This improve previous commit by making sure we
never consider the thread_id/model values in the
In-Reply-To/References header if the host name
did not match.
Also fixes the tests that were using the
6.1 compatibility mode to post in a mail group
thread instead of specifying the right
message-id.
bzr revid: odo@openerp.com-20140411142429-y0rpkzqbrsabxqsg
When emails transit from an OpenERP system to another
there is a chance of wrongly considering the foreign
model,id pair as local, possibly attaching a reply
to a wrong parent message.
This naive heuristic ignore the foreign identifiers
unless the hostname of the message-id mentioned in
the In-Reply-To header matches the local one.
This may fail after a change of local hostname but
is still better than the alternative.
In OpenERP v8 we should stop including the model,res_id
in outgoing Message-Id headers and only rely on the
stored, opaque, unique Message-Ids, preventing
any future mismatch.
bzr revid: odo@openerp.com-20140411101349-h3jmkdzvbr7lut26
found during processing. This way, all emails in the queue are managed and
all failed emails have their own exception in the logs, allowing easier
debugging.
Note that a failed email is set as seen to avoid processing it every time
the scheduler runs.
bzr revid: tde@openerp.com-20140213095717-tcwgkl143i3ujw8h
Indeed assert are statements meant to be used when developping, for
debug purpose. In a production environment it is safer to use real
exceptions that can be managed accordingly.
bzr revid: tde@openerp.com-20140212152737-c7q339psd9hi4iwd
Make sure the user is subscribed before the super() call in mail_thread. This allows to have follower based record rules working (eg: note.note)
bzr revid: mat@openerp.com-20140206104714-anpehaik4wajnfrw
get_filename method from email lib returns sometimes decoded value, sometimes still encoded.
for instance, get_filename cannot decode iso encoded filename
Recently, decode method has been applied on filename string, to handle iso encoded file.
Sadly, one cannot decode already decoded values.
Therefore, decoding already decoded windows or utf-8 encoded filename with special chars failed.
bzr revid: dle@openerp.com-20140102161149-x268xj7bzvvsit2h
- avoid having to browse the record in message_subscribe: refactored message_subscribe
to work on ids without browsing, try to lessen the number of queries to do
- update _set_followers, method of the function field message_follower_ids, to use
message_subscribe and message_unsibscribe
- in create: add author as follower in creation values, not after create, allowing
create access rules based on followers
- updated message_unsubscribe to be the lower method, not calling _set_followers anymore
bzr revid: tde@openerp.com-20131126171752-1irizm7mhv2y1wrn
This new code allows faster computation, because it avoids
- searching in a many2many relationship that is very costly in production
databases, because it is replaced by an 'id in [ids]' equivalent leaf
that can be huge.
- browsing in a create/write (should also lessen the number of issues
about read access rights in create)
A new optionnal parameter is given to the method, that are the
modified values. If this is not given (compatibility-mode) the
records are browsed to fetch the value.
bzr revid: tde@openerp.com-20131114113231-oyk16t9i3m3wul6k
alternative mode is computed when browsing the parts, not from
the message content type.
Added tests.
Also added some notification_email_send to none to avoid sending
emails in demo/data/update.
bzr revid: tde@openerp.com-20130823120611-0n4ull3c8gvwug2u
Previous implementation did a search on mail.followers with subtype_ids=False, leading
to a very bad computation time. This search is now avoided.
bzr revid: tde@openerp.com-20130613082029-lmd63s3mqwxasy4j
[FIX] mail, project, project_issue, crm_lead: fixed 'XX Created' subtype not triggered because of condition
based on state=new and not state=draft.
For tasks, issues and leads, no generic 'Document created' message is posted anymore
because of the 'XX Created' message with subtype automatically logged.
Generic creation message is logged before automatic subscription to enable
message pushing to responsibles.
bzr revid: mat@openerp.com-20130529143022-wy76srwb2nwkspe3
The way attachments are linked to the document has been cleaned. Posting method message_post may receives 2 arguments:
- attachment_ids: those linked to the wizard model (mail.compose.message) are linked to the document (res_model, res_id)
- attachments: the tuples are used to create new attachments linked to the document
The wizard does not set the res_model and res_id of attachments anymore, delegating this job to message_post.
The mail.message and mail.compose.message now use their respective attachment_ids field when possible. This is done instead of reading/creating new attachments based on the attachments tuple each time a mail.compose.message is processed. Email templates now also use attachment_ids, in particular when generating emails, instead of using the attachments tuple. Only reports are still generated on the fly and put into attachments instead of attachment_ids.
A cron job has been added to unlink 'lost' attachments. Those are attachments:
- linked to 'mail.compose.message'
- with res_id=0 (due Chatter used in minimal mode or reports generated by templates before the wizard has an ID)
- with no activity for more than one day (create_date and write_date)
bzr revid: tde@openerp.com-20130411112033-mqph9vjlcjkoolfs
Document created text was not translatable
Subtype was fetched without context, therefore not translated
Removed odd override of _t introduced in mail_followers at revision 7885
bzr revid: tde@openerp.com-20130408092447-3ri41v6xluuj0wha
Before trying to find possible routes, check that the incoming email's
message_id is not already present in mail.message table.
If it is the case, return False.
Parsing of the message has been moved before routing, to avoid looking
for routes for emails we will discard.
Tests have been added and updated.
bzr revid: tde@openerp.com-20130322124809-ven2p5kxpqfjqxb5
Jacques-Etienne Baudoux