Ha! I have frequently asked customers to send me a dump, so that I could examine the entrails of the beast (and hopefully track its spoor).
David de Jongh
After describing his own debugging practices, Gerhard Postpischil writes:
<begin snippet>
For production programs, I prefer to save the registers, if useful,
issue one or more explanatory messages, and then issue a user abend.
</end snippet>
and this is enough.
Debugging--as opposed to prior and subsequent systematic testing,
which are not--is a highly personal matter. Debuggers are and should
be free to use any technique they find effective; they should indeed
be free, as I said here long ago, to consult a haruspex (plural
haruspices) if they judge it helpful to do so (privately, of course,
lest the animal-rights people target their colleagues too).
--
John Gilmore, Ashland, MA 01721 - USA
<begin snippet>
For production programs, I prefer to save the registers, if useful,
issue one or more explanatory messages, and then issue a user abend.
</end snippet>
and this is enough.
Debugging--as opposed to prior and subsequent systematic testing,
which are not--is a highly personal matter. Debuggers are and should
be free to use any technique they find effective; they should indeed
be free, as I said here long ago, to consult a haruspex (plural
haruspices) if they judge it helpful to do so (privately, of course,
lest the animal-rights people target their colleagues too).
--
John Gilmore, Ashland, MA 01721 - USA
