Re: 6.0 NEWS; tls tests on gdb-6?

[David Carlton <carlton@kealia.com>]

On Sat, 20 Sep 2003 16:12:19 -0400, Andrew Cagney <ac131313@redhat.com> said:

>> A long-standing bug involving Java and dynamic libraries has been
>> fixed.  Support for Java debugging remains minimal, however.

> Do you have something more specific?  That could mean anything.

Yeah, you're right: that is vague.  The long-standing bug is that it
had been the case that, if you did something to trigger a dynamic
library being loaded after you started debugging a Java program with
GDB, then GDB would seg fault or do something else to make your life
miserable.  Bug 1322 is the latest problem in this area, as Michael
mentioned, but I'm really referring to earlier bugs, ones that were
present in released versions of GDB.  I fixed one or two bugs related
to this; I know my patch from 2003-05-02 was one of them, and there
might have been another one as well.  I don't have direct evidence as
to how old the problem is, but it's not clear to me that this ever
worked: it seems like it would have been present in the earliest
versions of those files in the public CVS tree.

The "minimal" comment is just there to not encourage people to get
their hopes up that Java debugging works well in GDB: my impression is
that it doesn't.

Anyways, that's what I meant.  As to what the NEWS file should say,
well... I'm not sure that referring to Bug 1322 is actually useful,
because it only had that particular manifestation for a couple of
months, and in particular no release of GDB had that bug.  You could
make my original statement a little more specific by saying "A
long-standing bug causing GDB to seg fault when loading a dynamic
library while debugging a Java program has been fixed."  That doesn't
read too well, though.

David Carlton
carlton@kealia.com