Execution of queue is stopped when an SMTPRecepientRefuse is encountered

Can you provide a traceback?

Of course ;) I replaced domain and email adress with asterisks...

2009-01-21 10:37:23 CRITICAL txn.1123420496 A storage error occurred during the second phase of the two-phase commit. Resources may be in an inconsistent state.
2009-01-21 10:37:23 ERROR Zope.SiteErrorLog http://***/portal_newsletters/channels/newsletter-stats.html
Traceback (innermost last):
  Module ZPublisher.Publish, line 125, in publish
  Module Zope2.App.startup, line 238, in commit
  Module transaction._manager, line 96, in commit
  Module transaction._transaction, line 395, in commit
  Module transaction._transaction, line 503, in _commitResources
  Module zope.sendmail.delivery, line 78, in tpc_finish
  Module collective.dancing.composer, line 413, in send
  Module zope.sendmail.mailer, line 45, in send
  Module smtplib, line 695, in sendmail
SMTPRecipientsRefused: {u'***@***': (450, '4.1.2 <***@***>: Recipient address rejected: Domain not found')}

Hello,

I notice exactly the same problem. If there is an STMTP exception raised the process queue stop but dispatch is not notice of that.
In http://svn.plone.org/svn/collective/collective.singing/trunk/collective/singing/message.py the try, catch doesn't work in that case because of the architecture of zope.sendmail I guess.
I think that collective.dancing.composer.SMTPMailer.send must have an try, cache in that case.
Exception in smtplib are :

"SMTPException","SMTPServerDisconnected","SMTPResponseException",
           "SMTPSenderRefused","SMTPRecipientsRefused","SMTPDataError",
           "SMTPConnectError","SMTPHeloError","SMTPAuthenticationError",

SMTPRecipientsRefused and SMTPSenderRefused are not critical error for the dispatch process and I think they can be catch in the SMTPMailer.
In my case , recipients received hundred message of the same newsletter because the queue failed of a empty email address. The queue is never purge and the @@dancing.utils/tick_and_dispatch finish by blocking zope server.

The traceback :

INFO collective.singing Dispatching is locked by another process.
CRITICAL txn.16292 A storage error occurred during the second phase of the two-phase commit. Resources may be in an inconsistent state.

2009-05-25T09:48:09 ERROR Zope.SiteErrorLog .../@@dancing.utils/tick_and_dispatch

Traceback (innermost last):

  Module ZPublisher.Publish, line 125, in publish

  Module Zope2.App.startup, line 238, in commit

  Module transaction._manager, line 96, in commit

  Module transaction._transaction, line 395, in commit

  Module transaction._transaction, line 503, in _commitResources

  Module zope.sendmail.delivery, line 87, in tpc_finish

  Module collective.dancing.composer, line 376, in send

  Module zope.sendmail.mailer, line 72, in send

  Module smtplib, line 695, in sendmail

SMTPRecipientsRefused: {}

Thanks for your work !
Regards Youenn

I would assume that queuedDelivery does not suffer from this problem.

http://pypi.python.org/pypi/collective.dancing#configuring-zope-sendmail-to-send-out-messages

Patches happily accepted.

I tried to reproduce this issue and it seems to work.

I injected the exception in the __call__ method of collective.singing.mail.Dispatch class

------------------------
from smtplib import SMTPRecipientsRefused
if '<email address hidden>' not in self._split(msg['To']):
    raise SMTPRecipientsRefused('-> SMTPRecipientsRefused raised <-')
-------------------

and all goes fine. An email was correctly delivered to "<email address hidden>", all other mails correcty failed.

Any idea?