A quickfill problem and an enhancement suggestion
Paul Abrahams
abrahams at acm.org
Thu Apr 7 00:32:29 EDT 2011
On Wednesday, April 06, 2011 08:58:20 PM Lincoln A Baxter wrote:
> I've been following this debate because I too have be very annoyed at
> times by the quick fill behavior, though I love it when it fills the
> right thing, and since I know the behavior, I plan my descriptions so
> that I don't have to encounter the problem very often.
>
> This _is_ the problem with using backspace... BOTH as an initial
> keystroke, and after quickfill starts. So some character that one does
> not normally type when entering text makes a huge amount of sense. How
> about ANY configurable meta-character (of which escape could be one --
> maybe even the default)?
To me, having the functionality is far more important than tha particular
typing sequence used to realize it. Here's what I see as essential:
1. When you type the "cancel quickfill sequence" (which might consist of just a
single character), quickfill is cancelled for as long as control resides within
the current transaction, and any quickfill already performed is reversed. That
applies, very specifically, to the creation of splits.
2. The "cancel quickfill sequence" is something quite different from anything
one would type normally while entering a transaction. Both ESC and initial
backspace (or backspacing to the beginning of the transaction) meet that
criterion, but there are other possibilities.
If you decide you want the quickfill after all, you reenter the transaction.
And if quickfill works as advertised, that should be pretty easy.
Consider that quickfill for Walmart is unlikely to produce the right result,
but quickfill for a gas station is quite likely to produce the right result.
(Walmart haters can surely find another example.)
Paul
More information about the gnucash-user
mailing list