Hi,
if you have thousands of test cases and want per test coverage results
you need thousands analysis calls. Maybe you can optimize a bit and e.g.
narrow the scope of the classes analyzed for a specific test.
Anyway, your code looks a bit strange: Why do you put the analysis into
the visitSessionInfo() method? This forces you to parse the cov byte
Array twice. This should be enough:
ExecFileLoader execFileLoader = new ExecFileLoader();
execFileLoader.load(new ByteArrayInputStream(cov));
final CoverageBuilder coverageBuilder = new CoverageBuilder();
final Analyzer analyzer = new
Analyzer(execFileLoader.getExecutionDataStore(), coverageBuilder);
analyzer.analyzeAll(new File("__classes_directory__"));
for (final IClassCoverage cc : coverageBuilder.getClasses()) {
// perfom my coverage analysis
}
If you need Session infos you can retrieve them from the ExecFileLoader:
execFileLoader.getSessionInfoStore();
Regards,
-marc
On 2015-10-15 13:44, José Carlos de Campos wrote:
Hi,
First of all, many thanks for your amazing work on JaCoCo. :)
I'm trying to build a tool on the top of JaCoCo, and one of the
features
requested is coverage per *unit* test case (and not the aggregated
coverage).
What I've so far: I've created my own TestListener, and every time an
*atomic* test finishes I execute
byte[] b = RT.getAgent().getExecutionData(true);
and then (as suggested here [1]) I create a _dump_ function that
performs
the coverage analysis. something like:
final ByteArrayInputStream in = new ByteArrayInputStream(cov); // cov
is the byte array from agent.dump()
final ExecutionDataReader reader = new ExecutionDataReader(in);
reader.setSessionInfoVisitor(new ISessionInfoVisitor() {
public void visitSessionInfo(final SessionInfo info) {
ExecFileLoader execFileLoader = new ExecFileLoader();
execFileLoader.load(new ByteArrayInputStream(cov));
final CoverageBuilder coverageBuilder = new CoverageBuilder();
final Analyzer analyzer = new
Analyzer(execFileLoader.getExecutionDataStore(), coverageBuilder);
analyzer.analyzeAll(new File("__classes_directory__"));
for (final IClassCoverage cc : coverageBuilder.getClasses()) {
// perfom my coverage analysis
}
}
});
reader.setExecutionDataVisitor(new IExecutionDataVisitor() {
public void visitClassExecution(final ExecutionData data) {
// empty
}
});
however, and because I'm calling my _dump_ function thousands of
times,
my coverage analysis is taking ages to finish. I've measured the CPU
time
and, 45% is
org.jacoco.core.internal.flow.MethodProbesAdapter.visitLabel(),
13% is org.jacoco.core.internal.analysis.SourceNodeImpl.incrementLine.
is there any better way of doing this? am I not using the right API
functions?
in your opinion, how can I improve the performance of my coverage
analysis?
--
Cheers,
Jose
--
You received this message because you are subscribed to the Google
Groups "JaCoCo and EclEmma Users" group.
To unsubscribe from this group and stop receiving emails from it, send
an email to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/jacoco/762ac31f-641b-46ed-9d66-61956778e2b8%40googlegroups.com
[2].
For more options, visit https://groups.google.com/d/optout [3].
Links:
------
[1] http://eclemma.org/jacoco/trunk/doc/examples/java/ExecDump.java
[2]
https://groups.google.com/d/msgid/jacoco/762ac31f-641b-46ed-9d66-61956778e2b8%40googlegroups.com?utm_medium=email&utm_source=footer
[3] https://groups.google.com/d/optout
--
You received this message because you are subscribed to the Google Groups "JaCoCo
and EclEmma Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/jacoco/73b0c2acd189320b1928df14856ee5c6%40mountainminds.com.
For more options, visit https://groups.google.com/d/optout.
