proposal to optimise the performance of the Jar utility

[Alan Bateman]

mike.skells at talk21.com wrote:
> Hi,
> Not sure if this is too late for Java 7 but I have made some optimisations for a 
> client to improve the performance of the jar utility in their environment, and 
> would like to promite then into the main code base
>   
It's been a long standing issue that creating JAR files with the jar 
command has been very slow. Sherman did some good improvements in jdk7 
[1], and I think they've been back-ported into Oracle jdk6 releases too. 
Also Martin Buchholz and Jeremy Mason fixed another performance issue 
for cases where the -C option is used to specify many directories [2]. 
Combined, this leads to some good improvements compared to older 
releases of the JDK. Sherman has a blog entry [3] on this topic that 
details some of the results.

I've no doubt more can be done and it sounds like you've made it your 
mission to get it as fast as possible (which is great). I would suggest 
sending a webrev or a patch so that the changes can be discussed. Also 
it would be good to include some performance figures comparing the 
implementation against jdk7. Just to set expectations, it probably is a 
bit late in jdk7 for extensive or risky changes but we probably need to 
see extend of the changes to judge that. If there are a lot of changes 
then one idea is to push it into jdk8 first and let it bake there for a 
while.

-Alan

[1] http://hg.openjdk.java.net/jdk7/jdk7/jdk/rev/57dc40ece164
[2] http://hg.openjdk.java.net/jdk7/jdk7/jdk/rev/ce55eb6668d9
[3] http://blogs.sun.com/CoreJavaTechTips/entry/superduper_slow_jar