Paradox DOS, Unexpected Condition Codes

[PhilUK]

When Paradox meets an unexpected condition, the message often includes a GP code (before the "-Leaving Paradox" message).

We have found that 0000, usually relates to a damaged file (.lib library files and damaged forms have caused this).

Does anyone know what these codes mean. In particular, codes starting with 17 or 18 (then two digits or letters). We suffer these in various processes.

Thanks

Phil

 

[footpad]

PhilUK,

Unfortunately, they're memory addresses indication the location in memory where the problem occurred. this means there's little consistency between them. There are a few that are consistent, but not many.

I used to have a list, but that was many years ago. I believe that information is available on Corle's KnowledgeBase, but I'm not sure how easy it would be locate.

By and large, most problems in Paradox/DOS distill down to:

1) Not using -extk 16000 under Windows.
2) Not enough disk space in private.
3) Corrupted table/index.
4) Corrupted form/report.

Hope this helps...

-- Lance

 

[PhilUK]

Thanks Lance.

I located the Corel KnowledgeBase on the Corel Website, under support and Support Services button. A few useful documents, but nothing on the GP codes.

The document on Memory management reminded me it may be worth allocating less memory to the table buffer and more to the Central Memory Pool (CPM)as our processes do tend to be more variable than table intensive. We are currently at Buffer 13620k (88%) and CPM 1781 (12%).

Phil

 

[footpad]

PhilUK,

It's in there...somewhere. Corel's KB contains, in part, material from Borland's TOPIC database, which I contributed to while I worked there. It may, however, not be in an inuitive location.

As far as memory management goes, I would avoid messing around with any command-line settings other than -extk, which must be capped at 16000. While it is possible to optimize Paradox/DOS memory management for less than 16MB of RAM, most modern machines have far beyond that, even if only in virutual memory.

There's a slim possibility that you're hampering performance by mucking about with the code pool and the table buffers. I used to do this with 8MB and 16MB machines, but I don't believe it's necessary on machines with more memory.

Your mileage, of course, may vary.

-- Lance