On Thu, Jul 04, 2002 at 03:55:15PM +0200, Thomas Klausner wrote:
> On Thu, Jul 04, 2002 at 09:45:29AM -0400, Thomas Dickey wrote:
> > Picking the thread up late - it appears to indicate that you stated that
> > terminfo is not extensible.  I addressed that two years ago...
> 
> Sorry for that. Good to know. I was only relaying third party
> information, you might want to clear that up: The mailing address is
> 	tech-userlevel@netbsd.org
> and you don't have to be subscribed to mail.
> 
> If you prefer, I can forward your comment instead.
> [Reading soda's comment: So the binary format on disk changed?]

yes/no.  I noticed (actually - checking, it was February 1999 - 3 years) that
terminfo's structure is well-defined enough to ignore data in the terminfo file
past the data which is expected.  So I defined an extension which checks the
terminfo filesize and interprets extra data.  (That complicated tic and
infocmp, but added about 2kb to the normal executable).  If the extension is
not used, the terminfo file is identical to the previous version.

For ncurses 5.0, I made that optional (normally not compiled in), but for
ncurses 5.1 that's normally in.  After I had the feature in 5.0, I worked for a
few months improving the terminfo/termcap translations by adding the
termcap-only capabilities (mostly booleans) to the terminfo source.  That
version went into OpenBSD, btw.

Along different lines, FreeBSD's been using some changes I made to allow
tputs() and tparm() accept termcap strings (since they have a number of
applications with hardcoded strings inside).

-- 
Thomas E. Dickey <dickey@invisible-island.net>
http://invisible-island.net
ftp://invisible-island.net