[prev in list] [next in list] [prev in thread] [next in thread] 

List:       majordomo-workers
Subject:    How much error checking is good, new config file release
From:       "John P. Rouillard" <rouilj () terminus ! cs ! umb ! edu>
Date:       1993-11-19 4:09:55
[Download RAW message or body]


I just got to fixing the last of the stupid niggling bugs in my config
file stuff (required the rewite of a function), any my regression
suite once again runs clear, so I am back to more theoretical issues.

My newest code supports array and enumerated types are supported. 
Array's must be	declared as an array, and are set up like:


	  array_keyword << EOAET
	  val1
	  val2
	  val3
	  val4
	
	  EOAET

The only whitespace only line in the array of elements can occur
before the end of array element token (EOAET above). I.E.

	  array_keyword << EOAET
	  val1
	  val2
	
	  val3
	  val4
	  EOAET

would fail with unexpected whitespace, and unknown keywords val3, val4
and EOAET. This allows me to resync on blank lines so that a forgotten
EOAET will be localized as much as possible. This may be a problem for
long descriptions (e.g.  info file stuff).  I may use an inital '.' on
the line to indicate a real blank line for info descriptions etc. Or I
may just say screw it, and digest the rest of the config file and
throw up at the end of it, while recording the start number. It will
miss any errors in the file after the missing EOAET, but that may not
be too bad.  I am wondering about the effects of a missing EOAET, with
other commands after the info setting command.

Does anybody have any thoughts on the above. I like being able to trap
errors close to their source, but there is something to be said for
not requiring random .'s at the beginnings of lines.

Note that my newest code doesn't rewrite the REPLY handling. If
anybody wants a copy of the patches (to version 1.62) to play with let
me know. Obviously, the patches can't be redistributed, but I would
like a few brave souls who wouln't mind setting up some lists with
this new stuff. I will be installing this version in the production
slot of majordomo@cs.umb.edu tomorrow probably. If anybody want's to
try banging on it tonite, its available at

	majordomo-test@cs.umb.edu

BTW majordomo@cs.umb.edu and majordomo-test@cs.umb.edu administer the
same lists.

				-- John
John Rouillard

Special Projects Volunteer	University of Massachusetts at Boston
rouilj@cs.umb.edu (preferred)	Boston, MA, (617) 287-6480
===============================================================================
My employers don't acknowledge my existence much less my opinions.

[prev in list] [next in list] [prev in thread] [next in thread]