Re: [PATCH] for Review: BufferedReader and InputStreamReader Optimizations (Updated)

[Bryce McKinlay <bryce at mckinlay dot net dot nz>]

Mohan,

Thanks for persuing this - the patch looks good to me. I presume you've 
run it through Mauve? Assuming it passes the tests, please go ahead and 
check it in to mainline.

Regards

Bryce.


On Jan 30, 2004, at 4:55 PM, Mohan Embar wrote:

> Knock knock. It's me again, with the >50% improvement for
> BufferedReader.readline():

> ...my recommendation would be to check my version in,
> enjoy the immediate performance gain, then let Bryce, Dalibor
> or whoever else tear this to pieces if he is so inclined.

> ChangeLog
> 2004-01-29  Mohan Embar  <gnustuff@thisiscool.com>
>
> 	* java/io/BufferedReader.java (sbuf): New field.
> 	(readLine): Use String.valueOf instead of new String() as per
> 	Per Bothner's suggestion. Use instance sbuf field instead of a
> 	local StringBuffer instance.
> 	* java/io/InputStreamReader.java (read(char[],int,int)): Pass the
> 	caller's buffer to refill().
> 	(read(void)): Pass our internal work buffer to refill if our
> 	input queue is empty.
> 	(refill): Changed return type to int. Use the specified buffer
> 	instead of our work buffer as per Bryce McKinlay's suggestion.
> 	Return the number of characters read or -1 for EOF.