Re: [HACKERS] error detail when partition not found - Mailing list pgsql-hackers

From Robert Haas
Subject Re: [HACKERS] error detail when partition not found
Date
Msg-id CA+TgmoYaQc867QPGUh35fk89nsYsio1PifUz9jusm23FZQTL+A@mail.gmail.com
Whole thread Raw
In response to [HACKERS] error detail when partition not found  (Amit Langote <Langote_Amit_f8@lab.ntt.co.jp>)
Responses Re: [HACKERS] error detail when partition not found  (Amit Langote <Langote_Amit_f8@lab.ntt.co.jp>)
List pgsql-hackers
On Tue, Feb 21, 2017 at 7:28 AM, Amit Langote
<Langote_Amit_f8@lab.ntt.co.jp> wrote:
> Simon pointed out in a nearby thread [0] that the detail part of
> partition-not-found error should show just the partition keys.  I posted a
> patch on that thread [1], but to avoid confusion being caused by multitude
> of patches over there I'm re-posting it here.

Thanks.  GetPartitionFailureData seems like a strange name for a
datatype, particularly the "Get" part.  How about
PartitionRoutingFailureInfo?  Or just two out parameters.

Spelling: BuildSlotPartitinKeyDescription (in comment).

ExecBuildSlotPartitionKeyDescription could have a comment saying that
it's LIKE BuildIndexValueDescription() instead of copy-and-pasting the
comments. And maybe BuildIndexValueDescription() could also get a
comment saying that if we change anything there, we should check
whether ExecBuildSlotPartitionKeyDescription() needs a similar change.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company



pgsql-hackers by date:

Previous
From: Peter Geoghegan
Date:
Subject: Re: [HACKERS] Should logtape.c blocks be of type long?
Next
From: Robert Haas
Date:
Subject: Re: [HACKERS] Should logtape.c blocks be of type long?