[comp.windows.x] new escape sequences for xterm

flh@cbnewsl.att.com (frederick.l.horman..iii) (05/10/90)

The following problem was recently noted by an AT&T customer:

	The AT&T xterm recognizes some escape sequences that
	are not recognized by the MIT xterm.  In particular,
	two escape sequences were added to signal when a
	program enters/exits curses mode so that xterm can
	inhibit window resizing while in curses mode.  As a result,
	someone trying to rlogin to an SVR4 system (where
	the AT&T xterm has been installed -- with its terminfo
	entry) will get extraneous output when running a curses
	program. 

We (AT&T) would like to donate the added functionality to MIT, so
that the MIT xterm will recognize the escape sequences and behave
appropriately.  To do this with minimal change, I need to pick
a set of escape sequences to be used for the smcup/rmcup terminfo
capabilities which are used to signal enter/leave curses mode.

I have tentatively chosen to use the \E[< and \E[> sequences because
they are designated by ANSI for private use and are not currently
used by the MIT and AT&T xterms.  I would like feedback on this
choice of escape sequences.

If you see a conflict that I'm missing, please let me know.  Feel
free to respond to this newsgroup or directly to me.


Thanks,

Fred Horman
AT&T
flh@attunix.att.com

mouse@LARRY.MCRCIM.MCGILL.EDU (der Mouse) (06/02/90)

>> I think that the desired behavior would be for AT&T to modify their
>> applications that use curses to detect that the window has changed
>> size (i.e. with SIGWINCH), [...].

> What about all third-party applications that is the reason that most
> people goes with UNIX anyway?  AT&T could surely modify all *their*
> applications (if they haven't already done so in SVR4), but then
> there are all those obstinare third-party software companies that
> refuse to do so.

It seems to me that a third-party vendor failing to properly handle
window size changes is on a par with a third-party vendor failing to
properly use terminfo/termcap/curses and instead wiring in knowledge of
escape sequences.  Proper handling of SIGWINCH will simply become a
necessary part of a complete curses application, much the way setting
the input hint correctly is a necessary part of a complete X
application.

					der Mouse

			old: mcgill-vision!mouse
			new: mouse@larry.mcrcim.mcgill.edu