laser-lovers@uw-beaver (06/10/85)
From: Alan Crosswell <US.ALAN@CU20B.ARPA> We have a LaserWriter that for now is plugged into a Vax on the DB-25 connector and Appletalk on the DB-9 (No, we don't use both at the same time). Most of the time it's on the Vax, but occasionally it's switched to Appletalk. The problem is, that when on the Vax (rotary switch set to 9600), the DB-9 connector is active at 9600 baud and gets ioerrors anytime there is some Appletalk traffic. I figured I'd fix this by turning off the DB-9 port when in "batch" mode (that is, selector on 9600). So, I ran this program: 0 serverdict begin exitserver % exit server mode statusdict begin % open the status dictionary 9 0 0 setsccbatch % set port 9: 0 baud, 0 parity end This set the baud rate to 0 for the DB-9, which the Advanced Users Supplement of Inside LaserWriter claims is the proper method to turn off a port. At this point, the LaserWriter appeared to hang; no more jobs would print on it. Upon rebooting it, no startup page would come out. I was, however, able to connect to the DB-25 port with a terminal, enter executive mode and confirm that the sccbatch settings for port 9 had indeed been set to 0 baud, 0 parity and that the dostartpage boolean was still true. I was also able to get it to respond to a showpage command. I then set the baud rate back to 9600 for the DB-9 port and it rebooted fine: the startup page printed, and it accepted jobs just fine. Any thoughts on why this happened? My only conjecture is that the server is somehow getting goofed up because it normally polls both ports to select the next job and is getting wedged by one port being turned off. This doesn't explain why the startup page didn't print or why I was still able to have a conversation with the machine over the DB-25. Presumably, if I can talk to it, so can my spooler (TranScript's psif). Alan Crosswell Columbia University us.alan@cu20b.arpa -------
laser-lovers@uw-beaver (06/10/85)
From: adobe!taft@Glacier (Ed Taft) There appear to be at least two bugs that strike when serial channel 9 is turned off: (1) The start page doesn't print. This is due to a silly bug in the PostScript program for the start page: it is trying to compute the log of the baud rate to determine the height of the bar in the bar graph. If the baud rate is zero, this causes a rangecheck error. This bug has no side-effects other than failure to print the start page. (2) Something else goes wrong with serial communication on channel 25. Somehow this problem is not visible when communicating from a terminal but it fouls up spooling from TranScript, as you observed. I do not yet understand what the problem is. In view of these problems, I suggest that people not attempt to turn off channel 9; instead, disconnect the AppleTalk cable while running in serial mode. If I discover a software workaround for this bug, I will let you know. Sorry for any inconvenience this bug may have caused. Ed Taft Adobe Systems, Inc.