Re: Berkeley Regex - was: fixincludes jumbo patch 2

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Per Bothner]

Zack Weinberg <zack@rabi.columbia.edu> writes:

> I am nearly certain that Henry Spencer has a newer version of this code
> available but I was unable to find it.

Recent versions of Tcl include a newer version of his code:

/*
 * re_*exec and friends - match REs
 *
 * Copyright (c) 1998, 1999 Henry Spencer.  All rights reserved.
 * 
 * Development of this software was funded, in part, by Cray Research Inc.,
 * UUNET Communications Services Inc., Sun Microsystems Inc., and Scriptics
 * Corporation, none of whom are responsible for the results.  The author
 * thanks all of them. 
 * 
 * Redistribution and use in source and binary forms -- with or without
 * modification -- are permitted for any purpose, provided that
 * redistributions in source form retain this entire copyright notice and
 * indicate the origin and nature of any modifications.
 * 
 * I'd appreciate being given credit for this package in the documentation
 * of software which uses it, but that is not a requirement.
 * 
 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
 * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
 * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL
 * HENRY SPENCER BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
 * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
 * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
 * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 *
 */

-- 
	--Per Bothner
per@bothner.com   http://www.bothner.com/~per/

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Per Bothner]

Zack Weinberg <zack@rabi.columbia.edu> writes:

> I am nearly certain that Henry Spencer has a newer version of this code
> available but I was unable to find it.

Recent versions of Tcl include a newer version of his code:

/*
 * re_*exec and friends - match REs
 *
 * Copyright (c) 1998, 1999 Henry Spencer.  All rights reserved.
 * 
 * Development of this software was funded, in part, by Cray Research Inc.,
 * UUNET Communications Services Inc., Sun Microsystems Inc., and Scriptics
 * Corporation, none of whom are responsible for the results.  The author
 * thanks all of them. 
 * 
 * Redistribution and use in source and binary forms -- with or without
 * modification -- are permitted for any purpose, provided that
 * redistributions in source form retain this entire copyright notice and
 * indicate the origin and nature of any modifications.
 * 
 * I'd appreciate being given credit for this package in the documentation
 * of software which uses it, but that is not a requirement.
 * 
 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
 * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
 * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL
 * HENRY SPENCER BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
 * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
 * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
 * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
 * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 *
 */

-- 
	--Per Bothner
per@bothner.com   http://www.bothner.com/~per/

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Jeffrey A Law]

  In message <386B9E0B.1DA22D0B@datadesign.com>you write:
  > > I don't think we want to fight that kind of battle right now.
  > 
  > OK.  When?  ;-)  Are there issues with the above copyright?
In general, RMS wants all code in GCC to be assigned to the FSF.  He is willing
to make exceptions when no alternative exists.

Alternatives do exist in this case (gnu-regex) and I won't have the time to
argue with RMS about the virtues of one regex package over another.

jeff

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Martin v. Loewis]

> Are you sure that the regexp code is copyrighted by the UCB, not
> Henry Spencer? If the later only Henry could change the license,
> Berkeley cannot do it by fiat.

According to
http://www.FreeBSD.org/cgi/cvsweb.cgi/src/include/regex.h?rev=1.3 ,
it is

 * Copyright (c) 1992 Henry Spencer.
 * Copyright (c) 1992, 1993
 *      The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * Henry Spencer of the University of Toronto.

So I guess UCB can (and did) change the license on that code. Henry
might have other versions of that code that he did not contribute to
Berkeley.

Regards,
Martin

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Andi Kleen]

On Thu, Dec 30, 1999 at 02:10:23PM +0100, Martin v. Loewis wrote:
> > Are you sure that the regexp code is copyrighted by the UCB, not
> > Henry Spencer? If the later only Henry could change the license,
> > Berkeley cannot do it by fiat.
> 
> According to
> http://www.FreeBSD.org/cgi/cvsweb.cgi/src/include/regex.h?rev=1.3 ,
> it is
> 
>  * Copyright (c) 1992 Henry Spencer.
>  * Copyright (c) 1992, 1993
>  *      The Regents of the University of California.  All rights reserved.
>  *
>  * This code is derived from software contributed to Berkeley by
>  * Henry Spencer of the University of Toronto.
> 
> So I guess UCB can (and did) change the license on that code. Henry
> might have other versions of that code that he did not contribute to
> Berkeley.

