While cleaning up my temp directory I noticed files named
apache-tika-fork-*.jar
The jar file is created in ForkClient.createBootstrapJar().
It is used by TikaCLITest.
One could fix this by making a change in TikaCLI.OutputType.process(),
by adding
if (fork) {
((ForkParser) p).close();
}
at the end.
I don't know enough about the code to be 100% sure that the jar file
isn't used at a later time so I didn't make the change myself.
Btw I also tried to run the fork parser from the command line and only
succeeded with small PDF files, I got an OutOfMemoryError, even with
-Xmx16g when trying bigger PDF files. The ForkClient doesn't seem to set
any -Xmx value for the fork.
Tilman
