On Sun, 2011-03-20 at 10:50 -0400, Matt Elson wrote: > I'm having the problem on an Intel 32-bit Linux machine running 5.8.8 > with the same version of re2c, so it looks like the common thread is > Intel 32 bit + re2c. I'll see if I can throw up 64 bit machine to test > further.
We saw the problem on an x86_64 machine, running Perl 5.10.0. # uname -a Linux [redacted] 2.6.26-2-amd64 #1 SMP Tue Mar 9 22:29:32 UTC 2010 x86_64 GNU/Linux # re2c -version re2c 0.13.5 # spamassassin -V SpamAssassin version 3.3.1 running on Perl version 5.10.0 # dpkg -l perl-base spamassassin Desired=Unknown/Install/Remove/Purge/Hold | Status=Not/Inst/Cfg-files/Unpacked/Failed-cfg/Half-inst/trig-aWait/Trig-pend |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad) ||/ Name Version Description +++-==============-==============-============================================ ii perl-base 5.10.0-19lenny minimal Perl system ii spamassassin 3.3.1-1~bpo50+ Perl-based spam filter using text analysis (it's a Debian Lenny machine, with SpamAssassin from Lenny backports.) # grep model.name /proc/cpuinfo | uniq model name : Intel(R) Xeon(R) CPU 5160 @ 3.00GHz So it doesn't seem to be exclusive to 32-bit x86 architecture. On the other hand, we had another machine running the same Debian version, also 64-bit, also a Xeon (though different speed), with the same versions of perl-base and spamassassin, and also using compiled rules, which did *not* see the problem. So it's puzzling. At first I commented out the rules and recompiled, but then I discovered that adding meta __PILL_PRICE_1 (0) meta __PILL_PRICE_2 (0) meta __PILL_PRICE_3 (0) to /etc/spamassassin/local.cf as suggested by Karsten Bräckelmann worked fine --Jay
