Re: multithreaded zstd backup compression for client and server - Mailing list pgsql-hackers

Hi Robert,

I haven't reviewed the meat of the patch in detail, but I noticed
something in the tests:

Robert Haas <robertmhaas@gmail.com> writes:
> diff --git a/src/bin/pg_verifybackup/t/009_extract.pl b/src/bin/pg_verifybackup/t/009_extract.pl
> index 9f9cc7540b..e17e7cad51 100644
> --- a/src/bin/pg_verifybackup/t/009_extract.pl
> +++ b/src/bin/pg_verifybackup/t/009_extract.pl
[…]
> +        if ($backup_stdout ne '')
> +        {
> +            print "# standard output was:\n$backup_stdout";
> +        }
> +        if ($backup_stderr ne '')
> +        {
> +            print "# standard error was:\n$backup_stderr";
> +        }
[…]
> diff --git a/src/bin/pg_verifybackup/t/010_client_untar.pl b/src/bin/pg_verifybackup/t/010_client_untar.pl
> index 487e30e826..5f6a4b9963 100644
> --- a/src/bin/pg_verifybackup/t/010_client_untar.pl
> +++ b/src/bin/pg_verifybackup/t/010_client_untar.pl
[…]
> +        if ($backup_stdout ne '')
> +        {
> +            print "# standard output was:\n$backup_stdout";
> +        }
> +        if ($backup_stderr ne '')
> +        {
> +            print "# standard error was:\n$backup_stderr";
> +        }

Per the TAP protocol, every line of non-test-result output should be
prefixed by "# ". The note() function does this for you, see
https://metacpan.org/pod/Test::More#Diagnostics for details.

- ilmari