> From: blymn@baea.com.au (Brett Lymn)
>
>
> This is a bit over the top.  Have you actually seen the diff to fix
> the pr you reported?  It was a change of the WINDOW structure - moving
> one int to the end to make the memory layout the same as it was
> before, only longer.  The fact that vi grovels through the curses
> internal data structures is somewhat of a concern to me.

Sorry if it sounded like "the sky is falling". FWIW, I wish I had phrased it
all more carefully -- sorry!

But on the technical point, you seem to be missing something ...  curses.h
will access those fields via macros!  It exports the structure, and then
compiles into the client open-coded references to the fields.

*Every* binary "grovels through the curses internal data structures" as a
result.  It's not "vi groveling" at all.

> >The libcurses update done today, 6/28 is INCOMPATIBLE with the update done
> >4/13;
>
> True, it will but you cannot have it all.

Of course not. I wasn't complaining at all there, just notifying the users
about the moving interface. Sorry--I didn't mean that to sound so negative.
You are allowed to break -current, my only points were w.r.t notice and also
a possibly overstated point about testing.

> Vi broke because of a habit you have that I do not share.  I rarely ^Z
> out of vi because I either use X or virtual consoles.  It was not an
> obvious test to me, one of the other people on the curses team did
> spot the problem and we were working on a solution.  The problem was
> hardly a show stopper - if you went forward a page the screen would
> update properly.


Well, OK, but it did affect more than just vi, and more than just me.

And actually, I do usually edit in 50-70 line xterms, in a virtual desktop.
That actually makes this problem worse in a way, because if the file is
less than 60 lines (and most of the admin files are) then I *can't* space
forward or backwards, and if the file has been modified, I can't even quit
and restart easily.

Anyway, overall I'm sure our libcurses is much nicer now. I'm a bit confused
as to why the PR-related comments count as notices about when a mega-change
actually hits -current. (Really, Matt should have done that.) If you had
said "read source-changes", I could have no argument, but "read the PR and
magically know when it got checked in" ... I dunno.

Anyway, all's well that ends well, etc...

	Ross.Harvey@Computer.Org