debuging idea (was: Re: Stumped: 0 "not a quantity")

Subject: debuging idea (was: Re: Stumped: 0 "not a quantity")
From: Joerg Wittenberger <Joerg.Wittenberger@xxxxxxxxx>
Date: 27 Jul 1999 10:46:20 +0200
--text follows this line--
Hello OpenJade folks,

>>>>> "BI" == Brandon Ibach <bibach@xxxxxxxxxxxxxx> writes:

BI>    Gotcha! :) On a serious note, I'd advise that the first step in
BI> tracking down something like this (any kind of "invalid argument
BI> or value" message from Jade) is to stick (debug) around the
BI> expression in question.  Sure, Jade may print out the "value" in
BI> question as part of the error message, but (debug) will tell you
BI> more.  In this case, it would have shown you that it wasn't the

This is todays way of debuging DSSSL with jade.  Often enough I
introduced new bugs (superflous ")", missplaced deletions of debuging
code etc.) when removing those debug statements.

(debug) is non-standard anyway, isn't it?  It shouldn't be too hard to
change the reader to accept some non-standard "breakpoint-syntax",
which could make debuging less error prone.  If I could write say #?
*in front* of an expression with the effect of sticking (debug)
*around* that expression - that would be cool.

Example: currently we do
(let auths-loop
    ((auth-idx 0))
  (if (<= num-auths
          (debug auth-idx))

proposed:
(let auths-loop
    ((auth-idx 0))
  (if (<= num-auths
          #? auth-idx)


Sure it's much more useful for complicated expressions to debug; it's
also not the same simple syntax as the rest of DSSSL.  If introduced,
another command line argument should enable/disable this kind of
debuging code.

Or is this idea total nonsense?

/Jerry
 


 DSSSList info and archive:  http://www.mulberrytech.com/dsssl/dssslist


Current Thread