Re: Chainfail equivalent in SQL

David Gibbs wrote:

Aaron Bartell wrote:

<sarcasm-combater>
Or maybe do the following:

> D // just trying to ruin Aaron's day

D SQLFound c const(2001)

> D // oopsie

/free

if SQLFound;
// do something REALLY exciting
endif;

/end-free
</sarcasm-combater>

Just in the interest of combating the combater. :)

Actually, "if SQLFound" shouldn't compile - SQLFound is not a boolean expression. At best it would have to be something like SQLSTATE = SQLFound. Also, SQLFound is the wrong name; it should be SQLNotFound. Not busting Aaron's chops, just showing that making code pretty isn't always the answer, either.

On a more serious note ... I've never really liked the way RPG implemented embedded SQL. I would far prefer to have first class functionality for SQL instead of the pre-compiler 'hack' (as I see it).

In RPG's defense, the syntax for embedded SQL goes back a long way; I've seen papers from the early 90s on embedded SQL in C, and the syntax is exactly the same. Now, if your complaint is that you want all the logic of the SQL precompiler built into the RPG compiler in such a way that embedded SQL directly generates binary code, rather than acting as a macro pre-processor, I can sympathize with your position, but that won't change the ugliness of the embedded SQL syntax.

Joe