RFR: 8224202: Speed up Properties.load

Claes Redestad claes.redestad at oracle.com
Thu May 23 12:57:21 UTC 2019


Hi Ivan!

On 2019-05-23 02:11, Ivan Gerasimov wrote:
> Hi Claes!
> 
> A few comments:
> 
> 1)
> 
>   571                     if (len == lineBuf.length) {
>   572                         int newLength = 
> ArraysSupport.newLength(lineBuf.length, 1, lineBuf.length);
> 
> could be written slightly shorter as
> 
>   571                     if (len == lineBuf.length) {
>   572                         int newLength = 
> ArraysSupport.newLength(len, 1, len);
> 
> 2)
>   585                     if (len == 0) {
>   586                         skipWhiteSpace = true;
>   587                         len = 0;
> 
> no need to update len here

nice catches, fixed.

> 
> 3)
> In loadConvert(char[] in, int off, int len, StringBuilder out):
> 
> It may make sense to first scan the input buffer for backslash, and if 
> not found (should be a common case), return new String(in, off, len), so 
> a copying to the StringBuilder will be avoided.
> 
> Otherwise:
>      out.append(in, off, posOfSlash - off);
>      off = posOfSlash;
> 
> and continue as before...

Right, this got a bit murky since String(char[], int, int) runs through
a noisy compress method and execute more bytecode in profiles. Still it
comes out a bit faster in a few real measurements, probably due to doing
fewer calls from interpreter to compiled code in the end:

http://cr.openjdk.java.net/~redestad/8224202/open.03/

Thanks!

/Claes


More information about the core-libs-dev mailing list