Open simosentissi opened 7 years ago
I've actually been thinking of rewriting the whole extension in C++, to get rid of all dependencies :)
Basically, the one function that needs Python/plv8/etc is the function which does more involved json processing, which isn't really possible in pl/pgsql.
Actually, now that I'm thinking of it, I could just replace that one function with a native/compiled function.
Anyway, I'll make this a priority soon, hopefully. Of course I'd be open to a pull-request too.
Oh, I should clarify, I'm not fundamentally opposed to a plv8 version, but the one obstruction I found was that plv8 isn't necessarily available in common package managers without pulling in third-party repos.
If that situation has changed we could re-assess.
I think you are correct in your train of thought/reasoning, however looking into the postgresql adoption from service providers like aws/compose/.... (it seems like there is a growing #) they all limit the extensions to whatever they deem more secure .... so to be pragmatic and increase adoption I am thinking that a second look into what are the most common PLs among these providers is probably due in order to increase adoption. I looked into your code much deeper and it seems like one function is involvec, once your start on conversion to a language I amet familiar with I will be more than to contribute & test on aws and compose along with my team!
Currently most cloud offerings do not allow for python based extensions (example: aws RDS) but allow for plv8. Any plans to break that dependency ? I noticed that there is only one function with that dependency...
Thanks -Simo