This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: ZipInputStream brokenness
- From: jmr at ugcs dot caltech dot edu (Jesse Rosenstock)
- To: tromey at redhat dot com
- Cc: java at gcc dot gnu dot org
- Date: Sat, 31 Aug 2002 01:51:47 -0700
- Subject: Re: ZipInputStream brokenness
Jesse> However, when I compile with -g, I get a different error:
Jesse> ; gcj -g ZISTest.java --main=ZISTest -o zistest
Tom> I get the same error with -g. That's what I would expect, since -g
Tom> shouldn't affect the generated code at all (this is a general gcc
Tom> rule, though of course there could be bugs). Could you try to find
Tom> out why this makes a difference for you?
After a bit more experimentation, I see that it's intermittent, and
happens occasionally regardless of how it's run. Run it about 20 times
and see if you can reproduce it.
Some highly unscientific testing revealed the following failure rates:
gij: 16%
gcj: 53%
gcj -g: 35%
Sample size about 100 in each case.