Isn't .dovecot.sievec a binary file that should be recreated on the target platform? /Lars
Julian Cowley wrote: > I just transferred the files in my home directory from an x86_64 machine > to an i686 machine. This included the .dovecot.sieve and > .dovecot.sievec files in my home root, and the Dovecot index files in my > ~/Maildir directory. Everything seemed to be working until I tried > sending an email that triggers a Sieve "redirect" command, then I got > the following in the logs. I then tried the same dovecot.sieve file > using a separate account which has always been on the i686 machine (and > the index files have always been written there too), and the Sieve > "redirect" worked fine. There were core files written in my home > directory which I can post if necessary. > > Both the x86_64 and i686 machines are CentOS 4.5 running Dovecot 1.0.5 > and Dovecot Sieve plugin 1.0.1. > > Sep 15 21:44:10 myhost deliver(julian): Trying to allocate 0 bytes > Sep 15 21:44:10 myhost deliver(julian): Raw backtrace: > /usr/libexec/dovecot/deliver(i_syslog_panic_handler+0x24) > [0x80a08b4] -> > /usr/libexec/dovecot/deliver [0x80a06a9] -> > /usr/libexec/dovecot/deliver [0x80a91bd] -> > /usr/libexec/dovecot/deliver [0x80574f8] -> > /usr/libexec/dovecot/deliver(duplicate_check+0x51) [0x8057671] -> > /usr/lib/dovecot/lda/lib90_cmusieve_plugin.so [0x133f1f] -> > > /usr/lib/dovecot/lda/lib90_cmusieve_plugin.so(sieve_execute_bytecode+0x549) > [0x141b59] -> > /usr/lib/dovecot/lda/lib90_cmusieve_plugin.so(cmu_sieve_run+0x3fd) > [0x134c4d] -> > /usr/lib/dovecot/lda/lib90_cmusieve_plugin.so [0x1339f8] -> > /usr/libexec/dovecot/deliver(main+0x10c8) [0x8056e28] -> > /lib/tls/libc.so.6(__libc_start_main+0xd3) [0x506de3] -> > /usr/libexec/dovecot/deliver [0x80550a1] > Sep 15 21:44:10 myhost postfix/local[4924]: E1947344AE: > to=<[EMAIL PROTECTED]>, > relay=local, delay=0.35, delays=0.17/0.01/0/0.17, dsn=5.3.0, > status=bounced (Command died with signal 6: > "/usr/libexec/dovecot/deliver") >
