Re: Triage on old commitfest entries - JSON_PATH - Mailing list pgsql-hackers

From Pavel Stehule
Subject Re: Triage on old commitfest entries - JSON_PATH
Date
Msg-id CAFj8pRDCSCAzwJ-9nYYGeZANx4jOQTzRB9GFHOSb03Vc61szYA@mail.gmail.com
Whole thread Raw
In response to Re: Triage on old commitfest entries - JSON_PATH  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers


ne 3. 10. 2021 v 22:16 odesílatel Tom Lane <tgl@sss.pgh.pa.us> napsal:
Erik Rijkers <er@xs4all.nl> writes:
> Op 03-10-2021 om 21:14 schreef Tom Lane:
>> I looked at entries that are at least 10 CFs old, as indicated by
>> the handy sort field.  That's a pretty small population: 16 items
>> out of the 317 listed in the 2021-09 CF.  A quick look in recent
>> CFs shows that it's very rare that we commit entries older than
>> 10 CFs.

> May I add one more?

> SQL/JSON: JSON_TABLE started 2018 (the commitfest page shows only 4 as
> 'Age in CFs' but that obviously can't be right)

Hm.  It's being actively worked on, so I wouldn't have proposed
killing it even if its age had been shown correctly.  Unless you
think it has no hope of ever reaching committability?

This is a pretty important feature and a nice patch.

Unfortunately, it is a pretty complex patch - JSON_TABLE is a really complex function, and this patch does complete implementation. I checked this patch more times, and I think it is good. There is only one problem - the size (there are not any problems in code, or in behaviour) . In MySQL or MariaDB, there is a much more simple implementation, that covers maybe 10% of standard. But it is available, and people can use it. Isn't it possible to reduce this patch to some basic functionality, and commit it quickly, and later commit step by step all parts.

Regards

Pavel




                        regards, tom lane


pgsql-hackers by date:

Previous
From: Pavel Stehule
Date:
Subject: Re: Triage on old commitfest entries
Next
From: Yura Sokolov
Date:
Subject: Re: BufferAlloc: don't take two simultaneous locks