All copyright owners have to agree to a license change AFAIK.

-Andi

-- 
This is like TV. I don't like TV.

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[John Wehle]

> Zack Weinberg submitted a fixincl patch for many things.
> I have asked him under private email to break up and
> simplify the patch.  I am pulling out one issue here
> in order to get consensus before installing.  The issue
> is whether or not the Berkeley regex processing is okay
> for inclusion or not.  It will apparently cut processing
> time by another 60% yielding a 7-fold improvement over the
> original script.  I don't get particularly excited over the
> Berkeley vs. GNU license debate, but I would hate to
> replace the gnu-regex only to have to put it back.

Another option is to use GNU rx which appears to be a more
current, faster, and correct regular expression library
than GNU regex.  I don't know how it compares to the Berkeley
implementation.

-- John
-------------------------------------------------------------------------
|   Feith Systems  |   Voice: 1-215-646-8000  |  Email: john@feith.com  |
|    John Wehle    |     Fax: 1-215-540-5495  |                         |
-------------------------------------------------------------------------

Berkeley Regex - was: fixincludes jumbo patch 2

[korbb]

This probably belongs on the GCC list more than gcc-patches:)


Zack Weinberg submitted a fixincl patch for many things.
I have asked him under private email to break up and
simplify the patch.  I am pulling out one issue here
in order to get consensus before installing.  The issue
is whether or not the Berkeley regex processing is okay
for inclusion or not.  It will apparently cut processing
time by another 60% yielding a 7-fold improvement over the
original script.  I don't get particularly excited over the
Berkeley vs. GNU license debate, but I would hate to
replace the gnu-regex only to have to put it back.

Comments, please?

Thanks,
	Bruce

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Martin v. Loewis]

> The issue is whether or not the Berkeley regex processing is okay
> for inclusion or not.

The good news is that earlier objections about incompatibilities
between GPL and the Berkeley license don't apply anymore, see

ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change

GPL requires that no additional restrictions may be made, and the only
additional restriction in BSD code has been revoked. I don't know for
sure whether the regex code falls under this rule, but I guess it
will.

Of course, you cannot expect the authors of this code assign their
copyright to the FSF, which might be a problem for inclusion into GCC.
The same problem would occur with any other general-purpose regex
engine that is not already copyright by the FSF.

Regards,
Martin

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Andi Kleen]

martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis) writes:

> > The issue is whether or not the Berkeley regex processing is okay
> > for inclusion or not.
> 
> The good news is that earlier objections about incompatibilities
> between GPL and the Berkeley license don't apply anymore, see
> 
> ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change
> 
> GPL requires that no additional restrictions may be made, and the only
> additional restriction in BSD code has been revoked. I don't know for
> sure whether the regex code falls under this rule, but I guess it
> will.

Are you sure that the regexp code is copyrighted by the UCB, not
Henry Spencer? If the later only Henry could change the license,
Berkeley cannot do it by fiat.

-Andi

-- 
This is like TV. I don't like TV.

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Harvey J. Stein]

Bruce Korb <korb@datadesign.com> writes:

 >   > The issue is whether or not the Berkeley regex processing
 >   > is okay for inclusion or not.  It will apparently cut
 >   > processing time by another 60% ...
 > 
 > Per Bothner writes:
 > > 
 > > > I am nearly certain that Henry Spencer has a newer version of
 > > > this code available but I was unable to find it.

I did some regex lib comparisons about a year ago (for guile).  At
that time Henry Spencer had a new version of the libs which were more
posix & had better character set support.  However, the library was
much slower (on the test cases I tried) than the version of Spencer's
library included with TCL.

