Re: Unnecessary array allocation and copying

[Sylvain Wallez]

Ugo Cei wrote:

While browsing our sources, I came across this snippet from 
o.a.c..components.pipeline.AbstractProcessingPipeline:

// execute the pipeline:
this.generator.generate();
byte[] data = os.toByteArray();
environment.setContentLength(data.length);
environment.getOutputStream(0).write(data);
But from the javadocs of java.io.ByteArrayOutputStream#toByteArray, I 
read:

"Creates a newly allocated byte array. Its size is the current size of 
this output stream and the valid contents of the buffer have been 
copied into it."

Thus I wonder: why are we doing this copy instead of doing:

// execute the pipeline:
this.generator.generate();
environment.setContentLength(os.size());
os.writeTo(environment.getOutputStream(0));
If I'm not mistaken, this would avoid allocating and copying an array 
as large as the serializer's output. Or am I missing something subtle 
here?


I guess you just found a more efficient way to achieve exactly the same 
thing ;-)

It seems to me we can safely avoid this array allocation.

Sylvain

--
Sylvain Wallez  Anyware Technologies
http://www.apache.org/~sylvain   http://www.anyware-tech.com
{ XML, Java, Cocoon, OpenSource }*{ Training, Consulting, Projects }

Re: Unnecessary array allocation and copying

[Ugo Cei]

Il giorno 26/apr/04, alle 23:19, Bruno Dumon ha scritto:

The same pattern occurs at other locations also, e.g.
AbstractCachingProcessingPipeline line 246
Yes, there are four occurrences between AbstractProcessingPipeline and 
AbstractCachingProcessingPipeline. I've optimized all four locally and 
tested by running random samples. So far, so good. Unfortunately, our 
test suite does not compile ATM (gr), so I'll probably commit and 
see if someone complains.

	Ugo

Re: Unnecessary array allocation and copying

[Bruno Dumon]

On Mon, 2004-04-26 at 22:39, Ugo Cei wrote:
> While browsing our sources, I came across this snippet from 
> o.a.c..components.pipeline.AbstractProcessingPipeline:
> 
>  // execute the pipeline:
>  this.generator.generate();
>  byte[] data = os.toByteArray();
>  environment.setContentLength(data.length);
>  environment.getOutputStream(0).write(data);
> 
> But from the javadocs of java.io.ByteArrayOutputStream#toByteArray, I 
> read:
> 
> "Creates a newly allocated byte array. Its size is the current size of 
> this output stream and the valid contents of the buffer have been 
> copied into it."
> 
> Thus I wonder: why are we doing this copy instead of doing:
> 
>  // execute the pipeline:
>  this.generator.generate();
>  environment.setContentLength(os.size());
>  os.writeTo(environment.getOutputStream(0));
> 
> If I'm not mistaken, this would avoid allocating and copying an array 
> as large as the serializer's output. Or am I missing something subtle 
> here?

I don't think so, this seems like a good optimalisation.

The same pattern occurs at other locations also, e.g.
AbstractCachingProcessingPipeline line 246

-- 
Bruno Dumon http://outerthought.org/
Outerthought - Open Source, Java & XML Competence Support Center
[EMAIL PROTECTED]  [EMAIL PROTECTED]

Unnecessary array allocation and copying

[Ugo Cei]

While browsing our sources, I came across this snippet from 
o.a.c..components.pipeline.AbstractProcessingPipeline:

// execute the pipeline:
this.generator.generate();
byte[] data = os.toByteArray();
environment.setContentLength(data.length);
environment.getOutputStream(0).write(data);
But from the javadocs of java.io.ByteArrayOutputStream#toByteArray, I 
read:

"Creates a newly allocated byte array. Its size is the current size of 
this output stream and the valid contents of the buffer have been 
copied into it."

Thus I wonder: why are we doing this copy instead of doing:

// execute the pipeline:
this.generator.generate();
environment.setContentLength(os.size());
os.writeTo(environment.getOutputStream(0));
If I'm not mistaken, this would avoid allocating and copying an array 
as large as the serializer's output. Or am I missing something subtle 
here?

	Ugo