Hello Thomas,

It was foretold that on Sunday, November 2, 2003 at 22:37 GMT +0100,
Thomas Woelk [TW] would type:

TW>>> SB! is freezing and say good bye. Maybe it is a SecureBat!
TW>>> Problem... or whatever. :-(

>> Can you please post *exactly* what you have in your DelLS2 template?
<snip template>

Hmm, I tried this and I'm getting it to remove the leading space from
the first line only.  So good news, it shouldn't be crashing.  Bad
news, it still doesn't work.

TW> I used the DelLS with my Quotes-Cleaner "Quotes_DELLS" and it
TW> works fine with SB! V1.6.

I have a sneaking suspicion that there is a bug in the regexp
implementation in v2.  I've noticed some weird greedy/non-greedy
interactions*, and I suspect the same thing is happening here.  I'm
very confused why SB and TB would manifest such a bug in different
ways.

I'll try re-writing the template for V2 at some point.  I just have to
find time to squeeze it into my schedule.

* For anyone who is interested, the bug seems to occur when you have a
  greedy modifier between two ungreedy modifiers.  If the strings
  being matched can overlap, then the last modifier wins, not the
  greedy modifier.  For example, suppose I have the string:
  xxxyyyzzz123123123xxxyyyzzz
  And the regexp: ^(.*)(\d*?)(.*)$
  Or: ^(.*?)(\d*)(.*?)$

  In the first case, the entire string is where it should be, in
  subpattern 1.  In the second case, logic suggests that the
  *preferred* match is with xxxyyyzzz in subpatterns 1 and 3 and
  123123123 in subpattern 2.  But what you get if you run it is that
  the entire string ends up in subpattern 3.  This isn't technically
  wrong, but it isn't the optimum match.

-- 
Thanks for writing,
 Januk Aggarwal



________________________________________________________

http://www.silverstones.com/thebat/TBUDLInfo.html

Reply via email to