Re: Database I/O Modernization

Nathan

If I may, I think that you use RLA as a term differently from what I do - I think of RLA as native IO in RPG (the only language in which I use access to data - I do know about the C++ functions, however).

It sounds as if you refer to the DB component of the system when you use the term - this is probably causing some of the controversy here - talking at different levels.

Both SQL and native I/O use the same low-level SLIC functionality - these are exposed in the various higher-level calls - the GETM, GETK, GET, etc., kinds of things.

I suspect several of us are agreeing violently!

Cheers
Vern

On 9/17/2017 5:17 PM, Nathan Andelin wrote:

On Sun, Sep 17, 2017 at 8:45 AM, D*B <dieter.bender@xxxxxxxxxxxx> wrote:

that's a contradiction in itself.

I don't know what you mean. What is a contradiction in itself?

The rlaColGet() function is stateless in that like an HTTP server it
doesn't require any maintenance of state between calls. Stateless doesn't
mean the total absence of state (e.g. a file being open).

Since there may be many hundreds or thousands of calls to rlaColGet(), the
file(s) referenced and opened in the procedure remain open in order to
ensure quick runtime performance. The procedure will always produce the
right response, no matter where is may be called from.

Having a look to the complete Thread and an article of itjungle - could it
be, that you are a fabulist?

I had to look up the definition of fabulist. I don't recall ever hearing
that before.

No, I have a reputation for stating things in a forthright manner. I try
not to confuse or mislead anyone. What made you ask?

In regard to the SQE using RLA to access files, it might help for you to
review the RLA procedures exported from the QDBXIO service program. A lot
of the programs and service programs in the SQL runtime environment bind to
QDBXIO. Quite a few also bind to service program QC2IO, which contains the
C functions that I prefer to use for RLA.