From: Ilya Zakharevich <nospam-abuse@ilyaz.org>
To: help-gnu-emacs@gnu.org
Subject: Re: will we ever have zero width assertions in regexps?
Date: Sat, 29 Jan 2011 22:28:48 +0000 (UTC) [thread overview]
Message-ID: <slrnik9550.1ae.nospam-abuse@powdermilk.math.berkeley.edu> (raw)
In-Reply-To: jwvzkqke8n4.fsf-monnier+gnu.emacs.help@gnu.org
On 2011-01-29, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
>> As I said, I put some optimizations which in most (AFAIK) practical
>> senses remove such pathologies. (The underlying problems remain; the
>> optimizations are only "heuristic"; but one needs to be extra
>> inventive to circumvent the optimizations.)
>
> A typical case could look something like "foo *(.*?) *bar". when
> matching "foo ..<many space>.. baZ".
No, this is a polynomial-time problem. My optimization does nothing
for such cases. And I do not think such a REx would provide any
problem in real life - unless you have many hundreds of consecutive
spaces. (And unless Emacs' REx engine is particularly slow per OPCODE.)
But I start to see the difference - it is in usage scenarios. Many
Perl REx matches are done "per-line", not "per-file". And in Emacs,
one would rarely narrow the buffer before a match. This creates a
major skew in REx matches - in Emacs the string to match against would
be quite often a couple of orders of magnitude larger.
Essentially, in Perl, a sloppy REx which leads to a non-linear
polynomial time matching would be mostly unnoticed speed-wise. Only
those which lead to exponential-time match bite hard.
So my patches for Perl improve only the exponential cases of
"sloppyness". In Emacs, polynomial-time may give quite noticable
problems too. Hmm, AFAICS, one can slighly modify my patches to
handle polynomial-time matches too. I would think about it more (but
do not promise any action! ;-).
> Basically, provide a primitive like (match-string RE STRING LIMIT) that
> can not only say "matched between START and END", but also "reached
> LIMIT within yet finding a match, here's the suspended SEARCH-STATE at
> LIMIT", so you can later resume the search starting at LIMIT by passing
> that state.
match-with-continuation. An interesting idea. I already implemented
it for Perl (to support (??{}), but it is not exposed to the user.
Would one want this in non-interactive situations?
Thanks,
Ilya
next prev parent reply other threads:[~2011-01-29 22:28 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <mailman.1.1296054361.23496.help-gnu-emacs@gnu.org>
2011-01-26 15:58 ` will we ever have zero width assertions in regexps? Stefan Monnier
2011-01-27 1:45 ` Le Wang
[not found] ` <mailman.6.1296092730.6982.help-gnu-emacs@gnu.org>
2011-01-27 2:21 ` Stefan Monnier
2011-01-27 6:34 ` Ilya Zakharevich
2011-01-27 16:10 ` Stefan Monnier
2011-01-28 23:49 ` Ilya Zakharevich
2011-01-29 2:51 ` Stefan Monnier
2011-01-29 22:28 ` Ilya Zakharevich [this message]
2011-01-31 16:08 ` Stefan Monnier
2011-01-31 17:10 ` Ilya Zakharevich
2011-01-31 21:29 ` Stefan Monnier
2011-02-02 15:09 ` Ilya Zakharevich
2011-02-07 20:30 ` Stefan Monnier
2011-02-08 22:41 ` Ilya Zakharevich
2011-01-26 14:55 Le Wang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: https://www.gnu.org/software/emacs/
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=slrnik9550.1ae.nospam-abuse@powdermilk.math.berkeley.edu \
--to=nospam-abuse@ilyaz.org \
--cc=help-gnu-emacs@gnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).