comp.lang.ada
 help / color / mirror / Atom feed
From: "Randy Brukardt" <randy@rrsoftware.com>
Subject: Re: QtSql problem
Date: Mon, 9 Feb 2009 21:09:57 -0600
Date: 2009-02-09T21:09:57-06:00	[thread overview]
Message-ID: <gmqr68$f04$1@munin.nbi.dk> (raw)
In-Reply-To: 498dfa7f$0$30231$9b4e6d93@newsspool1.arcor-online.net

"Georg Bauhaus" <see.reply.to@maps.futureapps.de> wrote in message 
news:498dfa7f$0$30231$9b4e6d93@newsspool1.arcor-online.net...
...
> However obvious this generic might be to those well versed
> in Ada programming, TB_Wrap continues to deserve 5 stars. :-)
>
> At least one non-GNAT Ada compiler produces tracebacks by
> default, that is, ObjectAda. What do other compilers do?

Janus/Ada has *always* generated complete tracebacks with names and line 
numbers. Even on the compiler that ran on and targeted 48K Z80 CP/M machines 
back in 1981. We learned early on (from using a several compilers that had 
such tracebacks and some that did not at the University of Wisconsin) how 
much easier it was to debug programs with proper tracebacks. I recall 
beating my head against a wall with the original Unix C compiler, whose only 
run-time error message was "bus error - core dumped" (that's the complete 
message!) -- didn't want to do that with our own compiler.

We do provide an option for turning the overhead of the trace back off, but 
it generally is needed only in the most speed-critical routines. (There is a 
also an easy method for turning off the overhead in the runtime for our 
embedded customers, but hardly any of them have used it.)

I recall complaining to Tucker and the rest of the Ada 9x team that the 
definition of Ada.Exception_Information made it hard to include a full 
walkback. The definition was changed somewhat, and indeed, Janus/Ada 
includes a truncated version of the walkback in Exception_Information.

My limited experience with other Ada compilers is that most include a 
traceback by default -- a common feature that I link to think came about in 
part because their customers complained that the "cheap" Janus/Ada had this 
useful feature, and their expensive compilers didn't. (I know that happened 
because several customers have told me that they made such complaints at one 
point or another.)

                                                      Randy Brukardt.





      reply	other threads:[~2009-02-10  3:09 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-02-06  4:43 QtSql problem tworoses1
2009-02-06  8:27 ` Vadim Godunko
2009-02-06  8:31   ` Gautier
2009-02-06 18:21     ` tworoses1
2009-02-06 19:56       ` Gautier
2009-02-06 20:44         ` tworoses1
2009-02-06 22:51       ` sjw
2009-02-08  0:09         ` tworoses1
2009-02-09 14:04           ` Vadim Godunko
2009-02-09 15:55             ` Ivan Levashew
2009-02-10 10:32               ` Vadim Godunko
2009-02-13  4:21                 ` tworoses1
2009-02-07 21:17     ` Georg Bauhaus
2009-02-10  3:09       ` Randy Brukardt [this message]
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox