🚀 go-pugleaf

In article <2snn46$f2s@access3.digex.net> kbennett@access.digex.net (Keith R. Bennett) writes:


   An application I'm working on will run on a network, probably IBM's
   TCP/IP.  We're looking for a low cost, low overhead data base package to
   use.  Currently we're considering CodeBase for the following reasons:

   C/C++ support in OS/2

The C++ `wrapper' is a pile of crap.

   Industry-standard file format (xbase)

If you like fixed length records.

   Works with record/file locking of dBase/FoxBase/Clipper
                                           ^^^^^^^
Foxpro

   In addition to the manipulation of these files via our C++ programs, they 
   will be available to users to generate reports u sing off the shelf 
   packages such as Paradox.

   My questions:

   Has anyone worked with CodeBase?  What is your opinion of it?

You need to write your own interface class (C++), the supplied one is
terrible.  The report generator seemed ok (although difficult to use), but
I ended up replacing it because it used up too much memory.  It contains
some nasty routines (in the DOS version) that attempt to allocate 160k of
memory the first time you open an index file.  If the memory isn't available
it halves the amount and tries again.

   Can you recommend any other packages?

'fraid not - I wrote my own, but mine was application specific, also has fixed
length records, has slower indices and is owned by the company that was
contracting me at the time.... but it does run with as little as 512 bytes of
buffer space and supports virtual indices.

   Any other advice will also be greatly appreciated.

Now if I could only remember the name of the other suggested package....
B-trieve... that was it, I was told that B-trieve is shit-hot.

Good hunting.
--
Brian <brian@awfulhak.demon.co.uk>

Don't _EVER_ lose your sense of humour....