Re: Teach pg_receivewal to use lz4 compression - Mailing list pgsql-hackers

From gkokolatos@pm.me
Subject Re: Teach pg_receivewal to use lz4 compression
Date
Msg-id S2rA_l89FgB0CtpkaBoPe0DTjvkvo46zUzlZxilMGzlDCICb5rwmrKXg1bzpp5ZO75GF-b7ayvG9dihbzRwOKVawLLreJbWWGfY6OdhSpho=@pm.me
Whole thread Raw
In response to Re: Teach pg_receivewal to use lz4 compression  (Magnus Hagander <magnus@hagander.net>)
Responses Re: Teach pg_receivewal to use lz4 compression  (Michael Paquier <michael@paquier.xyz>)
List pgsql-hackers

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐

On Thursday, July 1st, 2021 at 15:58, Magnus Hagander <magnus@hagander.net> wrote:

> On Thu, Jul 1, 2021 at 3:39 PM gkokolatos@pm.me wrote:
>
> > ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
> >
> > On Thursday, July 1st, 2021 at 12:28, Magnus Hagander magnus@hagander.net wrote:
> >
> > > On Wed, Jun 30, 2021 at 8:34 AM Dilip Kumar dilipbalaut@gmail.com wrote:
> > >
> > > > On Tue, Jun 29, 2021 at 8:15 PM gkokolatos@pm.me wrote:
> > > >
> > > > > Hi,
> > > > >
> > > > > The program pg_receivewal can use gzip compression to store the received WAL.
> > > > >
> > > > > This patch teaches it to be able to use lz4 compression if the binary is build
> > > > >
> > > > > using the -llz4 flag.
> > > >
> > > > +1 for the idea
> > > >
> > > > Some comments/suggestions on the patch
> > > >
> > > > @@ -90,7 +91,8 @@ usage(void)
> > > >
> > > > printf((" --synchronous flush write-ahead log immediately
> > > >
> > > > after writing\n"));
> > > >
> > > > printf((" -v, --verbose output verbose messages\n"));
> > > >
> > > > printf(_(" -V, --version output version information, then exit\n"));
> > > >
> > > > -   printf(_(" -Z, --compress=0-9 compress logs with given
> > > >
> > > >     compression level\n"));
> > > >
> > > > -   printf(_(" -I, --compress-program use this program for compression\n"));
> > > >
> > > >
> > > > Wouldn't it be better to call it compression method instead of
> > > >
> > > > compression program?
> > >
> > > I came here to say exactly that, just had to think up what I thought
> > >
> > > was the better name first. Either method or algorithm, but method
> > >
> > > seems like the much simpler choice and therefore better in this case.
> > >
> > > Should is also then not be --compression-method, rather than --compress-method?
> >
> > Not a problem. To be very transparent, I first looked what was already out there.
> >
> > For example `tar` is using
> >
> > -I, --use-compress-program=PROG
> >
> > yet the 'use-' bit would push the alignment of the --help output, so I removed it.
>
> I think the difference there is that tar actually calls an external
>
> program to do the work... And we are using the built-in library,
>
> right?

You are very correct :) I am not objecting the change at all. Just let you know
how I chose that. You know, naming is dead easy and all...

On a more serious note, what about the `-I` short flag? Should we keep it or
is there a better one to be used?

Micheal suggested on the same thread to move my entry in the help output so that
the output remains ordered. I would like the options for the compression method and
the already existing compression level to next to each other if possible. Then it
should be either 'X' or 'Y'.

Thoughts?



>
>
------------------------------------------------------------------------------------------------------------------------------------------------
>
> Magnus Hagander
>
> Me: https://www.hagander.net/
>
> Work: https://www.redpill-linpro.com/



pgsql-hackers by date:

Previous
From: Greg Sabino Mullane
Date:
Subject: Mention --enable-tap-tests in the TAP section page
Next
From: Simon Riggs
Date:
Subject: Re: PG 14 release notes, first draft