Horton

Re: SQL:2023 JSON simplified accessor support

Chao Li <li.evan.chao@gmail.com>

View thread
From: Chao Li <li.evan.chao@gmail.com>
To: Alexandra Wang <alexandra.wang.oss@gmail.com>
Cc: jian he <jian.universality@gmail.com>, Nikita Malakhov <hukutoc@gmail.com>, Vik Fearing <vik@postgresfriends.org>, Mark Dilger <mark.dilger@enterprisedb.com>, Matheus Alcantara <matheusssilv97@gmail.com>, Peter Eisentraut <peter@eisentraut.org>, Andrew Dunstan <andrew@dunslane.net>, Nikita Glukhov <glukhov.n.a@gmail.com>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>, "David E. Wheeler" <david@justatheory.com>
Date: 2025-09-02T02:46:35Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Add test coverage for indirection transformation

  2. Fix typo in comment

  3. Implementation of subscripting for jsonb


> On Aug 27, 2025, at 15:26, Chao Li <li.evan.chao@gmail.com> wrote:
> 
> diff --git a/src/include/parser/parse_node.h b/src/include/parser/parse_node.h
> index f7d07c84542..58a4b9df157 100644
> --- a/src/include/parser/parse_node.h
> +++ b/src/include/parser/parse_node.h
> @@ -361,7 +361,7 @@ extern SubscriptingRef *transformContainerSubscripts(ParseState *pstate,
>                                                                                                          Node *containerBase,
>                                                                                                          Oid containerType,
>                                                                                                          int32 containerTypMod,
> -                                                                                                        List *indirection,
> +                                                                                                        List **indirection,
>                                                                                                          bool isAssignment);
> 
> Here we change “indirection” from * to **, I guess the reason is to indicate if the indirection list is fully processed. In case of fully processed, set it to NULL, then caller gets NULL via **.
> 
> As “indirection” is of type “List *”, can we just set “indirection->length = 0”? I checked pg_list.h, there is not a function or macro to cleanup a list (free elements and reset length to 0). There is a “list_free(List *list)”, but it will free “list” itself as well. Maybe we can add a new function, say “list_reset(List *list)” or “list_cleanup(List *list)”. This way will keep the function interface unchanged, and less changes would be involved.


After revisiting the changes, I take back this comment. Given the nature that List are always used in pointer form, my comment would just make things much more complicated.

Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/