Thread: Supported languages

Supported languages

From
Thomas Hallgren
Date:
Joe Conway mentions five procedural languages shipped with PostgreSQL in
his article. Perl, Tcl, Python, Pgsql, and C. What would it take to add
Java to those languages? IMHO it would be rather natural given the large
precentage of PostgreSQL clients that use Java. Another strong argument
is the ability to write portable code.

Needless to say, I'd be happy to submit my Pl/Java project.

Kind regards,

Thomas Hallgren

Re: Supported languages

From
Josh Berkus
Date:
Thomas,

> Joe Conway mentions five procedural languages shipped with PostgreSQL in
> his article. Perl, Tcl, Python, Pgsql, and C. What would it take to add
> Java to those languages? IMHO it would be rather natural given the large
> precentage of PostgreSQL clients that use Java. Another strong argument
> is the ability to write portable code.
>
> Needless to say, I'd be happy to submit my Pl/Java project.

Is either PL/Java or PL/J ready for prime-time?    As far as I'm concerned,
any PL which is "production ready" should be in the same place.

However, I'd the impression that both you and Gavin were still bug-hunting.

--
-Josh Berkus
 Aglio Database Solutions
 San Francisco


Re: Supported languages

From
Thomas Hallgren
Date:
Josh Berkus wrote:

>Is either PL/Java or PL/J ready for prime-time?    As far as I'm concerned,
>any PL which is "production ready" should be in the same place.
>
>However, I'd the impression that both you and Gavin were still bug-hunting.
>
>
>
I have no known bugs at present but only a limited crowd of early
adopters uses Pl/Java. Consequently it has not been thoroughly tested
and to say that it's "production ready" would probably be a bit
optimistic. I would be very interested in working together with someone
who could write demanding tests. It's a known fact that tests written by
outsiders often find more bugs than tests written by the product developers.

 From a feature standpoint, I think Pl/Java is ready. It includes fully
functional support for functions, triggers, complex types (parameters
and return values), returning sets, a JDBC driver on top of SPI,
deployment/undeployment descriptors that executes SQL code, and more.

Right now I'm aligning Pl/Java with the upcoming 7.5 release and adding
gcj 3.4 as a possible choice of jvm (gcj is non proprietary and fits
right in from several other aspects as well).

I guess that "production ready" is a somewhat fuzzy measure. What would
it, in your opinion, take to claim that Pl/Java has reached it? Any
advice on how to get there?

Kind regards,

Thomas Hallgren


Re: Supported languages

From
Josh Berkus
Date:
Thomas,

> I guess that "production ready" is a somewhat fuzzy measure. What would
> it, in your opinion, take to claim that Pl/Java has reached it? Any
> advice on how to get there?

Well, it sounds like all you need is testing.   Also, it would have been nice
if you'd brought this up a few weeks earlier.   If you tell Tom, tonight,
that you want to move PL/Java into the main tree, I think the answer's going
to be "no" just because it's 12 hours to feature-freeze.

Also, it will precipitate an argument about whether the PLs really belong in
the main tree at all, or whether they should be seperate.   And that
*definitely* won't wind up in 12 hours.

--
Josh Berkus
Aglio Database Solutions
San Francisco

Re: Supported languages

From
Thomas Hallgren
Date:
Josh Berkus wrote:
> Thomas,
>
> Well, it sounds like all you need is testing.   Also, it would have been nice
> if you'd brought this up a few weeks earlier.   If you tell Tom, tonight,
> that you want to move PL/Java into the main tree, I think the answer's going
> to be "no" just because it's 12 hours to feature-freeze.
>
My intention was not to get it in for 7.5. Right now I'm more interested
in the procedure of getting there and if it's at all feasible to make a
submission. I think Pl/Java will be really good with gcj 3.4 and I don't
think it's been released on many platforms yet, so I'm in no real hurry.

> Also, it will precipitate an argument about whether the PLs really belong in
> the main tree at all, or whether they should be seperate.   And that
> *definitely* won't wind up in 12 hours.
>
Yes, I've seen discussions about that earlier. My objective is to get
Pl/Java promoted the same way the other languages are (i.e. one of the
"supported languages"). If it's in the main tree or not doesn't really
matter.

Kind regards,

Thomas Hallgren