Re: Missing program_XXX calling in pgbench tests

On 05.06.25 05:00, Hayato Kuroda (Fujitsu) wrote:
> Dear Fujii-san,
> 
>> A bit similar discussion came up before regarding pgbench and program_xxx_ok
>> in [1], but it seems that change was never applied.
> 
> I didn't know that, thanks for sharing. ISTM, it tried to extend function to test the
> shorter options.
> 
> While verifying the idea, I found that pg_config and pg_bsd_indent have not been
> supported "-V" option yet. They must address the option ro improve test functions.
> 
> Attached patch set implemented the idea. 0001 is my original point and can be
> backported. 0002-0004 needs API changes so they aim to be applied for HEAD.
> 
> If possible, I want to fork another thread to discuss 0002-0004 and want to
> concentrate 0001 here.

Patch 0001 looks very sensible.

I don't think we need to bother we the other ones.  pg_config works 
differently than the other programs anyway, because --version does not 
exit the program.  And pg_bsd_indent is an externally maintained 
program.  So I think it is ok if these two are intentionally different.