I think it's been committed to HEAD (and I know it's on my branch). Tom said it had a small but noticeable effect on Rails perf. Bit by bit, we're getting there.On 6/6/06,
Nick Sieger <[EMAIL PROTECTED]> wrote:
Looks good to me.Interesting what a useful abstraction "CharSequence" is, and how long
I've kicked the idea around, but the truth is that Java's String is pretty blazing fast. Sticking with a normal String also leaves open the option to support unicode out-of-the-box, which we wouldn't get with a byte[]. There's a bit of a mismatch between how Ruby strings work and how Java's String
this.
Anyway, rambling... =)
/O
- Original Message -
From: Nick Sieger <[EMAIL PROTECTED]>
Date: Wednesday, June 7, 2006 11:17 pm
Subject: Re: [Jruby-devel] More performance numbers and a committable patch
To: [email protected]
> Looks good to me.
>
> Inter
Looks good to me.Interesting what a useful abstraction "CharSequence" is, and how long it took to find its way into the JDK.On 6/4/06, Charles O Nutter
<[EMAIL PROTECTED]> wrote:
Blast, forgot the patch.
___
Jruby-devel mailing list
[email protected]
Have there been any thoughts of implementing RubyString as byte[]
instead of String? Or maybe java.nio.ByteBuffer?
This fits better with the Ruby paradigm of what a String *is*, saves
memory (since RubyStrings are binary data, there is one wasted byte for
every byte of real data in a Java string)
ObjectSpace is slow.I'm optimizing RubyString and seeing again and again how slow OS makes object creation. I've attached a patch that doesn't break anything and optimizes a few aspects of RubyString. Without OS enabled, this patch gives 15-20% improvement in the given microbenchmark. It's not quit
Blast, forgot the patch.On 6/4/06, Charles O Nutter <[EMAIL PROTECTED]> wrote:
ObjectSpace is slow.I'm optimizing RubyString and seeing again and again how slow OS makes object creation. I've attached a patch that doesn't break anything and optimizes a few aspects of RubyString. Without OS enabled,
