an unmotivated IllegalArgumentException
Bryce McKinlay
bryce@waitaki.otago.ac.nz
Thu Jun 6 19:24:00 GMT 2002
Roland.Hautz wrote:
>please help me with the interpretation of the appended gdb session. There is a
>crash in java.util.Hashtable.Hashtable(int, float), and I can't think of any
>reason for that (besides of a gcj bug in general ;-)) .
>
This looks like a miscompilation of some kind - as if the "if (!
(loadFactor > 0))" was incorrectly constant-folded to "true" - (there
was a bug like this on the mainline that was fixed recently). However it
would be very strange if this did not show up with a simple test case
that called the Hashtable constructor.
I would say that gdb is playing tricks on you if it wern't for the
actual exception being thrown. Perhaps try again with the latest gcc
tree to see if it still occurs. You can also try building libgcj with
GCJFLAGS=-g to get a non-optimized build.
>xalan is compiled without -O; I assume this implies -no-inline?
>
Right. In any case, code from libgcj should never get inlined into your
application binary.
regards
Bryce.
More information about the Java
mailing list