Date: Sun, 25 Aug 2002 22:18:36 -0400 From: Jay Sachs <jay@eziba.com> To: Archie Cobbs <archie@dellroad.org> Cc: freebsd-java@FreeBSD.ORG Subject: Re: FreeBSD-specific jikes problem Message-ID: <3D698FFC.3040904@eziba.com> References: <200208232330.g7NNUod23140@arch20m.dellroad.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Freshly -supped ports tree, and -stable as of Aug 7.
Attempt 1: I build the port with no optimization and debugging
on, as follows:
make CXXFLAGS='-g -pipe'
No core dump, just a (jikes) warning about use of '$' in an identifier.
So apparently the optimization in the 4.6-stable c++ compiler is
a bit flakey.
Attempt 2: Turn on optimization, building jikes as follows:
make CXXFLAGS='-g -O -pipe'
Ok, got core dump. Ran gdb, and the stack was filled
with repetitions of
#177 0x82071d2 in Semantic::ProcessBinaryExpression (this=0x8d18000,
expr=0x8d1e474) at expr.cpp:6640
#178 0x81f30c8 in Semantic::ProcessPLUS (this=0x8d18000, expr=0x8d1e4c0)
at semantic.h:1398
883 frames in total. So it looks like the optimizer inserted an infinite
loop. It looks like the gcc 2.95.4 import of June 20 could be the
culprit. The RELENG_4_6 branches (4.6-RELEASE, 4.6.1, 4.6.2 etc)
do NOT have 2.95.4 imported, so if someone could test one of those
systems we could pretty much nail down that it's a 2.95.4 optimizer bug.
jay
> Hi,
>
> If anyone who knows C++ and is interested in fixing a jikes core dump
> problem, please take a look at this bug:
>
> http://www.freebsd.org/cgi/query-pr.cgi?pr=41961
>
> This bug does *not* occur on Linux, so it appears that the bug
> may actually be part of FreeBSD rather than jikes.
>
> I know zero C++ and so can't do much about it..
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-java" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3D698FFC.3040904>