Since you seem to have Java regexp knowledge: is there a way to get ORO to work on a byte[] more efficient than converting it into a String first?
-- Salut,
Jordi.
En/na Stefan Bodewig ha escrit:
On Mon, 24 Nov 2003, Jordi Salvat i. Alabart <[EMAIL PROTECTED]> wrote:
I've found java.util.regex to be much simpler and slightly better performing than the ORO classes for the kind of usage we need.
As a complete outsider, but one with a bit of Java regexp knowledge as I've coded larger parts of Ant's regexp facade that abstracts away the choice: I'd always prefer ORO.
We've found java.util.regexp to have some strange platform dependent bugs that are hard if not impossible to work around, from <replaceregexp>'s manual:
,----
| There are cross-platform issues for matches related to line
| terminator. For example if you use $ to anchor your regular expression
| on the end of a line the results might be very different depending on
| both your platform and the regular expression library you use. It is
| 'highly recommended' that you test your pattern on both Unix and
| Windows platforms before you rely on it.
| | * Jakarta Oro defines a line terminator as '\n' and is consistent
| with Perl.
| | * Jakarta RegExp uses a system-dependant line terminator.
| | * JDK 1.4 uses '\n', '\r\n', '\u0085', '\u2028', '\u2029' as a
| default but is configured in the wrapper to use only '\n'
| (UNIX_LINE)
| | We strongly recommend that you use Jakarta Oro.
`----
In many ways your choice depends on how much you want to expose to your users and what background your users are likely to have. If your users come from a Perl background, ORO may in fact be the better choice.
Stefan
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
