Thread: Re: pgsql: Server-side gzip compression.

Re: pgsql: Server-side gzip compression.

From
Robert Haas
Date:
On Tue, Jan 25, 2022 at 4:20 AM David Rowley <dgrowleyml@gmail.com> wrote:
> On Tue, 25 Jan 2022 at 09:14, Robert Haas <rhaas@postgresql.org> wrote:
> > src/backend/replication/basebackup_gzip.c | 309 ++++++++++++++++++++++++++++++
>
> This could do with the attached.  MSVC compilers need a bit more
> reassurance that ereport/elog ERRORs don't return.

Err, well, if we need it, we need it. It surprises me, though:
wouldn't this same consideration apply to a very large number of other
places in the code base?

-- 
Robert Haas
EDB: http://www.enterprisedb.com



Re: pgsql: Server-side gzip compression.

From
David Rowley
Date:
On Wed, 26 Jan 2022 at 07:12, Robert Haas <robertmhaas@gmail.com> wrote:
> wouldn't this same consideration apply to a very large number of other
> places in the code base?

All of the other places are handled. See locations with "keep compiler quiet".

This one is the only one that generates a warning:

basebackup_gzip.c(90): warning C4715: 'bbsink_gzip_new': not all
control paths return a value

David



Re: pgsql: Server-side gzip compression.

From
Robert Haas
Date:
On Tue, Jan 25, 2022 at 3:56 PM David Rowley <dgrowleyml@gmail.com> wrote:
> On Wed, 26 Jan 2022 at 07:12, Robert Haas <robertmhaas@gmail.com> wrote:
> > wouldn't this same consideration apply to a very large number of other
> > places in the code base?
>
> All of the other places are handled. See locations with "keep compiler quiet".
>
> This one is the only one that generates a warning:
>
> basebackup_gzip.c(90): warning C4715: 'bbsink_gzip_new': not all
> control paths return a value

OK. I'm still surprised, but it is what it is. I've committed this now.

FWIW, I would have been fine with you just committing this change. I
can't see the warning locally, so I'm not in a position to
second-guess your statement that it's needed.

Thanks,

-- 
Robert Haas
EDB: http://www.enterprisedb.com



Re: pgsql: Server-side gzip compression.

From
David Rowley
Date:
On Fri, 28 Jan 2022 at 08:44, Robert Haas <robertmhaas@gmail.com> wrote:
> OK. I'm still surprised, but it is what it is. I've committed this now.

Thanks

> FWIW, I would have been fine with you just committing this change.

That's good to know, thanks for mentioning it.  FWIW, I just held back
as I wasn't 100% sure on the etiquette.

David