Remove Bill/Vendor

[Derek Atkins]

linas at linas.org (Linas Vepstas) writes:

> On Wed, Mar 26, 2003 at 11:30:29AM -0500, Derek Atkins was heard to remark:
> > The problem is that you would need a hook into every FooCommitEdit()
> > function, and then some way to walk in reverse.  
> 
> Yes, that's exactly it.  I didn't say its not a fair bit of work.
> 
> > Also note that the
> > only data object that keeps track of its original state between
> > BeginEdit() and CommitEditis() a Transaction.
> 
> It also makes me think in nebulous programming terms: 
> -- where can I find a language or system with the popularity &
> robustness of C, but still allows me to describe a struct
> so that at beginEdit() I could say "remember thyself", and then
> at commitEdit() be able to simply say "save thyself to the audit trail".

You could do it in C++ or Java.  In fact a lot of code would probably
be a lot easier if we used C++..  But I don't want to think about
doing that. ;)

> You may have noticed the rather unpleasently large complexity in
> CommitTransaction(), which I don't like.  But I don't think there's
> any other way, at least not in C (except possibly some interesting stunt
> using gnome  glib gobjects).

I dont think gobjects would help here.

> while on topic: there *may* (??) be some benefits to converting
> gnucash structs to glib gobjects.  I'm not sure what these are, just
> yet, but if you had time to play, this might be a worthy game ...

Maybe.  I'd rather see some of the core functionality completed (like
getting lots finished, and getting stock accounts to work right).
But we're all busy....

> --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