Msg: 2399 *Conference*
04-17-92 13:16:30
From: LEX JENKINS
To : RICK HANSON
Subj: GREMLINS
Hi, Rick, I've run across some gremlins in using UR-II this week and wonder whether you've experienced or heard of these critters. While merging comma-delimited data files (created with T-Base) into T-Word files, I'm pushing T-Base to maximum capacity, but staying well within the limitations specified in the manual: i.e., observing the 255 character limit, minus number of fields, description lines, etc. I've used all of UR-II's features for several months and am pretty familiar with them. (And, knowing one of your pet peeves, I researched the manuals THOROUGHLY!!!) When I DISPlayed (F4) merged files prior to printing, and when I printed files, everything worked normally. However, when I used the PLOT (F2) option, the first plotted "screen" scrolled partially and then an "Out of Memory" error message appeared and the laptop locked up, but didn't cold start. Neither BREAK nor F8 (Menu) would unlock it, so I'd have to press reset. The UR-II menu would appear and I could continue working as if nothing had happened. Or so I thought. When I examined the .DO files GRPH and CODE characters had been randomly substituted for alpha-numeric characters. In the worst "crash" pressing reset sent me all the way back to the standard menu and EVERY file (.DO, .BA and .CO) had swapped filenames! All the files were intact. All filenames were intact. They'd just played musical chairs! After a while, even straight text files (no merging) wouldn't operate properly. When I'd press the DISPlay or PLOT keys it'd jump back to the UR-II menu rather than going into the document. Most of these incidents resulted in a few randomly substituted garbage characters. This occurred on both my 102 with Booster Pak and my "stock" 100 with extRAM, using UR-II on both. I cold-started the 100 and gave UR-II a test run, without trying any complex merging. It worked OK. However the gremlins have continued on my 102. I killed UR-II, cold started the RAM workspace and tried again: same problems. I finally decided to kill and reformat the entire RAMdisk on the Booster Pak and start over. Same problems, plus a new one: when I entered CLEAR 256, MAXRAM to clear HIMEM (which I've done many, many times without problem) the 102 either locked up or cold started. Let me summarize my questions: 1. Is crashing while trying to PLOT-preview a merged document one of those hidden "features" of UR-II you've alluded to? 2. Can crashes of this type write data to addresses above and below RAM which could compound problems? I recall you and Tracy Allen have mentioned that certain ROMs - Multiplan and some of PCSG's ROMs - can write data to the laptop's ROM space during NORMAL usage. Unless I misunderstand the concept, it seems possible that ROM-based programs could write data (even meaningless data) above and below user RAM during ABnormal circumstances. 3. Can crashes write data to, or delete data from, EPROMs? 4. Would errors of this sort be unaffected by cold starts? Would they defy the usual attempts at de-gremlin-ing? 5. Does this sound like a hardware problem, possibly in the UR-II ROM? Is there a utility, similar to RAMTST.BA or RAMCHK.BA, to test the integrity of ROMs? Finally, wouldn't you rather be enjoying a lovely Easter weekend? I know I would. I'll be playing in the dirt in the garden this weekend. Anything to get my face out of this screen for a while! <Lex>