[Mono-devel-list] Nant / OJB.NET / Maverick.NET / NVelocity

Todd Berman tberman at gentoo.org
Tue Oct 7 13:02:59 EDT 2003


> -----Original Message-----
> From: mono-devel-list-admin at lists.ximian.com [mailto:mono-devel-list-
> admin at lists.ximian.com] On Behalf Of Paolo Molaro
> Sent: October 7, 2003 10:45 AM
> To: 'Mono-devel'
> Subject: Re: [Mono-devel-list] Nant / OJB.NET / Maverick.NET /
NVelocity
> 
> On 10/06/03 Todd Berman wrote:
> > Lupus, the issue is the entire Regex engine of mono, a known, and
> > documented mono issue. Filing a mono bug doesn't seem to make sense.
> 
> So, if someone has taken the time to actually find that the 'entire'
> Regex engine is at fault (that means reviewing 5000+ lines of code,
> writing tests and benchmarks), why hasn't he taken an additional 30
> seconds to file a bugzilla report? As is almost always the case,
> probably the nant issue depends mainly on few separated issues.

I am not sure who took said time to figure it out, when I decided to go
with NAnt for building purposes, and spoke to some people on irc about
its seeming slowness, I was told it was the Regex engine. I assumed
since it was known, bugs were filed, if that isn't the case, I will
happily file a bug.

My main point was not that the regex engine of mono completely sucks,
and that all 5000 lines of code need to be re-written, and more that it
needs the attention of someone who knows optimizations, and can tweak it
here and there for better performance.

Now, testing your patch, I ran my NAnt build 4 times, once from the
clean target (full rebuild required) with and without the patch, and
once from a fully built target (no rebuilding required atall) with and
without the patch. Here are my `time` based results:



More information about the Mono-devel-list mailing list