seurer+@rchland.ibm.com (Bill Seurer) (05/08/91)
I saw this all the time when I was using JPI v 1.17 (or was it 1.14?) on my floppy-only portable system. It seemed to occur when the disks were almost full and I was editting a large file. The editor writes out some temp files and if it can't, it croaks taking the disk with it. Not a very robust design. Swapping disks is also a definite No No since the editor leaves files open and with the way DOS works that means when you swap disks the old disk's FAT is written over the new disk. On my Big System I have had the JPI editor die and trash the files I was editting three times (in about a year) though it doesn't trash the whole harddisk (thank goodness!). It is obvious when this happens as the files on the screen are all jumbled up. If you immediately re-boot without doing anything else the copies on disk probably won't be trashed (though any changes are, of course, lost). I did this the last two times successfully. If the files do get trashed (as mine did the first time when I quit from the editor), run CHKDSK and save the disk area it finds as files. These files will have the correct source in them though it will be all mixed up. I managed to piece my source back together from these and what was left in the "real" files. It happens so infrequently that I haven't been able to see any pattern as to when it happens other than the files I was editting were very large. I only use 2.0 for a couple of programs that can't be compiled with 1.17 so I don't know if it suffers from the same problems. 2.0 won't work on a floppy-only system; it's just too large. - Bill Seurer Programming Support IBM Rochester, MN Prodigy: CNSX71A Internet: seurer@rchland.vnet.ibm.com
eric@zen.maths.uts.edu.au (Eric Lindsay) (05/08/91)
We have had many instances of damaged directory and fat tables on floppy disks used by our students running JPI Modula2 v1.14 and v1.16. The students are on a Novell 2.15 network, and normally don't have access to a hard disk, so all their work files are on floppy. I originally assumed that JPI were checking for a (hardware provided) disk change signal, which of course does not exist on older style XT disk controllers, and we reduced the incidence of problems by telling students never to remove their diskettes until after they had exited from Modula 2. Now I seem to have evidence of the same problem on an AT (which does have disk change signals). Has anyone else seen this? Is there anyone else actually running this software under similar conditions? Does anyone know if the same thing happens in Version 2? eric@zen.maths.uts.edu.au -- eric@zen.maths.uts.edu.au Eric Lindsay, Sch of Maths, Uni of Tech Don't take life too seriously. It is only temporary.