Обсуждение: Extending Regular Expression bounds limit of 255.

In the docs <a
href="http://www.postgresql.org/docs/8.3/interactive/functions-matching.html#POSIX-EMBEDDED-OPTIONS-TABLE">http://www.postgresql.org/docs/8.3/interactive/functions-matching.html#POSIX-EMBEDDED-OPTIONS-TABLE</a><br
/>it says that Regular Expression bounds {m,n} that m and n can be 0-255.  Is there a way to extend the upper limit. 
Weare trying to be consistent between Regular Expressionimplementation an the other implementations do not have this
limit.<br/><br clear="all" />Brent DeSpain<br />Schweitzer Engineering Laboratories, Inc.<br /> 

On 28 May 2010 07:33, Brent DeSpain <bd.postgres@gmail.com> wrote:
> In the docs
> http://www.postgresql.org/docs/8.3/interactive/functions-matching.html#POSIX-EMBEDDED-OPTIONS-TABLE
> it says that Regular Expression bounds {m,n} that m and n can be 0-255.  Is
> there a way to extend the upper limit.  We are trying to be consistent
> between Regular Expressionimplementation an the other implementations do not
> have this limit.
Interesting.  The POSIX standard for REs dictates this limit; which
implementations
you're using don't adhere to it?


Cheers,
Andrej




--
Please don't top post, and don't use HTML e-Mail :}  Make your quotes concise.

http://www.american.edu/econ/notes/htmlmail.htm

Excerpts from Brent DeSpain's message of jue may 27 18:55:55 -0400 2010:
> It looks like most of our tools are using the Perl version of regular
> expressions with an upper limit of a bound being 32766.  Is there any way to
> change this in PG?  Or can I change from POSIX to Perl?

See regcomp.c; you'll have to recompile.  The interesting constant is
DUPMAX (defined in regguts.h as per POSIX), but note that there's
arithmetic in repeat() with those values.  I wouldn't recommend doing
any of this unless you have time to figure out whether the algorithms
are efficient for larger values though.

-- 
Álvaro Herrera <alvherre@commandprompt.com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support

On 2010-05-27, Brent DeSpain <bd.postgres@gmail.com> wrote:
> --00504502c13812967604879b4ba3
> Content-Type: text/plain; charset=ISO-8859-1
>
> It looks like most of our tools are using the Perl version of regular
> expressions with an upper limit of a bound being 32766.  Is there any way to
> change this in PG?  Or can I change from POSIX to Perl?

perhaps you can do something in pl-perl?

posix regular expressions are different to perl regular expressions in
several ways.

Jasen Betts <jasen@xnet.co.nz> wrote:

>> It looks like most of our tools are using the Perl version of regular
>> expressions with an upper limit of a bound being 32766.  Is there any way to
>> change this in PG?  Or can I change from POSIX to Perl?

> perhaps you can do something in pl-perl?

> posix regular expressions are different to perl regular expressions in
> several ways.

Another last resort possibility would of course be to "pre-
compile" the regular expressions from "A{2000}" to
"A{255}A{255}A{255}A{255}A{255}A{255}A{255}A{215}" (with the
headaches of "A{1000,2000}" left as an exercise to the read-
er :-)).

Tim

On 2010-05-29, Tim Landscheidt <tim@tim-landscheidt.de> wrote:
> Jasen Betts <jasen@xnet.co.nz> wrote:
>
>>> It looks like most of our tools are using the Perl version of regular
>>> expressions with an upper limit of a bound being 32766.  Is there any way to
>>> change this in PG?  Or can I change from POSIX to Perl?
>
>> perhaps you can do something in pl-perl?
>
>> posix regular expressions are different to perl regular expressions in
>> several ways.
>
> Another last resort possibility would of course be to "pre-
> compile" the regular expressions from "A{2000}" to
> "A{255}A{255}A{255}A{255}A{255}A{255}A{255}A{215}" 
(A{200}){10}might work better.

> (with the headaches of "A{1000,2000}" left as an exercise to the read-
> er :-)).

easy enoungh to write, but probably easy to write an expression with
factorial complexity too, but this one should work, having at worst
two ways of matching any string.
(A{200}){5,9}A{0,200} 

but 
(A{5,10}){200}
is asking for trouble.