Hi,

Barry, regarding the hang of the browser, I've found the problem: there were many lock files for that mailing list in the locks/ directory (I've killed qrunner with -9 since it was eating 99% CPU). Removing them solved the hang.

I'd like to contribute to the testing of the current development release, I've tryed anonymous access to the CVS and it works :).

Now my question is: can I simply issue a "configure && make install" in the CVS tree or there are some other action that I must consider? :)

TIA. Regards, Luca

On Fri, Aug 31, 2001 at 01:34:03AM -0400, Barry A. Warsaw wrote:

...

...

...

...

...

"LM" == Luca Maranzano <liuk@publinet.it> writes:

LM> I've already submitted a bug to SF about this, but I can
LM> confirm that if I set dont_respond_to_post_requests to NO then
LM> this seems to arise a bug in Hold.py, i.e. the variable lang
LM> is not set and I get the following error:

Thanks, but this is already out-of-date w.r.t. cvs :). I've closed the bug, but appreciate your report.

-Barry

On Fri, Aug 31, 2001 at 10:45:04AM -0400, Barry A. Warsaw wrote:

LM> Now my question is: can I simply issue a "configure && make
LM> install" in the CVS tree or there are some other action that I
LM> must consider? :)

Depends on whether you're upgrading an existing site or testing it out in a separate installation (which I recommend for now). If the latter, then yup that should do it. If you're upgrading over an existing site, then read the instructions in the UPGRADING manual. Part of "testing it out" is also testing that upgrading is as painless (or as documented <wink>) as possible.

I've upgraded from 2.0.5 to 2.1a2 since I read on the Web site that this was the (literally from http://www.list.org/MM21/index.html) "current stable GNU release". Only after the upgrade I digged into the list archives of mailman and I discovered that version 2.1 is still in alpha stage :-/ Ok, no matter, I like to live on the bleeding edge :) (I hope my users of the lists like too ;-) so I've decided to go on with CVS and to help as far as I can in this testing phase.

I'm running a site with about 10 lists, all low traffic except one with ~500 subscribers and a traffic of some messages a day.

This morning I came back from one week on holyday and since the bug about dont_respond_to_post_requests has created some troubles, I've updated to 2.1a3 of the latest CVS tree (updated just this morning).

BTW, the upgrade seemed painless :).

But I've to report this traceback from logs/error, that showed up some time after I restarted mailmanctl:

[This happened 1 time only] Sep 08 11:14:09 2001 (17075) Uncaught runner exception: EOF read where object expected Sep 08 11:14:09 2001 (17075) Traceback (most recent call last): File "/home/mailman/Mailman/Queue/Runner.py", line 104, in __oneloop self.__onefile(msg, msgdata) File "/home/mailman/Mailman/Queue/Runner.py", line 152, in __onefile keepqueued = self._dispose(mlist, msg, msgdata) File "/home/mailman/Mailman/Queue/IncomingRunner.py", line 111, in _dispose status = self._dopipeline(mlist, msg, msgdata, pipeline) File "/home/mailman/Mailman/Queue/IncomingRunner.py", line 134, in _dopipeline sys.modules[modname].process(mlist, msg, msgdata) File "/home/mailman/Mailman/Handlers/Hold.py", line 159, in process hold_for_approval(mlist, msg, msgdata, NonMemberPost) File "/home/mailman/Mailman/Handlers/Hold.py", line 234, in hold_for_approval id = mlist.HoldMessage(msg, reason, msgdata) File "/home/mailman/Mailman/ListAdmin.py", line 176, in HoldMessage self.__opendb() File "/home/mailman/Mailman/ListAdmin.py", line 77, in __opendb self.__db = marshal.load(fp) EOFError: EOF read where object expected

While this is repeated several times (~10):

Sep 08 13:09:59 2001 (17080) Uncaught runner exception: 'None' object has no attribute 'startswith' Sep 08 13:09:59 2001 (17080) Traceback (most recent call last): File "/home/mailman/Mailman/Queue/Runner.py", line 104, in __oneloop self.__onefile(msg, msgdata) File "/home/mailman/Mailman/Queue/Runner.py", line 152, in __onefile keepqueued = self._dispose(mlist, msg, msgdata) File "/home/mailman/Mailman/Queue/CommandRunner.py", line 99, in _dispose if mlist.bounce_processing and
File "/home/mailman/Mailman/Bouncers/BouncerAPI.py", line 63, in ScanMessages addrs = sys.modules[modname].process(msg) File "/home/mailman/Mailman/Bouncers/SMTP32.py", line 48, in process if not mailer.startswith('<SMTP32 v'): AttributeError: 'None' object has no attribute 'startswith'

From the logs I can say that mailmanctl has prcessed some pending requests, but I don't know how to trace down the problem. Let me know if you need other info.

My be sooner or later I'll be able also to submit some patch ;-) for the moment I can only report problems as soon as they happen.

Thanks for your attention. Regards, Luca

...

...

...

...

"RFK" == Ricardo F Kustner <ricardo@rixhq.nu> writes:

>> I've upgraded from 2.0.5 to 2.1a2 since I read on the Web site
>> that this was the (literally from
>> http://www.list.org/MM21/index.html) "current stable GNU
>> release".

RFK> I also noticed alot of new mailman users start with 2.1a
RFK> instead of going for 2.0.6 ... maybe this isn't advertised
RFK> strongely enough?  I'd hate to see losing new users because
RFK> they think it isn't stable enough.

Good point! I'll update the index.html for the 2.1 pages.

Thanks, -Barry