The fastest was the regex library included with awk, which was snarfed
from some version of glibc.  The regex lib in glibc on my system
(linux, Redhat 5.x or 4.x at the time) was in between.  You can
probably find my timing tests in the guile mailing list archive.

-- 
Harvey Stein
Bloomberg LP
hjstein@bfr.co.il

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Alain Magloire]

Bonjour

> > I don't think we want to fight that kind of battle right now.
> 
> OK.  When?  ;-)  Are there issues with the above copyright?
> 
> > What I would suggest is that if the GNU regex routines are that lame is
> > for someone to get in touch with whomever maintains them and get them
> > improved.

GNU Regex, is now maintain inside GNU Lib C, with various contribution
by GNU *utils package using it ,  but you will not get much sympathy
from U. D., if you are just demanding for them to be improve without
any commitement on your part.

> 
> OK.  Who?  Henry?  What constraints are there on the fixincl code
> that are not covered by the above copyright?  What _is_ Henry's
> eaddress anyway? ...  :-}

   Henry Spencer
henry@zoo.toronto.edu 

But That was my last email exchange almost 3 years now.  If my memory
is correct, Mr Spencer implementation with locale etc .. was  more complete
then GNU Regex, But you will not gain anything more when it comes to speed.
By the way GNU regex also offers the full Posix API. Things will get quite
gruesome if you intented to modify the code to make it understand the full
locale range as define in Posix.


--
au revoir, alain
----
Aussi haut que l'on soit assis, on est toujours assis que sur son cul !!!

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Bruce Korb]

  > The issue is whether or not the Berkeley regex processing
  > is okay for inclusion or not.  It will apparently cut
  > processing time by another 60% ...

Per Bothner writes:
> 
> > I am nearly certain that Henry Spencer has a newer version of
> > this code available but I was unable to find it.
> 
> Recent versions of Tcl include a newer version of his code:
> 
> /*
>  * re_*exec and friends - match REs
>  *
>  * Copyright (c) 1998, 1999 Henry Spencer.  All rights reserved.
>  * 
>  * Development of this software was funded, in part, by Cray Research Inc.,
>  * UUNET Communications Services Inc., Sun Microsystems Inc., and Scriptics
>  * Corporation, none of whom are responsible for the results.  The author
>  * thanks all of them. 
>  * 
>  * Redistribution and use in source and binary forms -- with or without
>  * modification -- are permitted for any purpose, provided that
>  * redistributions in source form retain this entire copyright notice and
>  * indicate the origin and nature of any modifications.
>  * 
>  * I'd appreciate being given credit for this package in the documentation
>  * of software which uses it, but that is not a requirement.
>  * 
>  * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
>  * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
>  * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL
>  * HENRY SPENCER BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
>  * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
>  * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
>  * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
>  * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
>  * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
>  * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
>  *
>  */

Jeff Law wrote:
> I don't think we want to fight that kind of battle right now.

OK.  When?  ;-)  Are there issues with the above copyright?

> What I would suggest is that if the GNU regex routines are that lame is
> for someone to get in touch with whomever maintains them and get them
> improved.

OK.  Who?  Henry?  What constraints are there on the fixincl code
that are not covered by the above copyright?  What _is_ Henry's
eaddress anyway? ...  :-}

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Alain Magloire]

Bonjour

> > I don't think we want to fight that kind of battle right now.
> 
> OK.  When?  ;-)  Are there issues with the above copyright?
> 
> > What I would suggest is that if the GNU regex routines are that lame is
> > for someone to get in touch with whomever maintains them and get them
> > improved.

GNU Regex, is now maintain inside GNU Lib C, with various contribution
by GNU *utils package using it ,  but you will not get much sympathy
from U. D., if you are just demanding for them to be improve without
any commitement on your part.

