest, that we wait for
Release 1019.

Many other news groups use this media to exchange information
on manufacturers. Why is comp.lang.ada so special???  No one is
preventing the manufacturers from defending their product. What is
YOUR vested interest?? If you don't like the discussion, then hit
the 'n' key and continue to mind your own business. 

What part of the previous paragraph don't you understand?

			Dreez

=================================================================
=================================================================
               Michael J. Endrizzi
	Secure Computing Technology Corp.
	   1210 W. County Road E #100
	      Arden Hills, Mn. 55112
	        endrizzi@sctc.com
	          (612) 482-7425
	
*Disclaimer: The opinions expressed above are not of my employer
             but of the American people.
=================================================================
=================================================================

Re: was compiler bashing

[Thomas A Peterson]

=You know, at a certain point one has to decide, "Do I prefer to be right, or
=do I prefer to be effective?"  I think that what Karl is saying (pardon me if
=I put words into his mouth), is if you want to be effective, 
=
=	G O    D I R E C T L Y    T O    T H E    V E N D O R
=
=What part of the previous line don't you understand?

Some of us are tired of going to the vendor. We wish to notify other
users about the problems we have been experiencing and force the
vendor to fix the compiler via this coercive action.  I am certain
that my current Ada vendor would much rather port their system to a
new platform to get yet more unsuspecting customers than fix the
mistakes that exist in the current compiler.

Tom

Re: was compiler bashing

[John Townsend]

vladimir@prosper (Vladimir G. Ivanovic) writes:

>	G O    D I R E C T L Y    T O    T H E    V E N D O R

>What part of the previous line don't you understand?

and may I add...

	        A N D   T H E   M A R K E T P L A C E.

Verdix (and nobody else either) is not going to make an expensive change to
their compiler just because one user (who has already paid for the product
anyway) has a problem.  They have too many other important revenue-producing
jobs to do that.  From my experience as a compiler validator for the Language
Control Facility at WPAFB, it was obvious that if a vendor wanted to cheat on
a validation, it would be disgustingly easy.  The real "validation" of
compilers
is done by the marketplace, and you vote with your dollars (or pounds,
or marks,
or francs, or rubles ;) or whatever).  Granted, going public with a gripe
without going to the vendor is downright discourteous.  Nevertheless, to be
truly "effective" in a free market, you may need to include leverage in your
tools.

-------------------------------------------------------------------------------

                       John Townsend (johnt@mead.UUCP)

                            c/o Mead Data Central
 		            9443 Springboro Pike
                                P.O. Box 933
                             Dayton, Ohio, 45401

                               (513) 865-7250

 "Any opinions expressed herein are my own, and not necessarily those of MDC."

-------------------------------------------------------------------------------

Re: was compiler bashing

[Lennart AtternaesTT/UO]

vladimir@prosper (Vladimir G. Ivanovic) writes:

>You know, at a certain point one has to decide, "Do I prefer to be right, or
>do I prefer to be effective?"  I think that what Karl is saying (pardon me if
>I put words into his mouth), is if you want to be effective, 

>	G O    D I R E C T L Y    T O    T H E    V E N D O R

>What part of the previous line don't you understand?

This is insulting. You have to realize that there are people out
here that have serious problems that cost their employer a lot of
money. Of course we have sent reports to the vendor, however, my
employer would rather buy a more expensive compiler rather than
having the employees working part time for the Verdix cooperation
debugging their software. If there aren't any better Ada
compilers we of course have to change back to fortran (or C if we dare
changing language one more time).


--
Lennart Atternaes

Ericsson Telecom AB
Dep. KK/ETX/TT/UO	   Telephone		 Fax
S-126 25 STOCKHOLM	   Nat	  08-7197564	 Nat	08-7196000, 7080
SWEDEN			   Int +46 8 7197564	 Int +46 8 7196000, 7080

E-mail: ytlat@kk.ericsson.se

Re: Flame on Verdix, Any other good Unix compilers???

[John Ostlund]

JIMBO!!!!!!!!!!!!!!!

I thought I'd run into you some day in this group...

Reply if this is the James Thiele I think it is...

Later,  JJ