On 14.10.25 13:13, Matěj Klonfar wrote:
certain instrumentation tools do prefix each statement with an
informational comment, typically to provide some tracing information to
logs (datadog for example). While this works for SQL statements, it's
not possible with logical replication statements because their grammar
doesn't support comments and it is causing unnecessary syntax errors.
I can imagine this limitation is likely a holdover from the system's
evolution from physical replication where comments make no sense.
However, in logical replication walsender mode both SQL and replication
statements can be issued [1], so the current state brings the necessity
to distinguish when to inject the comment and when not to. What do you
feel, are there any unexpected impacts of extending the replication
grammar with comments?
Another approach could be to get rid of repl_scanner.l and use the main
scanner. This would be similar to how plpgsql works.