> 
> OK.  Who?  Henry?  What constraints are there on the fixincl code
> that are not covered by the above copyright?  What _is_ Henry's
> eaddress anyway? ...  :-}

   Henry Spencer
henry@zoo.toronto.edu 

But That was my last email exchange almost 3 years now.  If my memory
is correct, Mr Spencer implementation with locale etc .. was  more complete
then GNU Regex, But you will not gain anything more when it comes to speed.
By the way GNU regex also offers the full Posix API. Things will get quite
gruesome if you intented to modify the code to make it understand the full
locale range as define in Posix.


--
au revoir, alain
----
Aussi haut que l'on soit assis, on est toujours assis que sur son cul !!!

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Harvey J. Stein]

Bruce Korb <korb@datadesign.com> writes:

 >   > The issue is whether or not the Berkeley regex processing
 >   > is okay for inclusion or not.  It will apparently cut
 >   > processing time by another 60% ...
 > 
 > Per Bothner writes:
 > > 
 > > > I am nearly certain that Henry Spencer has a newer version of
 > > > this code available but I was unable to find it.

I did some regex lib comparisons about a year ago (for guile).  At
that time Henry Spencer had a new version of the libs which were more
posix & had better character set support.  However, the library was
much slower (on the test cases I tried) than the version of Spencer's
library included with TCL.

The fastest was the regex library included with awk, which was snarfed
from some version of glibc.  The regex lib in glibc on my system
(linux, Redhat 5.x or 4.x at the time) was in between.  You can
probably find my timing tests in the guile mailing list archive.

-- 
Harvey Stein
Bloomberg LP
hjstein@bfr.co.il

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Bruce Korb]

  > The issue is whether or not the Berkeley regex processing
  > is okay for inclusion or not.  It will apparently cut
  > processing time by another 60% ...

Per Bothner writes:
> 
> > I am nearly certain that Henry Spencer has a newer version of
> > this code available but I was unable to find it.
> 
> Recent versions of Tcl include a newer version of his code:
> 
> /*
>  * re_*exec and friends - match REs
>  *
>  * Copyright (c) 1998, 1999 Henry Spencer.  All rights reserved.
>  * 
>  * Development of this software was funded, in part, by Cray Research Inc.,
>  * UUNET Communications Services Inc., Sun Microsystems Inc., and Scriptics
>  * Corporation, none of whom are responsible for the results.  The author
>  * thanks all of them. 
>  * 
>  * Redistribution and use in source and binary forms -- with or without
>  * modification -- are permitted for any purpose, provided that
>  * redistributions in source form retain this entire copyright notice and
>  * indicate the origin and nature of any modifications.
>  * 
>  * I'd appreciate being given credit for this package in the documentation
>  * of software which uses it, but that is not a requirement.
>  * 
>  * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
>  * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY
>  * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL
>  * HENRY SPENCER BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
>  * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
>  * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
>  * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
>  * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
>  * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
>  * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
>  *
>  */

Jeff Law wrote:
> I don't think we want to fight that kind of battle right now.

OK.  When?  ;-)  Are there issues with the above copyright?

> What I would suggest is that if the GNU regex routines are that lame is
> for someone to get in touch with whomever maintains them and get them
> improved.

OK.  Who?  Henry?  What constraints are there on the fixincl code
that are not covered by the above copyright?  What _is_ Henry's
eaddress anyway? ...  :-}

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Andi Kleen]

On Thu, Dec 30, 1999 at 02:10:23PM +0100, Martin v. Loewis wrote:
> > Are you sure that the regexp code is copyrighted by the UCB, not
> > Henry Spencer? If the later only Henry could change the license,
> > Berkeley cannot do it by fiat.
> 
> According to
> http://www.FreeBSD.org/cgi/cvsweb.cgi/src/include/regex.h?rev=1.3 ,
> it is
> 
>  * Copyright (c) 1992 Henry Spencer.
>  * Copyright (c) 1992, 1993
>  *      The Regents of the University of California.  All rights reserved.
>  *
>  * This code is derived from software contributed to Berkeley by
>  * Henry Spencer of the University of Toronto.
> 
> So I guess UCB can (and did) change the license on that code. Henry
> might have other versions of that code that he did not contribute to
> Berkeley.

