In some cases when subscribing followers code may ask to force subtypes
of the subscription. It means existing followers should have only asked
subtypes independently from existing subtypes. Previously it was done
by unlinking and creating new followers for same partner or channel.
This commit proposes to instead update missing subtypes and remove extra
subtypes.
It has a small impact in real life as forcing subtype is not done often.
Some tests are improved thanks to this commit. This commit is linked to
task ID 1839043. Closes#24417.
Cache invalidation when subscribing people has been added at e6f038a821.
Indeed subscribing partners to a record may lead to an access right update
as some of them are based on followers. This is why a cache invalidation
is necessary to avoid access rights issues.
However subscribing people to a record should change their rights only on
the records involved in the subscription mechanism. We can therefore give
ids to the cache invalidation to limit to updated records.
Cache invalidation is also limited when writing on followers if writing
on model, res_id or partner_id fields. Indeed changing subtypes or
channel of a subscription should have no impact on cache and access rights.
Cache invalidation done manually in _message_subscribe is not necessary
as subscription create or update mail.followers records since f9c210923d.
Create and write of mail.followers records already ask for cache invalidation.
It is therefore not necessary to invalidate cache twice.
This commit allows to save a few queries on some tests, notably about
activities that deal with subscription and messages. As cache is now kept
it is not necessary to refetch some data, leading to a few query gain. We
gain about 1K queries on com runbot.
A side effect of limiting cache invalidation is that some unit tests require
a manual cache invalidation to have up to date results. Indeed record not
being up to date in cache was hidden by the invalidation we just removed.
This commit adds new methods to add or update followers to documents. Purpose
is to have new optimized methods that perform less queries. It is based on
experience and use cases of various addons in Odoo using it.
Technically two main methods are given on mail.followers model
* _add_followers: this method generates values required to create new
followers or update existing one. It takes several parameter controlling
how to deal with existing followers. This is used notably when auto
subscription should not erase existing subscriptions;
* _insert_followers: this method creates or updates followers for a given
set of documents. It uses result from _add_followers;
Performance gain come from the use of a SQL query to fetch followers data
as well as less use of record sets. Code can be more low-level in some
places. Purpose is to avoid browse and prefetch and just fetch data once
and use it in various computation.
New implementation is equivalent to the old one. It offers more flexibility
as caller can either receive values for followers or directly have records
created and updated. Various parameter allow to tune computation depending
on each specific call to the new follower methods.
Future commits will gradually replace the use of old _add_follower_command
by the new methods. Purpose is to replace them one by one in order to see
the gain in term of performance one replacement at a time.
Now that we're closer to switching to P3 for good, these helpers have
outlived their usefulness, and mostly add noise.
All remaining dict.iter*() or dict.view*() must be converted to the
normal keys(), values() or items() calls.
Whenever the result is likely to be used for more than the scope of a
loop, or when the dict needs to be modified during iteration, the calls
must be wrapped in a ``list()``, to protect the new P3 semantics.
Those cases are very exceptional.
Also removed some dead code or improved the API to remove unnecessary
conversions.
Purpose
=======
Currently an employee can
- Create mail.followers
- Write on mail.follower if it is related to his own partner
When creating a record with followers (example: An expense sheet), the followers are created but the function fields are recomputed and stored on the table (if store=True).
In that particular case (Create: ok, Write: ko), this recomputation will trigger a write and then an access right error with a transaction rollback.
Specification
=============
Adding the `compute_sudo=True` attribute on the `res_model` field, the recomputation will be allowed for users at the record creation.
This fix doesn't imply to update the modules.
In Python 3:
* various builtins and dict methods were changed to return
view/iterable objects rather than lists
* and the separate Python 2 view/iterable builtins and methods were
removed altogether
This is problematic when using these items as list (which the happens
repeatedly in Odoo), but more viciously when iterating *multiple times*
over them (which also happens, which I've messed up multiple times while
writing this, and which is a pain to debug even when you've just created
the issue).
Convert all code using these to semantics-matching cross-version
helper functions to get the LCD behaviour between P2 and P3, and
forbid the builtins via lint.
issue #8530
- Create a SO with a new product configured as service, Track Service
"Create task and create hours" and link to "Internal - GAP Analysis".
