Hi,
Am Donnerstag, den 18.09.2008, 09:42 +0100 schrieb Philip Hazel:
> This same question was just discussed on the exim-dev mailing list.
Yes, I just learned that the bug reporter for Octave has asked this here
as well. Sorry, I only looked at the pcre-dev list's archive.
> The OP then said that that it worked with matlab, so I tried it with
> Perl, and Perl was OK, so I then made this comment:
Yes, I've tried it with Matlab as well. It works, but takes more than
three seconds on my not-too-slow machine.
> The OP's comment about \Q is not right; however, I can get the regex to
> work by adding \q13000000 on the front of the data, though it is very
> slow. This sets MATCH_LIMIT (the default is 10000000). I discovered what
> number to use by adding \M on the front. This (as well as the fact that
> possessifying the \d+ makes it work) confirms the original diagnosis.
So, would a feasible solution look like the following:
1) Try with the default values on MATCH_LIMIT.
2) If that fails, check with (the equivalent of) \M and use that value
for a second calculation?
Thanks
Thomas
