On 11/6/07, Reg Me Please <regmeplease@gmail.com> wrote:
> Il Tuesday 06 November 2007 22:13:15 hai scritto:
> > On 11/6/07, Reg Me Please <regmeplease@gmail.com> wrote:
> > > Il Tuesday 06 November 2007 19:43:38 Scott Marlowe ha scritto:
> > > > On 11/6/07, Reg Me Please <regmeplease@gmail.com> wrote:
> > > > > That seems not to be the case.
> > > > > The last line has a \. by its own and the last but one is
> > > > > well formed.
> > > >
> > > > (Please don't top post...)
> > > >
> > > > Got a self contained test case you can post?
> > >
> > > Back to the original topic ...
> > >
> > > I'm trying to understand what and where.
> > > The point is that I have this 29M+ lines script telling me there's a
> > > problem somewhere.
> > >
> > > A self contained test, at the moment, would be that long!
> > >
> > > I'm considering a "branch and bound" approach ... but it'd be quite long
> > > and tedious as the program generating the script has not been written
> > > to do such things.
> >
> > Split it in half, with the appropriate sql on each end so the data
> > still works, and see which half causes a problem. Keep splitting the
> > one that causes a problem in half until you have a small one with the
> > problem still.
> >
> > I'm guessing the problem will become obvious then.
>
> That's the "branch and bound". Editing 29M+ lines file takes some time.
> But this is the way I'm going to go right now.
Oh, we called it half-splitting in the military.
Using something like head / tail in unix to do it. Should be fairly
fast, especially if you keep cutting it in half after the first test.
I can't imagine editing something that large even in vi being very
fast.
