MathGroup Archive 1996

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

Search the Archive

Exceptions handling

  • Subject: [mg3320] Exceptions handling
  • From: fdupont at irisa.fr (Florent Dupont-De-Dinechin)
  • Date: 26 Feb 1996 14:06:12 -0600
  • Approved: usenet@wri.com
  • Distribution: local
  • Newsgroups: wri.mathgroup
  • Organization: IRISA, Campus de Beaulieu, 35042 Rennes Cedex, FRANCE
  • Sender: daemon at wri.com


We are developping a relatively big program transformation environment
in Mathematica, and we are wondering what is the best way
of handling errors. Problems occur when a the evaluation of some function 
is blocked for some reason (in relation with an error in the input).
The caller function gets a whole expression tree instead of getting 
the expected result, which causes even more errors, and so on.

Currently we use various combinations of Catch/Throw, Message/Check
and strong type checking to handle these problems, but the result is quite
inconsistent and definitely not idiot-proof, and it is getting worse
as the size of the system increases (as one may expect). 

Is there an "official" way of correctly handling exceptions ? We are looking
for something like "good programming habits", from the litterature or 
other users' personal experience.


	Thanks in advance,

		Florent

           "Reality shouldn't constraint our formalisms."
--------------------------------------------------------------------------------
        Florent de Dinechin. Equipe Architectures Paralleles Integrees
             IRISA, Campus de Beaulieu, F-35042 Rennes, France
        tel: (33) 99 84 74 56		      e-mail: fdupont at irisa.fr
--------------------------------------------------------------------------------
   "Dans toute entreprise serieuse, l'intelligence est un don negligeable."





  • Prev by Date: Exceptions handling
  • Next by Date: Path find algorithm for undirected graph
  • Previous by thread: Exceptions handling
  • Next by thread: Path find algorithm for undirected graph