Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ncar!ames!apple!claris!jazzman
From: jazzman@claris.com (Sydney R. Polk)
Newsgroups: comp.sys.apple
Subject: Re: GSBug Questions
Message-ID: <10407@claris.com>
Date: 19 Jul 89 23:55:46 GMT
References: <36410002@hpindwa.HP.COM>
Organization: Claris Corporation, Santa Clara CA
Lines: 48

From article <36410002@hpindwa.HP.COM>, by tribby@hpindwa.HP.COM (David Tribby):
> Several comp.sys.apple readers were helpful the last time I had a GSBug
> question, so here are a couple more...
> 
> 1. Is there any way to set a "real time" breakpoint for a toolset call?
>    Since I do most of my programming in TML Pascal, I would like to set
>    a breakpoint on a particular toolset call, then let the program run
>    in real time until the invocation.
Ooh, yuck!  OK, you have to use something like Nifty List or Mangler to
find where the toolset is loaded.  You have to look in the tollset jump
table to figure out where the tool is.  If you suceed in this, you still
may fail, because you can't set a breakpoint to an address in ROM.

Now that I have thought a second, you might get the TSPtr by finding
and area in memory that is blank, assembling something to make the correct
call, and then setting a BP their.  You still may run into the ROM
problem, although you could set a bp in the dispatcher at the appropriate
place.

It would be much easier to set a breakpoint at the routine you are calling
the tool from and then turn mp off for e1000-ffff and step into it.
You do NOT want to do this in GS/OS calls; they have real-time code in them.

> 2. Does the "init" version of GSBug interfere with the "run time" version?
>    I have the init version enabled at startup, and sometimes peculiar
>    things happen when I invoke the run-time version and try to load & debug.
>    For example, I loaded my program, then tried to invoke the loader/dumper
>    CDA to get the seqment addresses. When I pressed OA/Ctrl/esc, the display
>    froze (the cursor quit blinking and no characters echoed on the screen).
>    Things seemed to start working again when I pressed the option key (by
>    itself), but it turned out I had invoked the init version of GSBug! I
>    Restored from the init version, then could work with my program in the
>    run-time GSBug.
What we do:  We set up a check at the beginning of our program to
see if Opt-Ctl are down.  If they are, we brk, which drops into the init
and we can set things up from there.  I NEVER use the run-time debugger
anymore;  if I have to, I run off of a disk without the init.

> Thanks in advance for any help!
> 
> --Dave Tribby
>                                    - - - - -
>         ARPA: tribby%hpda@hplabs.HP.COM   UUCP: hplabs!hpda!tribby
-- 
Syd Polk           | Wherever you go, there you are.
jazzman@claris.com | Let the music be your light.
GO 'STROS!         | These opinions are mine.  Any resemblence to other
GO RICE!           |  opinions, real or fictitious, is purely coincidence.