Re: pg_basebackup check vs Windows file path limits

Andrew Dunstan <andrew@dunslane.net>

From: Andrew Dunstan <andrew@dunslane.net>
To: PostgreSQL-development <pgsql-hackers@postgresql.org>
Date: 2023-07-03T14:12:49Z
Lists: pgsql-hackers
On 2023-07-02 Su 09:15, Andrew Dunstan wrote:
>
> The buildfarm animal fairywren has been failing the tests for 
> pg_basebackup because it can't create a file with a path longer than 
> 255 chars. This has just been tripped because for release 16 it's 
> running TAP tests, and the branch name is part of the file path, and 
> "REL_16_STABLE" is longer than "HEAD". I did think of chdir'ing into 
> the directory to create the file, but experimentation shows that 
> doesn't solve matters. I also adjusted the machine's settings related 
> to long file names, but to no avail, so for now I propose to reduce 
> slightly the name of the long file so it still exercises the check for 
> file names longer than 100 but doesn't trip this up on fairywren. But 
> that's a bandaid. I don't have a good solution for now.
>

I've pushed a better solution, which creates the file via a short 
symlink. Experimentation on fairywren showed this working.


cheers


andrew

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

Commits

  1. Skip pg_baseback long filename test if path too long on Windows