Re: pgsql: psql: add an optional execution-count limit to \watch. - Mailing list pgsql-hackers

From Andrew Dunstan
Subject Re: pgsql: psql: add an optional execution-count limit to \watch.
Date
Msg-id 58d71f37-a353-408f-950d-0c759e662d81@dunslane.net
Whole thread Raw
List pgsql-hackers
On 2023-04-07 Fr 10:00, Tom Lane wrote:
> Alexander Korotkov<aekorotkov@gmail.com>  writes:
>> On Thu, Apr 6, 2023 at 8:18 PM Tom Lane<tgl@sss.pgh.pa.us>  wrote:
>>> psql: add an optional execution-count limit to \watch.
>> This commit makes tests fail for me.  psql parses 'i' option of
>> '\watch' using locale-aware strtod(), but 001_basic.pl uses hard-coded
>> decimal separator.
> Huh, yeah, I see it too if I set LANG=ru_RU.utf8 before running psql's
> TAP tests.  It seems unfortunate that none of the buildfarm has noticed
> this.  I guess all the TAP tests are run under C locale?


[just noticed this, redirecting to -hackers]


When run under meson, yes unless the LANG/LC_* settings are explicitly 
in the build_env. I'm fixing that so we will allow them to pass through. 
When run with configure/make they run with whatever is in the calling 
environment unless overridden in the build_env.

We do have support for running installchecks with multiple locales.This 
is done by passing --locale=foo to initdb.

We could locale-enable the non-install checks (for meson builds, that's 
the 'misc-check' step, for configure/make builds it's more or less 
everything between the install stages and the (first) initdb step. We'd 
have to do that via appropriate environment settings, I guess. Would it 
be enough to set LANG, or do we need to set the LC_foo settings 
individually? Not sure how we manage it on Windows. Maybe just not 
enable it for the first go-round.


cheers


andrew


--
Andrew Dunstan
EDB: https://www.enterprisedb.com




pgsql-hackers by date:

Previous
From: Peter Eisentraut
Date:
Subject: Re: Use XLOG_CONTROL_FILE macro everywhere?
Next
From: Masahiko Sawada
Date:
Subject: Re: New committers: Melanie Plageman, Richard Guo