Thread: Feature Request: DDL + RegExp - definitions

Feature Request: DDL + RegExp - definitions

From
Richard Broersma
Date:
It'd be nice if the catalog entries that store the DDL constraints
definitions and view definitions with regular expressions strings
would preserved any escape string syntax (i.e.E) for RegExp
backslashes.

For example, PGAdmin uses these catalog entries to produce Create or
Replace scripts to that allow one to alter table or view definitions.
However, when certain regular expressions syntax are added to these
definitions, the resulting scripts will not commit.

The following email illustrates the behavior that is undesirable (to me).
http://www.mail-archive.com/pgadmin-support@postgresql.org/msg11482.html

--
Regards,
Richard Broersma Jr.

Re: Feature Request: DDL + RegExp - definitions

From
Merlin Moncure
Date:
On Wed, Aug 24, 2011 at 11:30 AM, Richard Broersma
<richard.broersma@gmail.com> wrote:
> It'd be nice if the catalog entries that store the DDL constraints
> definitions and view definitions with regular expressions strings
> would preserved any escape string syntax (i.e.E) for RegExp
> backslashes.
>
> For example, PGAdmin uses these catalog entries to produce Create or
> Replace scripts to that allow one to alter table or view definitions.
> However, when certain regular expressions syntax are added to these
> definitions, the resulting scripts will not commit.
>
> The following email illustrates the behavior that is undesirable (to me).
> http://www.mail-archive.com/pgadmin-support@postgresql.org/msg11482.html

this is IMO not a feature, but a bug.

merlin

Re: Feature Request: DDL + RegExp - definitions

From
Richard Broersma
Date:
On Wed, Aug 24, 2011 at 10:14 AM, Merlin Moncure <mmoncure@gmail.com> wrote:

> this is IMO not a feature, but a bug.

Would the feature I'm requesting be a bug or is it the current
behavior that would be considered a bug?


--
Regards,
Richard Broersma Jr.

Re: Feature Request: DDL + RegExp - definitions

From
Tom Lane
Date:
Merlin Moncure <mmoncure@gmail.com> writes:
> On Wed, Aug 24, 2011 at 11:30 AM, Richard Broersma
>> The following email illustrates the behavior that is undesirable (to me).
>> http://www.mail-archive.com/pgadmin-support@postgresql.org/msg11482.html

> this is IMO not a feature, but a bug.

No, it's not a bug, it's just an unfortunate side-effect of the pushups
we've had to go through to get to standard conforming strings.

I seriously doubt that we would add a feature of the sort Richard
suggests, because with standard_conforming_strings turned on (which is
the default as of 9.1), there is no issue.  Forcing E'' syntax would
make the results of pg_getviewdef less standards-compliant, not more so,
and that's not a direction we want to go in.

If you find the warnings about backslashes annoying and unhelpful, you
can turn off escape_string_warning --- they're only there to help
identify code that is likely to have an issue when moved to an
environment with standard_conforming_strings turned on.

            regards, tom lane