>>Well, I may have finally found what my C0000005 problem has been. After days of testing for this problem alone, I've been able to recreate it. I got a more meaningful error message this time, and I have determined that it involves a simple index file being corrupted-- reindexing(DEL *.IDX and then INDEX ON again) fixes the problem. The only index open at this time(the one at fault) is a simple index on a logical field(PAID). The PAID field is changed in only 3-4 places in the program-- only once per record number. All it does is change it to .T. after locking the record(only if RLOCK returns true, of course). What could I be doing wrong to corrupt this index? Are there some rules in updating index keys that I'm just not following? Any ideas? Thanks.

>Is it possible in some place in your code, that you don't have the .IDX file >open when there are new records added to the table?

That is, of course, the first thing I checked(extremely thoroughly-- I do this every time I add an index). But even still, I should get VFP errors or application quirkiness when an IDX file is out of sync(believe me, I've done so). I'm not exactly sure what it was, but I tried putting a 'FLUSH' right after I changed the index key-- hopefully that will take care of it..


>Why are you using .IDX files, anyways? With .IDX's, you have to remember to >have them open when new records get added/changed, etc. CDX tags make much >more sense (and easier to remember!) today.

The application I develop was ported to VFP from Clipper Summer '87. I try to retain as much code compatibility as possible between our Windows and DOS versions. Within a year, however, I'm sure I'll be migrating to the new VFP features such as CDX's, etc. Maybe, we'll even go to an SQL server setup-- I can only hope. <g>