Re: [GSoC] github repo and initial work - Mailing list pgsql-hackers

From Charles Cui
Subject Re: [GSoC] github repo and initial work
Date
Msg-id CA+SXE9uPEgDvz20kLc+rw3DJtxzeWQ5tGNRJ5rXR+iQ+M6aF+w@mail.gmail.com
Whole thread Raw
In response to Re: [GSoC] github repo and initial work  (Aleksandr Parfenov <a.parfenov@postgrespro.ru>)
List pgsql-hackers
Thanks for correcting me, will definitely study citext and see how a new type is registered in plugin. 


2018-05-24 23:23 GMT-07:00 Aleksandr Parfenov <a.parfenov@postgrespro.ru>:
On Thu, 24 May 2018 18:25:28 -0700
Charles Cui <charles.cui1984@gmail.com> wrote:
> The second is to provide thrift type just like json or jsonb. When you
> create a table, postgres knows ::thrift keywords.
> I think method one should be easier to implement because it only
> limits to this plugin. Method two needs modify postgres kernel to
> register a new type, which may time consuming,
> but more natural. Any ideas on this?
>
> Thanks, Charles

Hi Charles,

I prefer the second way with separate type. But I think it is good idea
to wait for an answer from your project mentor or someone other.

I'm not an expert in PostgreSQL user-defined types, but AFAIK, it
doesn't require changes in PostgreSQL core, since types can be created
in extensions. It doesn't require changes to grammar or something.
You can look at citext contrib as an example.

--
Aleksandr Parfenov
Postgres Professional: http://www.postgrespro.com
Russian Postgres Company

pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: Performance regression with PostgreSQL 11 and partitioning
Next
From: Charles Cui
Date:
Subject: Re: [GSoC] github repo and initial work