[bug #14747] major (10X) memory savings possible in some situations

2005-12-20 15:05:33

Update of bug #14747 (project mhonarc):

                  Status:                    None => In Progress            


Follow-up Comment #1:

Andrew creates a custom regex for the given charset and
passes the data through that versus the existing generic
regex.  This definitely appears to be more efficient in
how Perl's regex engine works (and simplifies the substitute
operation, which should be faster).

To generalize, a cache of charsets->regexes could be
created to avoid regenerating the custom regex each time
(improving performance even further).

Alternatively, I could auto-create perl routines before-hand
that define the more efficient regexes.  These routines are
required on-demand when needed.  I.e.  The process of creating
the regexes are done at mhonarc build-time instead of at
run-time since there are no run-time dependencies on the
creation of the regexes.

BTW, I do not think all charsets can
be handled in this manner, but for those that do, this
performance enhancement can be done.

Note, those that use TEXTENCODE should not encounter this
performance problem since char->html conversion is vastly
simplified and should avoid the routine(s) in question here.
This assertion is made on my understanding of the code versus
any actual testing.  Those using TEXTENCODE are free to
run tests and report their results back to the dev list.


Reply to this item at:


  Message sent via/by Savannah

To sign-off this list, send email to majordomo(_at_)mhonarc(_dot_)org with the

<Prev in Thread] Current Thread [Next in Thread>
  • [bug #14747] major (10X) memory savings possible in some situations, Earl Hood <=