[pgpool-hackers: 4087] Re: Fix regression test 075.detach_primary_left_down_node

[qianglj.fnst at fujitsu.com]

Hello Ishii san,

Thank you for your reply.

> I wonder why you use "grep -E" instead of egrep here. 

That reminds me, using egrep may be more reasonable.

> That maybe matter of taste but I just want to know.

You said it. Honestly I use "grep -E" more often than "egrep". What's more, I used "grep -E" because I want to keep consistent with the original test set as much as possible.

Best regards,
Qiang Lingjie

-----Original Message-----
From: Tatsuo Ishii <ishii at sraoss.co.jp> 
Sent: Tuesday, January 11, 2022 3:57 PM
To: Qiang, Lingjie/强 凌杰 <qianglj.fnst at fujitsu.com>
Cc: pgpool-hackers at pgpool.net
Subject: Re: [pgpool-hackers: 4085] Fix regression test 075.detach_primary_left_down_node

> Hello and Happy new year :),

Happy new year :-)

> I found the regression test case "075.detach_primary_left_down_node" was not working correctly.
> 
> If the "./startall" command on line 23 failed  for some reasons (port occupancy for example), the "$PGBIN/psql -c "show pool_nodes" test" command on line 36 would report errors resulting in "$PGBIN/psql -c "show pool_nodes" test|grep down" returned 1.
> 
> So this test case would finally output "ok" even if we had not start 
> the instance. The test purpose is not achieved. The detailed logs are 
> as follows：

I agree with you. Current test is broken. Thank you for pointing it out.

[snip]

> Fixed this by adding the STDOUT redirection and keyword error's judgment, details can be seen in the "0001-fix-regression-test-075.patch" attached.

I wonder why you use "grep -E" instead of egrep here. That maybe matter of taste but I just want to know.

Best reagards,
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese:http://www.sraoss.co.jp