Message-ID: <c8f93fc7-2f91-4130-99dc-de5fac734da8@technowledgy.de>
Date: Fri, 22 Mar 2024 09:02:04 +0100
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
From: Wolfgang Walther <walther@technowledgy.de>
Subject: Re: Regression tests fail with musl libc because libpq.so can't be
 loaded
Content-Language: en-US
To: Andrew Dunstan <andrew@dunslane.net>, Tom Lane <tgl@sss.pgh.pa.us>
Cc: Andres Freund <andres@anarazel.de>, Thomas Munro
 <thomas.munro@gmail.com>, Peter Eisentraut <peter@eisentraut.org>,
 Christophe Pettus <xof@thebuild.com>,
 PostgreSQL Bugs <pgsql-bugs@lists.postgresql.org>,
 Bruce Momjian <bruce@momjian.us>
References: <844665.1711088127@sss.pgh.pa.us>
 <2A001890-EDFA-4671-911D-42EF27B6B42F@dunslane.net>
In-Reply-To: <2A001890-EDFA-4671-911D-42EF27B6B42F@dunslane.net>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: 
 <https://www.postgresql.org/message-id/c8f93fc7-2f91-4130-99dc-de5fac734da8%40technowledgy.de>
Precedence: bulk

Andrew Dunstan:
>> On Mar 22, 2024, at 4:45 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> How was it passing?  The issue discussed in this thread has surely
>> been there for a long time, and Wolfgang mentioned that he sees
>> others.
> 
> The buildfarm client has a switch that delays running regression tests until after the install stages.

Hm. So while that switch makes the animal pass the build, it did hide 
exactly this problem. Not sure whether this switch should be used at 
all, then. Was this switch only implemented for the specific case of 
Alpine/musl or is there a different reason for it, as well?

The other issues I had been seeing were during make check-world, but not 
make check. Those were things around setlocale() / /bin/locale, IIRC. 
Not sure whether all of the tests are run by the buildfarm?

Best,

Wolfgang