You are viewing a plain text version of this content. The canonical link for it is here.
Posted to oro-user@jakarta.apache.org by "Matthias L. Jugel" <ma...@first.fhg.de> on 2003/03/24 14:17:41 UTC

Performance problem with regex

Hello,

up to now oro has worked great but we are now suffering a strange 
performance problem with the following regular expression:

"\\{([^:}]+):?(.*?)\\}(.*?)\\{(\\1)\\}" (Java String)

It is used to match macros in our wiki markup in the form of

{quote}
The quick brown fox ...
{quote}

We now found that this takes about 30s when we try to match a text with 
inlined macros like the above with about 28-30k

I have verified that that it is the matching that takes so long by 
creating a small test program that employs the oro matcher and javas 
java.util.regex. It looks like the java package is marginal faster but 
it still takes too long.

If this is because the regular expression is stupid or some other 
reason I cannot say and would be very happy to get help. I have 
appended the comparison test and a sample text.

Please completely ignore the contents of the example. It is not our 
text and we would like to stay neutral in this case. The text is for 
the sole purpose of the test only. So, no discussions please, there are 
other places.

Thank you,

	Leo.