All copyright owners have to agree to a license change AFAIK.

-Andi

-- 
This is like TV. I don't like TV.

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Martin v. Loewis]

> Are you sure that the regexp code is copyrighted by the UCB, not
> Henry Spencer? If the later only Henry could change the license,
> Berkeley cannot do it by fiat.

According to
http://www.FreeBSD.org/cgi/cvsweb.cgi/src/include/regex.h?rev=1.3 ,
it is

 * Copyright (c) 1992 Henry Spencer.
 * Copyright (c) 1992, 1993
 *      The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * Henry Spencer of the University of Toronto.

So I guess UCB can (and did) change the license on that code. Henry
might have other versions of that code that he did not contribute to
Berkeley.

Regards,
Martin

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Andi Kleen]

martin@loewis.home.cs.tu-berlin.de (Martin v. Loewis) writes:

> > The issue is whether or not the Berkeley regex processing is okay
> > for inclusion or not.
> 
> The good news is that earlier objections about incompatibilities
> between GPL and the Berkeley license don't apply anymore, see
> 
> ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change
> 
> GPL requires that no additional restrictions may be made, and the only
> additional restriction in BSD code has been revoked. I don't know for
> sure whether the regex code falls under this rule, but I guess it
> will.

Are you sure that the regexp code is copyrighted by the UCB, not
Henry Spencer? If the later only Henry could change the license,
Berkeley cannot do it by fiat.

-Andi

-- 
This is like TV. I don't like TV.

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[Martin v. Loewis]

> The issue is whether or not the Berkeley regex processing is okay
> for inclusion or not.

The good news is that earlier objections about incompatibilities
between GPL and the Berkeley license don't apply anymore, see

ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change

GPL requires that no additional restrictions may be made, and the only
additional restriction in BSD code has been revoked. I don't know for
sure whether the regex code falls under this rule, but I guess it
will.

Of course, you cannot expect the authors of this code assign their
copyright to the FSF, which might be a problem for inclusion into GCC.
The same problem would occur with any other general-purpose regex
engine that is not already copyright by the FSF.

Regards,
Martin

Re: Berkeley Regex - was: fixincludes jumbo patch 2

[John Wehle]

> Zack Weinberg submitted a fixincl patch for many things.
> I have asked him under private email to break up and
> simplify the patch.  I am pulling out one issue here
> in order to get consensus before installing.  The issue
> is whether or not the Berkeley regex processing is okay
> for inclusion or not.  It will apparently cut processing
> time by another 60% yielding a 7-fold improvement over the
> original script.  I don't get particularly excited over the
> Berkeley vs. GNU license debate, but I would hate to
> replace the gnu-regex only to have to put it back.

Another option is to use GNU rx which appears to be a more
current, faster, and correct regular expression library
than GNU regex.  I don't know how it compares to the Berkeley
implementation.

-- John
-------------------------------------------------------------------------
|   Feith Systems  |   Voice: 1-215-646-8000  |  Email: john@feith.com  |
|    John Wehle    |     Fax: 1-215-540-5495  |                         |
-------------------------------------------------------------------------

Berkeley Regex - was: fixincludes jumbo patch 2

[korbb]

This probably belongs on the GCC list more than gcc-patches:)


Zack Weinberg submitted a fixincl patch for many things.
I have asked him under private email to break up and
simplify the patch.  I am pulling out one issue here
in order to get consensus before installing.  The issue
is whether or not the Berkeley regex processing is okay
for inclusion or not.  It will apparently cut processing
time by another 60% yielding a 7-fold improvement over the
original script.  I don't get particularly excited over the
Berkeley vs. GNU license debate, but I would hate to
replace the gnu-regex only to have to put it back.

Comments, please?

Thanks,
	Bruce