CVS update: gnucash/src/backend/file

[Derek Atkins]

linas at linas.org (Linas Vepstas) writes:

> On Tue, Jul 29, 2003 at 01:11:00PM -0400, Derek Atkins was heard to remark:
> > > 
> > > Why can't you just set the error status, the way you did for the unlink
> > > errors? Somehwere higher up the stack, the gui routines do check for
> > > errors ... 
> > 
> > Because it's NOT an error -- it's a warning.  Also, I don't think the
> 
> Hmm.. I see nothing wrong with popping up a warning message.

True, but there is no way to do that.

> > The real problem is that the backend function is not continuable..  If
> > the backend wants user input (due to a warning, or even a continuable
> > error) then there is no way to warn the user via a dialog and then
> > continue the process.  There is no way to "continue the process".
> 
> Yeah, it wasn't supposed to be synchronous error delivery. 
> (Insert vague mumblings about remote servershere).  I guess 
> telling the user to 'retry' doesn't cut it?

No, that doesn't cut it.  Indeed, it wont even help because the
warning would happen the next time, too.

Matthew and I were discussing this on #gnucash the other day.  What we
really need is a state-machine so that you can interrupt the
processing to present the warning to the user and then jump back into
the processing engine where you left off (or cancel the process if you
need to).

> --linas

-derek

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available