- Confirm the SO
- Add a new line for the same product in the SO and save
A Odoo Server Error (Validation Error) appears "Error, a partner cannot
follow twice the same object"
It seems that the value of the related field `res_model` is not
correctly computed in the cache. As a workaround, we can use the
`res_model_id`.
Courtesy of @mart-e
opw-709790
Fixes#15589
Remove unrestricted "read" access. To make code internally using `ir.model`
work, add a private method `_get` on `ir.model` to retrieve the record
corresponding to a model name, without access rights issue.
Change signature of method `get_authorized_fields` to make it use a model name
instead of a model id. This removes the necessity of a search on `ir.model`.
This is a manual forward-port of commit 03ac47b9d7
Currently when someone is added as follow to a document default subtypes
are computed if no specific subtypes are given to the subscription. However
if internal subtypes are set as default those are added to all new
followers. It should take into account customers and shared users and avoid
adding them internal subtypes.
Customers and shared users are not notified of messages using an internal
subtype. So currently there is no information leak. However with this fix
we ensure internal subtypes are not followed by external people even if
no notification is created.
Compared to original commit tests are slightly updated because there are now
default internal subtypes in data. Not in 10.0 and previous versions.
Currently when someone is added as follow to a document default subtypes
are computed if no specific subtypes are given to the subscription. However
if internal subtypes are set as default those are added to all new
followers. It should take into account customers and shared users and avoid
adding them internal subtypes.
Customers and shared users are not notified of messages using an internal
subtype. So currently there is no information leak. However with this fix
we ensure internal subtypes are not followed by external people even if
no notification is created.
From #14704, backport of 03ac47b9
Currently when someone is added as follow to a document default subtypes
are computed if no specific subtypes are given to the subscription. However
if internal subtypes are set as default those are added to all new
followers. It should take into account customers and shared users and avoid
adding them internal subtypes.
Customers and shared users are not notified of messages using an internal
subtype. So currently there is no information leak. However with this fix
we ensure internal subtypes are not followed by external people even if
no notification is created.
When uninstalling a module followers stay in the database because there is no
real link between ir.model and followers. Adding a m2o ensure followers are
correctly removed when uninstalling a module.
res_model fields is kept as a stored related field. This allows to keep
compatibility with almost all code using / searching on followers. Only
followers creation is changed a bit. As follower creation is done mainly
through a dedicated method few changes are required. Yey. Mat should be
happier now.
fix test
Main changes :
- mail.notification model is removed. People do not receive notifications
anymore. Instead two ways of following documents exist
- using a channel; messages will be displayed on the channel itself
in a near future commit
- following with its partner; messages will all be considered as
needaction, using a new m2m table. People should receive less
needaction messages by following less records by themselves.
There is no more read / unread state anymore. Instead only needaction
messages are considered. Todo (Favorites) messages still exist, and are
stores on a new m2m table instead of using decorated notifciations.
- the main filter for documents is not message_unread anymore, but
message_needaction. A lot of views and filters have been updated
accordingly.
- the vote feature has been removed
Followers can now be partners or channels. Partners following a document
will receive needaction, as previously. However people can follow documents
through channels. Members of a channel are able to listen to a stream
of messages using the channel. Those messages do not create needaction
messages. It is therefore possible to follow documents without receiving
too much notifications. For interesting documents subscribing with its
partner will create notification.
message_follower_ids fields is udpated. It is now a many2many to
mail.followers, not to res.partner anymore. A subscription can be either
a partner (partner_id) or a channel (channel_id).
Some access rules have been updated accordingly.
Move the notification management from mail.notification to res.partner model.
Those methods belong more to the partner model, as they work on partners.
Moreover with the upcoming new model for NewChatter notification will be
heavily refactored.
model has been renamed to mail.channel to prepare the slack modeling.
In future commits the mail.group model will be merged with the channel
model from im_chat. The first move is to rename mail.group into
mail.channel to have a model that will unite both features.
Thibault Delavallée