RFR (S): 7185591: jcmd-big-script.sh ERROR: could not find app's Java pid.

[Staffan Larsen]

Katja,

test/lib/testlibrary/jdk/testlibrary/JcmdBase.java
 68      * Run jcmd standalone

I think you should expand a bit on what “standalone” means here. It took me a while to understand the difference.

test/lib/testlibrary/jdk/testlibrary/OutputAnalyzer.java
423     public int indexOf(String pattern) {

This seems very inefficient. Add all lines to an ArrayList and then walk through them one at a time to if it matches and then walk through them once again to find the index of that line.

469     public int shouldMatchByLine(String from, String to, String pattern) {

Same inefficiency here, but worse because both asLines() and indexOf() does the same work.

test/lib/testlibrary/jdk/testlibrary/Utils.java
65     public static final String TEST_SRC = System.getProperty("test.src").trim();

I wonder if this really works. Isn’t “test.src” different for different tests? A property that jtreg changes before invoking each test? Or does this work because each test is run in a different class loader and Utils.java will exist once in each class loader?


/Staffan 


On 10 jan 2014, at 13:50, Yekaterina Kantserova <yekaterina.kantserova at oracle.com> wrote:

> Hi,
> 
> Could I please have a review of this fix.
> 
> In this fix I've rewritten sun/tools/jcmd/* tests in pure java to get rid of all intermittent failures depending on for example MKS or race conditions (test application has not yet started when the test start to run).
> 
> 
> Webrev:
> http://cr.openjdk.java.net/~ykantser/7185591/webrev.00/
> 
> Primal bug:
> https://bugs.openjdk.java.net/browse/JDK-7185591
> 
> Similar bugs:
> https://bugs.openjdk.java.net/browse/JDK-6977426
> https://bugs.openjdk.java.net/browse/JDK-8020798
> https://bugs.openjdk.java.net/browse/JDK-7130656 (this one is blocked by https://bugs.openjdk.java.net/browse/JDK-8031482 so far)
> 
> Thanks,
> Katja