Re: Usage of pg_waldump - Mailing list pgsql-bugs

From David G. Johnston
Subject Re: Usage of pg_waldump
Date
Msg-id CAKFQuwapABHOoHYrG7LR0TdwLp5a9_7Av0b+2_Fwi3+6pJAPXg@mail.gmail.com
Whole thread Raw
In response to Re: Usage of pg_waldump  (Andres Freund <andres@anarazel.de>)
Responses Re: Usage of pg_waldump  (Peter Eisentraut <peter.eisentraut@2ndquadrant.com>)
Re: Usage of pg_waldump  (Peter Eisentraut <peter.eisentraut@2ndquadrant.com>)
List pgsql-bugs
On Tue, May 1, 2018 at 9:26 AM, Andres Freund <andres@anarazel.de> wrote:
Hi,

On 2018-05-01 16:41:04 +0900, Dang Minh Huong wrote:
> The syntax is known as,
>
>   pg_waldump [OPTION]... [STARTSEG [ENDSEG]]
>
> Mean that, it could work with no arguments.
> But the "no arguments specified" ERROR will be returned in that case.
>
>   $ pg_waldump
>   pg_waldump: no arguments specified
>   Try "pg_waldump --help" for more information.
>
> I wonder it should be,
>
>   pg_waldump { OPTION... [STARTSEG [ENDSEG]] } | { [OPTION]... STARTSEG [ENDSEG] }
>
> Have any comments?

To me that doesn't seem like an improvement, looks a good chunk more
complicated. You'll get the error and the docs also tell you that some
arguments are needed.

​Reading the v10 docs...

​note: there is either an extra space between the two closing brackets or a missing space before startseg

I suppose it would help to understand why an argument is required - providing a startseg implicitly sets path and timeline but both of those have defaults and so even without a startseg those two aren't required.  Since no other options are required when startseg is specified it doesn't make sense that there would be others required when it is not.

IOW, the error message in the no-arg command invocation seems incorrect.

David J.

pgsql-bugs by date:

Previous
From: Andres Freund
Date:
Subject: Re: Usage of pg_waldump
Next
From: Tom Lane
Date:
Subject: Re: BUG #15180: Alter table add column if not exists with unique constraint will add extra duplicate