cheung@vu-vlsi.UUCP (Wilson Cheung) (07/30/87)
Its been said that resource tracking would have placed too much of an overhead on the Amiga's operating system. Well with 2 megs and 5 different applications set up this overhead would be tolerable for me, although I fail to see what type of overhead other than memory usage this would cause. Take the following scenario I have Scribble! up with a FORTH program and 3 additional include files in there own windows. In the background I am downloading a 100K file at 1200 baud. In the meantime I have a Multi-Forth window popped up. I compile the FORTH program with no problems. Then I run it. I Wait, and wait and wait. Darn! my program hung up FORTH probably a stupid endless loop. The Multi-Forth now eats up huge CPU time causing my mouse pointer to become jerky and sluggish. Now do I pop up another FORTH evironment and put up with a much slowed down 68000 or do I reboot, lose my file transfer, and the work in setting up my applications the way I wanted. Here it would be nice to have a certain key combination to invoke a resource/task manager that would allow a user to cancel a task and all the resouces associated with it. Using the keyboard rather than the mouse should generally guarantee that the user can still invoke the manager even if the locked application has locked up the mouse pointer as well. I guess what one would have to do to create the manager is to intercept all requests to access devices or allocate memory and store the information in its own table or list. Anyone up to the task? Wilson Cheung
bryce@COGSCI.BERKELEY.EDU (Bryce Nesbitt) (07/31/87)
In article <1012@vu-vlsi.UUCP> cheung@vu-vlsi.UUCP (Wilson Cheung) writes: > >...Take the following scenario... >...In the meantime I have a Multi-Forth window popped up. I compile the FORTH >program with no problems. Then I run it. I Wait, and wait and wait. Darn! >my program hung up FORTH probably a stupid endless loop. The Multi-Forth >now eats up huge CPU time causing my mouse pointer to become jerky and >sluggish. Now do I pop up another FORTH evironment and put up with a much >slowed down 68000 or do I reboot, lose my file transfer, and the work in >setting up my applications the way I wanted? Neither. You either ran the compile/development task at proirity -1, or bumped all your other tasks up somewhat. Depends on what crashed in FORTH, but you will probably regain nearly %100 CPU for the use of your other programs. If you forgot this step, you could change the priorities of everthing else after the fact. (changetaskpri will do this for a CLI, nothing I know of will do this for any task.) >...Using the keyboard rather than the mouse should generally guarantee >that the user can still invoke the manager even if the locked application has >locked up the mouse pointer as well. The mouse and keyboard input flow via nearly the same paths. With the exception of CTRL-AMIGA-AMIGA the keyboard has little in the way of advantage. Of course with the mouse, there's probably nothing to point at. |\ /| . Ack! (NAK, EOT, SOH) {o O} . ( " ) bryce@cogsci.berkeley.EDU -or- ucbvax!cogsci!bryce U "Success leads to stagnation; stagnation leads to failure."
ali@rocky.STANFORD.EDU (Ali Ozer) (07/31/87)
In article <1012@vu-vlsi.UUCP> cheung@vu-vlsi.UUCP (Wilson Cheung) writes: >my program hung up FORTH probably a stupid endless loop. The Multi-Forth >now eats up huge CPU time causing my mouse pointer to become jerky and >sluggish. >Now do I pop up another FORTH evironment and put up with a much slowed down >68000 or do I reboot, lose my file transfer, and the work in setting up my >applications the way I wanted. A solution that works well is to set the priority of the runaway or otherwise hung task to some low value --- Minus hundred works well. There are several programs to do this, I believe. I use one named "cpri." I think it appeared on a Fish disk, but browsing through a full index didn't reveal it. (I either remember wrong about "cpri" being from Fish disks or maybe I renamed it...) In any case, with a priority of -100, the task will not interfere with any other tasks (unless it is truly runaway and eating memory)... Ali Ozer, ali@rocky.stanford.edu
randy@bcsaic.UUCP (Randy Groves) (08/04/87)
In article <1012@vu-vlsi.UUCP> cheung@vu-vlsi.UUCP (Wilson Cheung) writes: >68000 or do I reboot, lose my file transfer, and the work in setting up my ^^^^ > > Wilson Cheung FINALLY!! Somebody who knows how to spell 'lose'. (NOT 'loose') -- -randy groves - Boeing Advanced Technology Center UUCP: ..!uw-beaver!uw-june!bcsaic!randy USNail: Boeing Computer Services CSNET: randy@boeing.com PO Box 24346 M/S 7L-68 VOICE: (206)865-3424 Seattle, WA 98124