mark@spider.co.UK (Mark Valentine) (01/09/89)
I'm having a problem with xterm stopping halfway through a screen update, waking up (usually) after a few seconds or when I move the cursor out of the window or into the title bar. Details follow. Has anyone else seen (or solved!) this problem? I'm running X11r3 (+fixes 1&2 +purdue) on a pair of vaxen (Ultrix 2.2 on both). The server is running on a monochrome VAXstation (QVSS), and I run twm on a larger MicroVAX by way of an rsh invocation in my .xinitrc on the VAXstation (hmmm... however, the same thing happens when I start up twm directly from the other end). I start up my xterms from twm (i.e. using the VAXstation more or less as an X terminal, if you discount the initial rsh of twm). Occasionally, during a screen update (e.g. ^F in vi), output stops (usually only a few lines from the top of the page). This happens regardless of the size of my xterm. Usually, after a few seconds, the screen update will continue (it rarely sticks twice during a single update, despite the fact that it does so well up the page). Just about any event will wake it up during the time it's stuck (e.g. keyboard input, moving the mouse across a boundary such as into the title bar). The problem seems to be worse the more levels of software I stick in on the MicroVAX (?). For example, running "screen" (Oliver Laumann's multi-screen terminal emulator) in an xterm (I know, it's perverse, but I find the multiple screens easier to manipulate than the equivalent number of separate xterms), this occurs often enough to make it impractical. Even just running rn in an xterm, I have to jerk the mouse about every other article. On the network, the story goes like this (as told by a SpiderMonitor and interpreted by me)... At the point in question, the VAXstation has decided it's been swamped by the MicroVAX and has closed its TCP window. Five seconds later, the MicroVAX sends a probe byte, the VAXstation kicks back into life, and the pair then decide to speak DECspeak (?- type 1002 Ethernet packets) for a little while (at least in several packet traces these packets occurred just after the TCP recovery). Incidentally, the MicroVAX is the boot server for the VAXstation. Anyone with more knowledge of these bits and pieces care to point a finger? Mark. __ Mark Valentine, Spider Systems Limited, Edinburgh, UK. /\oo/\ <mark@spider.co.uk, mark%spider.co.uk@uunet.uu.net, uunet!mcvax!ukc!spider!mark>