[kaffe] Regression Test Failures on m68k Amiga

Archie Cobbs archie at dellroad.org
Thu Feb 27 21:14:01 PST 2003


Tony Wyatt wrote:
> Of the 129 tests in /test/regression, 112 run successfully on my Amiga. Of
> those that don't, several fail because they print their output lines in the wrong
> order, compared to the "expected output". 
> 
> My question is, is this really a failure? Does it represent some disastrous
> failure-to-schedule-threads-in-the-right-order?
> Or is it no problem at all? 
> 
> A typical example is /test/regression/CatchDeath.java. The expected output
> is:
> CD Caught java.lang.ThreadDeath
> CSD Caught java.lang.ThreadDeath
> CSWD Caught java.lang.ThreadDeath
> 
> What it gets is
> CD Caught java.lang.ThreadDeath
> CSWD Caught java.lang.ThreadDeath
> CSD Caught java.lang.ThreadDeath
> 
> - which is counted as a failure.

Tests for which the order of the output is indeteriminate should
have the "// Sort output" comment.

I think in this case test just needs a "// Sort output" line.

-Archie

__________________________________________________________________________
Archie Cobbs     *     Precision I/O      *     http://www.precisionio.com




More information about the kaffe mailing list