PL/Julia: clarification on IN array parameters issue

On 2/10/26 1:40 PM, abdelsalam mostafa wrote:

I’m interested in contributing to PL/Julia and would like to better
understand this issue:
https://gitlab.com/pljulia/pljulia/-/issues/5 <https://gitlab.com/
pljulia/pljulia/-/issues/5>

The issue is about handling arrays passed as IN parameters, but I’m
still not fully clear on the problem and the expected behavior. I’d
really appreciate it if someone could explain what currently happens and
what is missing, at a high level.

Welcome to PostgreSQL Hackers!

Unfortunately this is not the mailing list for development of PL/Julia
so I do not think we can provide that good answer to your question.
Maybe Mark Wong will see your email here but in general do not expect
anyone here to know much about PL/Julia.

That said one thing you can do is to take a look at what other
procedural languages (e.g. plpython, plperl, plv8, ...) do when with
array arguments.

I’m planning to contribute to PL/Julia as part of GSoC 2026, and this
explanation would help me approach the issue correctly.

Good luck!

Andreas

On Tuesday, February 10, 2026, abdelsalam mostafa <
abdelsalam.lance@gmail.com> wrote:

Hi all,

I’m interested in contributing to PL/Julia and would like to better
understand this issue:
https://gitlab.com/pljulia/pljulia/-/issues/5

The issue is about handling arrays passed as IN parameters, but I’m still
not fully clear on the problem and the expected behavior. I’d really
appreciate it if someone could explain what currently happens and what is
missing, at a high level.

I’m planning to contribute to PL/Julia as part of GSoC 2026, and this
explanation would help me approach the issue correctly.

A quick skim of the code (with blame enabled) turned up this commit:

https://gitlab.com/pljulia/pljulia/-/commit/9bef4757ab491a792e049846341c683ba5993a7c

The wording of the issue is basically “missing feature addition”. The
commit happened a month after the issue was raised. Personally I’d have
closed the issue as resolved. Probably have added a “limitations” note to
the docs stating that only arrays of base types are presently accepted
(that may not be a true claim anymore though - I didn’t look).

Suggest you create some more test functions with different kinds of inputs,
and skim the test suite, to see what limitations do exist. Can you get the
PL to crash or bug out with a simple echo function with various inputs?

David J.

P.S. I’m confused how you ended up on -hackers for the core project
talking about GSoC for a non-core extension. Shouldn’t your mentor be
providing this kind of guidance? Also, it seems problematic to contribute
to an abandoned project as a mentee unless working with the last
maintainer/owner (though I’m not really up on the goals/guidelines GSoC
puts out in that respect either…).

On Tue, Feb 10, 2026 at 08:35:40PM -0800, David G. Johnston wrote:

P.S. I’m confused how you ended up on -hackers for the core project
talking about GSoC for a non-core extension. Shouldn’t your mentor be
providing this kind of guidance? Also, it seems problematic to contribute
to an abandoned project as a mentee unless working with the last
maintainer/owner (though I’m not really up on the goals/guidelines GSoC
puts out in that respect either…).

Yes, there are some enthusiastic prospective contributors. We are in touch now
so we're catching up. :)

Regards,
Mark
--
Mark Wong <markwkm@gmail.com>
EDB https://enterprisedb.com