I would not expect a java sp/trigger/udf language to be compiled into BLR
but possibly to have a way of calling BLR (or a replacement) through a
possibly simplified jdbc.

david jencks

On 2001.12.13 19:15:05 -0500 Frank Ingermann wrote:
> Hi Ann/Ivan/...,
>
> "Ann W. Harrison" wrote:
> >
> > At 03:41 PM 12/13/2001 +0100, Ivan Prenosil wrote:
> >
> > >or
> > > [DECLARE VARIABLES]
> > > A,B INTEGER;
> > > C,D CHAR(1);
> >
> > or
> > int create function (int a, char * b)
> > as
> > int x, y, z;
> > char c, d[2];
> > begin
>
> ...or:
>
> create function (a: Integer, b: ^char ):Integer;
> var
> x,y,z: Integer;
> c:Char;
> d:String[2];
> begin
> {...}
> end;
>
> ... i really did *not* want to start a "religious war" on the various
> language
> syntaxes here (honestly, i didn't! i'm still just a lazy typer ;-)
> ... the key question here imho is Jim's:
>
> <quote> Should the shared procedure/trigger
> language be an independent, distinct language, or should it adopt another
> language as a model, eventually to converge with. </quote>
>
> If it were to be independant (which is my current way of thinking) then
> my initial idea wouldn't cause any harm, and leave the sp lng mainly as
> it is (hence my proposal).
>
> If it were to adopt another lng, then i'm quite sure there will be
> endless
> discussions about *which* other lng we're talking about. *My* favourite
> would
> surely be Delphi, others (expectedly the greater part of the fb-devels)
> would certainly be in favour of C, Jim would probably prefer Java...
>
> ...BUT: it would all compile into the same BLR code, right?
>
> conclusion: [*not* serious!] : we need a Delphi2BLR, C2BLR and Java2BLR
> parser/compiler plus a new sql keyword "CODED_IN", so we can write:
>
> create function foo CODED_IN DELPHI as
> (x,y,z: integer):integer
>
> or
>
> create function foo2 CODED_IN C as
> int (int x,y,z)
>
> or
>
> create function foo3 CODED_IN JAVA as
> (sorry can't recall that calling syntax atm... <g>
>
> heck, all i wanted was to save some keystrokes! - but never mind, my
> ib/fb
> tools all have code completion (dv -> "declare variable")... <bg>
>
> regards,
> fingerman
>
> P.S. (to completely ruin my own proposal :-) would Borland InterBase
> support
> > declare variable i1,i2,i3,i4 integer;
> in their "open" source version ??
> (within this century ????)
>
> To unsubscribe from this group, send an email to:
> IB-Architect-unsubscribe@onelist.com
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>
>
>