Ken Johanson wrote:
> Bruce Momjian wrote:
> > E'' is more a marker than a type. I realize making E a type might work,
> > but it seems unusual.
> >
> > What we could do is backpatch E'' to 8.0.X as a no-op like it will be in
> > 8.1.
> >
>
> Bruce,
>
> Is it possible in the 8.1 betas to 'switch on' on the standard SQL
> escape behavior? This is from the use-case perspective of someone who
> does not have backwards compatibility concerns, rather, I'd like to
> preemptively forward-port / certify an app from another databases, onto
> PostgreSQL -- so all I need to do is switch that config on, if possible.
>
> From the changelog:
>
> "While this release does not change the <default?> handling of
> backslashes in strings, it does add new configuration parameters to help
> users migrate applications for future releases:
>
> o standard_conforming_strings ......
>
> o escape_string_warning ......
>
> The standard_conforming_strings value is read-only. ...."
>
> The last quoted sentence seems to answer my question (as no), but
> hopeful optimism is my motto :-)
>
> If it is indeed readonly, can it be made 'writable' before the 8.3
> release where is would be made the default behavior? For that matter, if
> the current backslash behavior stayed as the default for pre-8.3
> releases, and the patches are backported, I don't see any harm to the
> old-style apps/users; yet the correct behavior option is a useful
> "opt-in" one (one that I would like to try, now, even on 8.1).
The configuration parameter will be writable in 8.2. We would have
liked to have it writable in 8.1 but found it too complex to do in a
limited amount of time.
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073