kelso@mimsy.umd.edu (Stephen Kelley) (09/27/90)
First the good news (and possibly a repeat) as they say: PopUpMenu (vers 34) works *great* w/ 2.0! My thanks to the author(s). Next the bad news: Path (as in the 2.0 command) hangs the system if given >7? 8? (I had about 10) arguments. Really *weird*! The hdisk blinks rythmically seemingly looking and not finding something crutial. Actually it acts the same as it did when I installed my extra memory chips in the wrong slots. NO! I did *not* hose it up by pounding on it or dropping it (;-} After disabling the startup-seq. (thanks for *that* option CBM!) and breaking the "path ... add" into 2 separate commands it booted fine (and perceptably quicker, huh!). Is there a limit to the no. of parms in a shell? (Oh, Std. 2.0 shell btw). Other odd news: Has anybody had a problem w/ "ls" in 2.0? I switch (depending on shells, etc) between Manx's ls, Csh4.01, & 4.0. Anyway, *one* of then grabs *gobs* of cpu delaying keystrokes and even mouse movements. It does finish (depending on my patience) but it's definately cpu bound (no disk light while thrashing). I'll try to catch and document this one better, but I thot I'd mention what I had found while I'm at it. Oh, my 1st reaction on it is that I'd done 2 or more 'ls' *quickly* (or a cd & an ls) and dir/file locks were not released/checked correctly.
rick@rsami.spoami.com (Rick Schaeffer) (09/30/90)
In regard to ls in csh4.01 hanging up, I use a version of csh (derived from 2.07m) that I have hacked on and had the same problem when using it on 2.0. It turns out that the code that creates the linked list of directory entries was allocating one (1) byte too few. This turns out to create a problem because it causes the next byte in memory to be zeroed out. Since addresses on the 3000 under 2.0 actually use all 32 bits (apparently some or all of the extended memory is addressed at something above 7fffff) this causes a "delayed action" crash. Delayed because it is AMigados's memory arena that has been corrupted! Under 1.3 addresses aren't usually above 0x7fffff so this bug didn't cause trouble. I don't know if this is the problem with csh4.01 but, since the code dealing with file lists was straight out of the original csh program it's possible that it is. -- Rick Schaeffer UUCP: uunet!isc-br.isc-br.com!ricks ISC-Bunker Ramo ricks@isc-br.isc-br.com Box TAF-C8 Phone: (509)927-5114 Spokane, WA 99220