On Oct 6, Thomas Lockhart mentioned:
> The following example shows psql correctly clearing its input buffer
> when a line containing *only* a comment is seen, but not completely
> clearing the buffer (or not realizing that it is cleared; note the
> changed prompt) if the comment is at the end of a valid query.
>
> postgres=> -- comment
> postgres=> select 'hi'; -- comment
> ?column?
> --------
> hi
> (1 row)
>
> postgres->
That has been noted by me as well. From looking at the code I see that
someone intended to do something quite different in this case, like print
the comment on top of the query being echoed, I think. But I couldn't
really follow that.
Anyway, I'm going to end up rewriting that parser anyway, so that will be
taken care of. I was almost about to use flex but the Windows crowd
probably wouldn't find that too funny. (The Windows crowd won't find this
thing funny anyway, since I have no clue what #ifdef's I need for that.
Someone else will have to do a looong compile&fix session.)
The question I have though is, is there a reason, besides efficiency, that
psql doesn't just send the comment to the backend with the query? The
backend does accept comments last time I checked. Perhaps someone will one
day write something that makes some use of those comments on the backend
(thus conflicting with the very definition of "comment", but maybe a
logger) and it would remove some load out of psql.
--
Peter Eisentraut - peter_e@gmx.net
http://yi.org/peter-e/
