Status of bug #974290 unknown encoding crashes qrunner?
What is the status of these bugs?
bug #974290 unknown encoding crashes qrunner
bug #926034 crashes on encoding errorsI seem to recall that Tokio Kikuchi or maybe Jim Tittsler had a fix, but there's nothing about it in the tracker, and I can't find it in my mailman-developers archive. There are a few posts with the words "unknown encoding" in them on mailman-users and mailman-developers, but no fixes. :-(
I can probably get my host to install the codecs (I assume that would help), but this really shouldn't crash the runner.
I'm seeing this under Mailman 2.1.5, on a fairly motley RedHat system (I think it was running 5.x when we started hosting our lists there, it's been upgraded piece-by-piece, often from source, since then). Of course it's all Chinese spam in GB2312, but it would be nice if it didn't fill the error log and the shunt queue, and simply evaporated instead.<wink>
-- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.
Hi,
Stephen J. Turnbull wrote:
What is the status of these bugs?
bug #974290 unknown encoding crashes qrunner bug #926034 crashes on encoding errorsI seem to recall that Tokio Kikuchi or maybe Jim Tittsler had a fix, but there's nothing about it in the tracker, and I can't find it in my mailman-developers archive. There are a few posts with the words "unknown encoding" in them on mailman-users and mailman-developers, but no fixes. :-(
Well, these problems are essentially from absence of codecs for the charset used. You will need to upgrade python to 2.2 for 'windows-1252' was introduced in python 2.2 and not present in 2.1 for the latter #926034 case. You may want to upgrade python to 2.4 because gb2312 and CJK charset was finally supported in the Python distribution.
I can probably get my host to install the codecs (I assume that would help), but this really shouldn't crash the runner.
I'm seeing this under Mailman 2.1.5, on a fairly motley RedHat system (I think it was running 5.x when we started hosting our lists there, it's been upgraded piece-by-piece, often from source, since then). Of course it's all Chinese spam in GB2312, but it would be nice if it didn't fill the error log and the shunt queue, and simply evaporated instead.<wink>
You can reduce the number of shunts by using privacy filters. In the forthcoming mailman 2.1.6, list-request address is more robust against the spams having alian charset.
Again, all developers are encourged to upgrade your installation to the latest CVS (Release_2_1-maint branch).
Cheers,
Tokio Kikuchi, tkikuchi@ is.kochi-u.ac.jp http://weather.is.kochi-u.ac.jp/
"Tokio" == Tokio Kikuchi <tkikuchi@is.kochi-u.ac.jp> writes:
What is the status of these bugs? bug #974290 unknown encoding crashes qrunner bug #926034 crashes on encoding errors
Tokio> Well, these problems are essentially from absence of codecs
Tokio> for the charset used.I was afraid of that.
I'll submit a patch in a few days.
-- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.
Hi,
Stephen J. Turnbull wrote:
"Tokio" == Tokio Kikuchi <tkikuchi@is.kochi-u.ac.jp> writes:
What is the status of these bugs? bug #974290 unknown encoding crashes qrunner bug #926034 crashes on encoding errors
Tokio> Well, these problems are essentially from absence of codecs Tokio> for the charset used.I was afraid of that.
I'll submit a patch in a few days.
Try the latest source from CVS before start hacking. CookHeaders.py in question is heavily patched for inclusion of sequence number in subject_prefix. The bug may be fixed as a side effect. ;-)
If you feel trouble downloading from CVS, here is an unofficial release of mailman-2.1.6a: http://mm.tkikuchi.net/mailman-2.1.6a.tgz
Cheers,
-- Tokio Kikuchi, tkikuchi@ is.kochi-u.ac.jp http://weather.is.kochi-u.ac.jp/
"Tokio" == Tokio Kikuchi <tkikuchi@is.kochi-u.ac.jp> writes:
Tokio> Try the latest source from CVS before start
Tokio> hacking. CookHeaders.py in question is heavily patched for
Tokio> inclusion of sequence number in subject_prefix. The bug may
Tokio> be fixed as a side effect. ;-)I don't have that choice; it's not my installation.
Seems a pretty unlikely side effect, anyway, since CookHeaders.py isn't in some of the traces for me. (I just noticed that on the most recent SHUNT.) What's happening is that outside of the try that catches the LookupError for the explicit call to unicode(), there is a call h.append(s,c) or so, which comes from email.Header, which is _not_ protected.
All of the traces I've seen so far have h.append() in them. There's at least one more of these crashes in CommandRunner.py, due to a call to make_header(). Basically, I'd say that anything that calls into email.Header is at risk.
-- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Ask not how you can "do" free software business; ask what your business can "do for" free software.
participants (2)
-
Stephen J. Turnbull -
Tokio Kikuchi