Re: [HACKERS] psql \copy warning

Tom Lane wrote:
> Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
> > The attached patch fixes the warning you received by adding E'' strings
> > to the \copy arguments, and adds it for the other backslash commands
> > like \d.
>
> Further comment on this: I don't think we want all these places
> individually making this sort of decision. What they should all be
> doing is using appendStringLiteralConn to generate the properly-quoted
> literal. (I fixed this already in describe.c, but not in those other
> places.)

Yes, I think so.

> Once we've got that done, we could argue about whether appendStringLiteral
> ought to prepend an E to silence escape_string_warning. I'd still vote
> no, but at least it would be a single place to change and not N of 'em.
> What's more, each place that is generating a variable literal without
> using appendStringLiteral or PQescapeString is at least potentially
> vulnerable to encoding issues, and so we should probably convert them
> anyway.

True. See the email I just sent about escape_string_warning. FYI, we
are finding these places because of escape_string_warning. Do we trust
users to turn that on and test before standard_conforming_strings
becomes true. One big problem is that people are having to use E'' if
they want to keep using backslashes, even if they have already tested
standard_conforming_strings. One nice thing about E'' is that it works
no matter what the value of standard_conforming_strings is.

--
Bruce Momjian http://candle.pha.pa.us
EnterpriseDB http://www.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +