On Fri, Jul 01, 2005 at 11:21:12PM +0300, Martin Vermeer wrote: > On Fri, Jul 01, 2005 at 09:23:59PM +0200, Andre Poenitz wrote:
... > > I don't agree with the idea of storing row and column in the cursor. > > This duplicates information and will lead to synchronization troubles. > > Actually I would have preferred to have _only_ row, col. But there are > dozens of statements in the code like "idx()= ..." with idx() on the > left side, which stop working without idx_. I didn't want to go there > at this point. > > (...and about "synchronization troubles": we already have those ;-) > > > Apart from that, the cursor is already far too fat to be nice. But > > that's my doing, I know. OK, I see it now: idx_type should be a struct with two int members, .row and .col. And suitable operator definitions added. No duplication of information, and still "idx() = ..." will work. Does that sound like a plan? Not for 1.4 though. - Martin
pgpk217g1riqe.pgp
Description: PGP signature
