Re: psql: show only failed queries

From: Rajeev rastogi <rajeev(dot)rastogi(at)huawei(dot)com>
To: Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>
Cc: Samrat Revagade <revagade(dot)samrat(at)gmail(dot)com>, Peter Eisentraut <peter_e(at)gmx(dot)net>, Fabrízio Mello <fabriziomello(at)gmail(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: psql: show only failed queries
Date: 2014-06-30 09:20:41
Message-ID: BF2827DCCE55594C8D7A8F7FFD3AB7713DE10F92@SZXEML508-MBX.china.huawei.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 30 June 2014 12:24, Pavel Stehule Wrote:

>>I have reviewed this patch. Please find my review comments below:

>>1. Command start-up option (e.g. -a/--echo-all for --ECHO=all), for new functionality is not provided.
>all not options entered via psql variables has psql option and psql comment. I'll plan add new decription to --help-variables list.
>If it is necessary I can add long option --echo-errors, I didn't a good char for short option. Any idea?

But the new option we are adding are on a track of existing option, so better we add start-up option for this also.
Yeah long option –echo-errors seems to be fine to me also. For short option, I think we can use “-b” stands for blunder. This is the closest one I could think of.

>>2. New Command start-up option should be added in "psql --help" as well as in documentation.
>depends on previous,
Right.
>>Also as I understand, this new option is kind of sub-set of existing option (ECHO=query), so should not we display
>>query string in the same format as it was getting printed earlier.
>>Though I also feel that prefixing query with STATEMENT word will be helpful to grep but at the same time I am worried
>>about inconsistency with existing option.

>This is question. And I am not strong in feeling what should be preferred. But still I am inclined to prefer a variant with STATEMENT prefix. Mode with -a is used with different purpose than mode "show errors only" - and output with prefix is much
> more consistent with log entry - and displaying error. So I agree, so there is potential inconsistency (but nowhere is output defined), but this output is more practical, when you are concentrated to error's processing.
Yeah right, I just wanted to raise point to provoke other thought to see if anyone having different opinion. If no objection from others, we can go ahead with the current prefixing approach.
Thanks and Regards,
Kumar Rajeev Rastogi

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Andres Freund 2014-06-30 09:24:01 Re: better atomics - v0.5
Previous Message Ronan Dunklau 2014-06-30 09:01:48 Re: IMPORT FOREIGN SCHEMA statement