Skip to contentSkip to conversation

SOLVED: mail accepted but not retransmitted?

DISCUSSION Andrew Kohlsmith 3 participants 4 messages Oct 18, 2004

lore
Share & tools
lore
#0 lore raw

Ok wrapping mlmmj-recieve (that is spelled wrong btw, "receive" is how to spell it in English) and drilling down into the strace output I see what is happenning.

- mlmmj-recieve gets called, places mail into incoming queue
- mlmmj-recieve calls mlmmj-process which does its magic and places mail into outgoing queue
- mlmmj-process calls mlmmj-send which takes the queued message and segfaults.

The exact command that is being executed is /usr/local/bin/mlmmj-send -L /var/spool/mlmmj/mlmmj-test/ -m /var/spool/mlmmj/mlmmj-test//queue/72e1474456fce86a

I can run that manually and get a segmentation fault. The strace output shows that mlmmj-send segfaults after trying to do some network calls:

connect(6, {sa_family¯_INET, sin_port=htons(53), sin_addr=inet_addr("192.168.1.1")}, 28) = 0
send(6, "#D\1\0\0\1\0\0\0\0\0\0\6gromit\7mixdown\2ca\0\0"..., 35, 0) = 35 gettimeofday({1098107544, 338999}, NULL) = 0
poll([{fd=6, events=POLLIN, revents=POLLIN}], 1, 5000) = 1
ioctl(6, FIONREAD, [35]) = 0
recvfrom(6, "#D\201\203\0\1\0\0\0\0\0\0\6gromit\7mixdown\2ca\0\0"..., 1024, 0, {sa_family¯_INET, sin_port=htons(53), sin_addr=inet_addr("192.168.1.1")}, [16]) = 35
close(6) = 0
socket(PF_INET, SOCK_DGRAM, IPPROTO_IP) = 6
connect(6, {sa_family¯_INET, sin_port=htons(53), sin_addr=inet_addr("192.168.1.1")}, 28) = 0
send(6, "#E\1\0\0\1\0\0\0\0\0\0\6gromit\0\0\1\0\1", 24, 0) = 24 gettimeofday({1098107544, 342537}, NULL) = 0
poll([{fd=6, events=POLLIN, revents=POLLIN}], 1, 5000) = 1
ioctl(6, FIONREAD, [24]) = 0
recvfrom(6, "#E\201\203\0\1\0\0\0\0\0\0\6gromit\0\0\1\0\1", 1024, 0, {sa_family¯_INET, sin_port=htons(53), sin_addr=inet_addr("192.168.1.1")}, [16]) = 24
close(6) = 0
--- SIGSEGV (Segmentation fault) @ 0 (0) ---
+++ killed by SIGSEGV +++

Drilling down a little more, I see that my /etc/hosts entry for this machine is wrong. I had renamed the machine but did not update /etc/hosts.

Fixing this such that

127.0.0.1 localhost
127.0.0.1 gromit.mixdown.ca gromit

was in there corrected the SIG11 and the list is now sending.

I am shortly going to submit a patch to mlmmj-1.0.0 which uses the RECIPDELIM define throughout; right now only about 1/3 of the code uses it. :-)

Thank you all for your help in tracking this problem down. It was a real bear of a problem and should probably make its way into a FAQ or README somewhere.

-A.

#1 lore raw
Quoted 6 lines · Andrew Kohlsmith

The exact command that is being executed is /usr/local/bin/mlmmj-send -L /var/spool/mlmmj/mlmmj-test/ -m /var/spool/mlmmj/mlmmj-test//queue/72e1474456fce86a

I can run that manually and get a segmentation fault. The strace output shows that mlmmj-send segfaults after trying to do some network calls:

[snip]

Quoted 2 lines

Drilling down a little more, I see that my /etc/hosts entry for this machine is wrong. I had renamed the machine but did not update /etc/hosts.

[snip]

Quoted 1 line

corrected the SIG11 and the list is now sending.

Ok.

Quoted 2 lines

I am shortly going to submit a patch to mlmmj-1.0.0 which uses the RECIPDELIM define throughout; right now only about 1/3 of the code uses it. :-)

We're looking forward to it. It is a well known issue.

Quoted 3 lines

Thank you all for your help in tracking this problem down. It was a real bear of a problem and should probably make its way into a FAQ or README somewhere.

Well, no, it should be patched, if it is a problem in mlmmj. I guess we will have to debug that a bit.

Morten

Signature · 2 lines
Morten K. Poulsen <[email protected]>
http://www.afdelingp.dk/
#2 lore raw
In reply to Andrew Kohlsmith
Quoted 5 lines · Morten K. Poulsen
Quoted 3 lines

I am shortly going to submit a patch to mlmmj-1.0.0 which uses the RECIPDELIM define throughout; right now only about 1/3 of the code uses it. :-)

We're looking forward to it. It is a well known issue.

I just did the patch, now I am just testing. :-)

Quoted 6 lines
Quoted 3 lines

Thank you all for your help in tracking this problem down. It was a real bear of a problem and should probably make its way into a FAQ or README somewhere.

Well, no, it should be patched, if it is a problem in mlmmj. I guess we will have to debug that a bit.

I'm not sure if it's an mlmmj issue or a glibc issue. I'll play a little more.

-A.

#3 lore raw
In reply to Andrew Kohlsmith
Quoted 7 lines · Andrew Kohlsmith
Quoted 5 lines
Quoted 3 lines

I am shortly going to submit a patch to mlmmj-1.0.0 which uses the RECIPDELIM define throughout; right now only about 1/3 of the code uses it. :-)

We're looking forward to it. It is a well known issue.

I just did the patch, now I am just testing. :-)

Awesome!

Quoted 9 lines
Quoted 6 lines
Quoted 3 lines

Thank you all for your help in tracking this problem down. It was a real bear of a problem and should probably make its way into a FAQ or README somewhere.

Well, no, it should be patched, if it is a problem in mlmmj. I guess we will have to debug that a bit.

I'm not sure if it's an mlmmj issue or a glibc issue. I'll play a little more.

Thanks for debugging this, we really appreciate it.

Signature · 4 lines
Mads Martin Joergensen, http://mmj.dk
"Why make things difficult, when it is possible to make them cryptic
 and totally illogical, with just a little bit more effort?"
                                -- A. P. J.