[kaffe] regression test stalled

Patrick Tullmann tullmann at cs.utah.edu
Fri Jan 31 14:43:01 PST 2003


> I downloaded the newest version using CVS. Then I built, installed, and
> tested it using the following command:
> make
> make install
> make check

Looks good.

> The test didn't progress for more than 2 hours after "PASS:
> SoTimeout.java" was printed. I stopped it using CTRL_C. Is it normal
> to wait for so long time or there is something wrong?

It shouldn't take that long on any one test, so it was probably hung.
Since SoTimeout passed, it was probably hung on the next test:
wc.java.  The only really interesting bit of that test is the stdin
redirection, so perhaps that was a problem?  You can try running the
test manually (i.e., "java wc < wc.java", it expects something
interesting on stdin to count lines/words/chars of).

> Moreover, the test failed for some programs: DoubleCvt.java,
> ThreadState.java, ProcessTest.java

These failures might be benign, the way to tell is to look at the
"*.fail" and compare with the corresponding "*.out" file.  ".fail" is
the actual output, and .out is the expected output.  You can just send
the .fail files to the list (assuming they're not too big).

Include the output of 'kaffe -fullversion', too.

> The OS is SunOS 5.8, and I am not a root user.

I'm actually a bit surprised it ran at all!  :)  I don't recall if
anyone has done much work on SunOS support recently...


----- ----- ---- ---  ---  --   -    -      -         -               -
Pat Tullmann                                       tullmann at cs.utah.edu
    Not many people realize just how well known I am -- Anonymous

More information about the kaffe mailing list