MathGroup Archive 2007

[Date Index] [Thread Index] [Author Index]

Search the Archive

Debugging with v6

  • To: mathgroup at
  • Subject: [mg82357] Debugging with v6
  • From: Art <grenander at>
  • Date: Thu, 18 Oct 2007 04:56:10 -0400 (EDT)
  • Organization: University of California, Berkeley

A few questions regarding debugging in v6.0.1 (Linux 64-bit). I am a new
user. Answers to any would be appreciated.

1. I work with large arrays and trivial typos in code can result in Mathematica
quickly consuming all memory, then all of swap before I can abort or kill
the kernel. Is there a global option to force abort when MemoryInUse[]
goes beyond a certain amount, a sort of global MemoryConstrained[]? How
about abort on any messages?

2. If the frontend becomes unresponsive because the kernel has used up all
memory, is it possible to kill the kernel process outside the frontend
with out killing the frontend? If I just kill the kernel process, a popup
appears telling me that Mathematica wasn't able complete a dynamic evaluation. If I
select abort on the popup, the frontend crashes.

3. I have difficulty getting the debugger tools to work. What is the
criteria for execution to stop at a breakpoint? Do you have to re-execute
the code (such as a function) that contains the breakpoint after you have
defined the breakpoint? Most of the time, I can't get the debugger to stop
at the defined breakpoint. This is true also in Workbench (where you
select a line rather than an expression). Often, the frontend hangs if I
look at any of the local variables in the stack display and stepping in
debug mode is often too slow to be usable.

The only helpful documentation I have found for the debugger is the
screencast for Workbench and a short post in this forum. It is pretty
buggy on my platform. I would love to use it. I will learn to use Trace[]


  • Prev by Date: Setting up external functions to be called from Mathematica
  • Next by Date: Re: Compile forgets assigned values inside Do unless they are initialised
  • Previous by thread: Re: Setting up external functions to be called from Mathematica
  • Next by thread: Baffling behavior of Factor[] in versions 5 and 6 -- a bug?