Re: Messages again

[John Darrington]

On Mon, May 01, 2006 at 10:34:21AM -0700, Ben Pfaff wrote:
     
     I mean what's currently done with a stack of "getl_location"s; in
     essence a description of what *provoked* the error and a trace of
     how we got there.  Right now it's just a single file and line
     number pair, but I mean to generalize it to be more useful.
     
     I think this concept is separate from the one you have in mind.
     What do you think?


It depends on how far you're going to take the generalisation.  But I
think that "what provoked the error [...] and  how we got there" is a
good description of what I'm trying to achieve too.

A file name and line number is certainly useful for errors generated
from syntax files.  It's not so useful for errors generated from
user's mistakes in a GUI (eg trying to enter a string in a cell
dedicated to numeric data).

The difference is  I suppose, in the way we've approached the problem.
In my proposal, I've put the burden on the UI programmer to predeclare
the actions of a block of code; "I'm about to enter something into a
cell".  Whereas perhaps you can think of a better way to determine that
an error was provoked by an attempt to enter data into a cell for
which it was not appropriate.

J'

-- 
PGP Public key ID: 1024D/2DE827B3 
fingerprint = 8797 A26D 0854 2EAB 0285  A290 8A67 719C 2DE8 27B3
See http://pgp.mit.edu or any PGP keyserver for public key.

signature.asc
Description: Digital signature