R4 Xterm (and color Xterm) implementation woes under ISC Unix 2.2

barry hannigan bgh at ice9.uucp
Fri Sep 21 13:21:12 AEST 1990


has ANYONE managed to get an xterm client running against the R4 server
( R4 + patch [1-9] + vga patch + patch [10-14]) ? 

the problem is isolated to xterm, all other clients (except xauth) work fine.
 
it is compiled as follows using gcc 1.37.1
 
  gcc -DNOSTDHDRS -Datt -Dix386 -DSYSV -DUSG -DUSE_ASM -fstrength-reduce \
		-fpcc-struct-return -fwritable-strings ....

to compile successfully (?)

	assigned "discipline = TIOCSETD" instead of "discipline = d_disipline"
	around line 1367 as d_disipline was declared in code not included

	commented out the code included by the symbol TIOCSETC. this is an ioctl
	call, the call just before it is ok.

	the reapchild() is rife with nightmares. as it stands, because SIGTSTP and
	USE_SYSV_SIGNALS are defined, a call to wait3() is included. isnt wait3()
	a bsdism? ok, fine. forcing the inclusion of the alternate ifdef code, 
	using a wait() call. that doesnt work either. turning on debugging does 
	provide some relief, the client is mapped, cant do anything because 
	apparently as soon as it maps, it exits. with debug on, and the wait()
	active, you get zillions of "Exiting.." lines in the logfile.

	and after all this joy, nasty messages about out of swap, needed 1 page,
	which forced a nap for the machine. i question this though, swap wasnt 
	being used until this client, and then there was still lots. one thing at
	a time.

is there an alternative to this god awful code?

thanks, barry
-- 
~~~~~~~ ..walkin on the beaches, lookin at the peaches



More information about the Comp.unix.sysv386 mailing list