Subject: Re: TAP tests and symlinks on Windows
To: Michael Paquier <michael@paquier.xyz>
References: <87sgf4pmb8.fsf@wibble.ilmari.org>
 <20200612065957.GA3362@paquier.xyz>
 <CAC+AXB0VJ9A4f6Ob9b3wV8V_VGND5X3MBLKQRd0+8L6t-gMiCQ@mail.gmail.com>
 <20200613060054.GM3362@paquier.xyz> <20200615062319.GF52676@paquier.xyz>
 <11c4ee09-3706-5912-ffd7-dc3657de7a5e@2ndQuadrant.com>
 <20200617074434.GB118592@paquier.xyz> <20200623105527.GA1974@paquier.xyz>
 <3ebcde2b-3b78-2e5f-9249-03c709dc23dd@2ndquadrant.com>
 <20200629075616.GC330401@paquier.xyz> <20200630121328.GB67289@paquier.xyz>
Cc: Andrew Dunstan <andrew.dunstan@2ndquadrant.com>,
 =?UTF-8?Q?Juan_Jos=c3=a9_Santamar=c3=ada_Flecha?=
 <juanjo.santamaria@gmail.com>, =?UTF-8?Q?Dagfinn_Ilmari_Manns=c3=a5ker?=
 <ilmari@ilmari.org>, pgsql-hackers <pgsql-hackers@postgresql.org>
From: Peter Eisentraut <peter.eisentraut@2ndquadrant.com>
Organization: 2ndQuadrant
Message-ID: <504f09d7-d600-44ca-a32b-0ed209cc2273@2ndquadrant.com>
Date: Fri, 3 Jul 2020 16:11:27 +0200
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.14; rv:52.0)
 Gecko/20100101 Thunderbird/52.9.1
MIME-Version: 1.0
In-Reply-To: <20200630121328.GB67289@paquier.xyz>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 2020-06-30 14:13, Michael Paquier wrote:
> Attached is an updated patch, where I have tried to use a better
> wording in all the code paths involved.

This new patch doesn't work for me on MSYS2 yet.

It fails right now in 010_pg_basebackup.pl at

     my $realTsDir      = TestLib::perl2host("$shorter_tempdir/tblspc1");

with chdir: No such file or directory.  Because perl2host requires the 
parent directory of the argument to exist, but here it doesn't.

If I add

     mkdir $shorter_tempdir;

above it, then it proceeds past that point, but then the CREATE 
TABLESPACE command fails with No such file or directory.  I think the call

     symlink "$tempdir", $shorter_tempdir;

creates a directory inside $shorter_tempdir, since it now exists, per my 
above change, rather than in place of $shorter_tempdir.

I think all of this is still a bit too fragile it needs further 
consideration.

-- 
Peter Eisentraut              http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services