Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!lll-winken!uunet!cbmvax!vu-vlsi!swatsun!jackiw
From: jackiw@cs.swarthmore.edu (Nick Jackiw)
Newsgroups: comp.sys.mac.programmer
Subject: Re: How do I get Discipline?
Message-ID: <2555@ilium.cs.swarthmore.edu>
Date: 16 Mar 89 17:07:18 GMT
References: <12678@dartvax.Dartmouth.EDU>
Reply-To: jackiw@ilium.UUCP (Nick Jackiw)
Organization: Visual Geometry Project, Swarthmore College, PA
Lines: 50

In article <12678@dartvax.Dartmouth.EDU> earleh@eleazar.dartmouth.edu (Earle R. Horton) writes:
> I have heard a lot about something called Discipline, a package which
> checks parameters to ToolBox calls.  Where is the best place to get it?
> 
> Earle R. Horton. 23 Fletcher Circle, Hanover, NH 03755--Graduate student.

Two that I know of:
The copy of TMON that I purchased came with Darin Adler's EUA (Extended
User Area), which includes two varieties of trap discipline: rigid and
lenient.  Lenient works pretty well at detecting odd-addresses and other
This-MUST-Be-Wrong sort of arguments (although most of the ones its flagged
in my code are SOOO wrong that they would've halted the processor shortly
thereafter anyway); Rigid works TOO well (in that it complains about a 
lot of things that are intentional and that work by bending but not
breaking the rules). An pre-6.0 Finder used to continually break, for
instance.

Steve Jasik's _TheDebugger_, included with MacNosy v2, has TrapDiscipline
available under its Bondage menu.  A relevant quote from the manual:

"The values of the various types [of trap args] are checked for range errors.
I. E. things that should point into the heap, or a particular heap are
checked as such, etc. ... Note that some ROM trap calls are flagged as bad
by Discipline, but are OK in the sense that the offending fields will be
properly setup when they are used. That is the local context in which
Discipline checks the arguments is insufficient relative to the global 
context of the use of the field." [p. 35]

While theDebugger looks feature-laden and more powerful, in general, than
TMON, its manual seems to have been composed by amphetamized howler
monkeys.  I don't object to all of the "unimplemented features," or even
to a program that comes with its internal-debugging-data still displayed
(cause of the notorious Internet worm...), but WHY OH WHY DO DEBUGGING
MANUALS HAVE TO BE SO BADLY WRITTEN?  The TMON manual I have is also
garbage, though not as bad (and I understand it's been revised). Debuggers
are supposed to save us time. A good manual would save lots.

I'm waiting on SADE, which I suspect will have at least intelligible, if
not well-written, documentation.  APDA's way behind on their MPW 3 
shipments though...  (I love paying a mandatory $12 FedEx bill for a
package I'm not going to see for four weeks anyway.)

Hope this helps.


-- 
+-------------------+-jackiw@cs.swarthmore.edu / !rutgers!bpa!swatsun!jackiw-+
|  nicholas jackiw  | jackiw%campus.swarthmore.edu@swarthmr.bitnet           |
+-------------------+-VGP/MathDept/Swarthmore College, Swarthmore, PA 19081--+
"Ah...I've got this CHRONIC pain."                             _True Believer_