Quickfill gripe -- again!

Paul Abrahams abrahams at acm.org
Wed Dec 21 00:12:02 EST 2011


On Tuesday, December 20, 2011 06:05:20 PM David T. wrote:
> ________________________________
>  From: Paul Abrahams <abrahams at acm.org>

> > There are two different issues going with Quickfill. The first is
> > getting
> > the description formed as we like; the other is getting the splits the
> > way we like. I will comment only on the first, since I do not know how
> > to avoid the second...
> > 
> > My experience is that if I type "Kmart.<b>" (note the trailing period
> > followed by the backspace) that the description field stays "Kmart". And
> > it does not need further additions on subsequent iterations. Paul, I
> > believe that was the point that Fred was making.
> 
> I believe that typing "Kmart.<bs>" is in every way equivalent to simply
> typing "Kmart".  Is that right?
> 
> No. If you have an imported transaction with "KMART" as the description and
> type in "Kmart", Gnucash will retain "KMART". once you put the period on
> the end, however, Gnucash reverts to "Kmart.", and when you backspace at
> this point, the form "Kmart" is retained. [My personal gripe is that this
> really is anti-Quickfill, but that's another thread altogether]

Yes, that's true.  I hadn't considered the case of an overlapping imported 
transaction.  And in that case, indeed, adding a period and backspacing over 
it is very useful


> > The issue of unwanted split configuration has come up before, and no one
> > has proposed a viable solution that meets everyone's needs.
> 
> Cancelling the quickfill by whatever method would perforce cancel an
> unwanted split.  And the method I recommend is that a trailing space <s>
> does the cancellation.
> 
> As far as what one wants: if the memorized transaction was a split, then
> there are three possibilities: you want the same split as before, you want
> a different split, or you want no split  In the first case, quickfill does
> exactly the right thing; in the third case, cancelling the quickfill does
> the right thing.  The second case is squishier -- is it easier to edit the
> previous split or just construct a new one?  Really, the user should have
> the choice in that case, and a quickfill cancellation convention provides
> that choice.
> 
> I'll note for the record that in Gnucash, every transaction has at least one
> split, so I don't quite understand your third case. As for your squishy
> middle case, how exactly do you propose implementing this? As I noted, this
> has come up in earlier threads, without anyone arriving at a solution that
> meets people's various needs--never mind anyone actually offering code to
> implement that ethereal solution.
> 
Again, I should have been more precise.   I was talking about the case where 
the split is explicitly shown as such, with the notation "Split Transaction".   
If you don't want a split, then just cancel the quickfill.  If you want the 
same split as before, allow the quickfill to proceed.

As to the squishy case: if you want to keep the split but modify it, then 
don't cancel the quickfill.  You can then modify the old split as desired.
That would often be easier than reconstructing the split from scratch.



More information about the gnucash-user mailing list