Re: Using defines for protocol characters

On 2023-Aug-09, Nathan Bossart wrote:

> On Wed, Aug 09, 2023 at 10:44:42AM -0600, Dave Cramer wrote:
> > On Wed, 9 Aug 2023 at 10:34, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> >> I agree with Peter: let's use the names in the protocol document
> >> with a single prefix.  I've got mixed feelings about whether that prefix
> >> should have an underscore, though.
> > 
> > Well, we're getting closer :)
> 
> I'm +0.5 for the underscore.

We use CamelCase_With_UnderScores in other places (PgStat_Counter,
AtEOXact_PgStat_Database).  It's not pretty, and at times it's not
comfortable to type, but I think it will be less ugly here than in those
other places (particularly because there'll be a single underscore), and
I also agree that readability is better with the underscore than
without.

So, I'm also +0.5 on having them, much as it displeases Robert.

-- 
Álvaro Herrera        Breisgau, Deutschland  —  https://www.EnterpriseDB.com/