Re: BUG #15558: NOTIFY max channel length is undocumented

Oh, that's a good point.  Do you think a hyperlink like the following might be an improvement?

channel
      Name of the notification channel to be signaled (any <a
href="https://www.postgresql.org/docs/11/sql-syntax-lexical.html#SQL-SYNTAX-IDENTIFIERS">identifier</a>)

If there had been a hyperlink I believe I might have clicked it and found this out sooner.

On Tue, Dec 18, 2018, at 6:57 PM, David G. Johnston wrote:
> On Tue, Dec 18, 2018 at 5:36 PM PG Bug reporting form
> <noreply@postgresql.org> wrote:
> >
> > The following bug has been logged on the website:
> >
> > Bug reference:      15558
> > Logged by:          Andy Edwards
> > Email address:      jedwards@fastmail.com
> > PostgreSQL version: 11.1
> > Operating system:   (irrelevant)
> > Description:
> >
> > After debugging why notifications weren't getting delivered, I discovered
> > the following:
> >
> > # LISTEN "pg/tables/TagNotificationTriggers/tag/_org/1/JSON
> > Test/battery/1/lifetimeWh";
> > NOTICE:  identifier "pg/tables/TagNotificationTriggers/tag/_org/1/JSON
> > Test/battery/1/lifetimeWh" will be truncated to
> > "pg/tables/TagNotificationTriggers/tag/_org/1/JSON Test/battery/"
> >
> > I haven't seen a mention of this channel length limit in any version of the
> > documentation for NOTIFY.  I wish it were in the documentation, because then
> > I would have found out about this limit before deciding to use Postgres
> > notifications, and chosen to use Redis notifications from the beginning.
> 
> Since it is just another name/identifier it is the same limit and
> behavior as is present for table names, column names, type names,
> etc...
> 
> The length of identifiers is covered in the general section covering syntax:
> 
> https://www.postgresql.org/docs/11/sql-syntax-lexical.html#SQL-SYNTAX-IDENTIFIERS
> 
> Default length for identifiers is identified as being 63 bytes -
> though technically (but not usefully) is system dependent at compile
> time.
> 
> It is intentional to not spam the documentation with such a
> wide-ranging dynamic everywhere it might come into play...I'm
> undecided whether NOTIFY is deserving of an exception...
> 
> David J.