This is the mail archive of the
mailing list for the Java project.
Re: [PATCH] java/12857: Fix class literals
- From: Jeff Sturm <jsturm at one-point dot com>
- To: Tom Tromey <tromey at redhat dot com>
- Cc: Andrew Haley <aph at redhat dot com>, <java-patches at gcc dot gnu dot org>, <gcc-patches at gcc dot gnu dot org>
- Date: Sat, 15 Nov 2003 00:35:49 -0500 (EST)
- Subject: Re: [PATCH] java/12857: Fix class literals
On 7 Nov 2003, Tom Tromey wrote:
> >>>>> "Jeff" == Jeff Sturm <email@example.com> writes:
> >> What Mauve failures?
> Jeff> See below. All of these appear to be serialization failures.
> Jeff> The patch surely did affect class layout, but I could not easily see how
> Jeff> these tests are affected. For example, running serialver on the mauve
> Jeff> tests before and after patching did not reveal any change.
> It looks to me that the tests rely on the serialver being consistent
> across compilers, as there is sample data checked in. This is not a
> valid thing to do... compilers generate different code and cannot
> easily be made serialization-compatible with Sun's compiler. See the
> various discussions of this point on this list in the past for more
That makes sense to me, although my problem turned out to have an
entirely different cause: java.io.ObjectStreamClass was miscompiled by
the patched gcj.