Odoo periodically updates the user online status. This is
done every minute and is throttled in order to ensure it won't
be triggered more than it should.
Since [1], the underscore `throttle` function has been replaced
by the `throttleForAnimation` function. This is not equivalent:
the function won't be triggered if there is no animation which
result in an outdated user status.
This issue was highlighted by the live chat: the button does
not show up when there is no available operator.
Steps to reproduce the issue:
- Log in on Odoo with live chat installed
- Go to a web page that uses the live chat widget
- Wait more that one minute and refresh the page
- The live chat button is not shown anymore: the only
operator is considered as disconnected.
This PR fixes that issue by re-introducing the underscore
throttle function: it is robust and can be needed, not all
throttles are linked to animations.
[1]: https://github.com/odoo/odoo/pull/120437.
opw-3578517
closesodoo/odoo#142160
X-original-commit: f64a7cbad3dee3ffc365b931306b1c13c8bff536
Signed-off-by: Matthieu Stockbauer (tsm) <tsm@odoo.com>
When the bus service is loaded, it connects to the websocket
worker and passes some information that are used to determine
whether the websocket connection is outdated.
Amongst them, the user id is used to detect when the user changed
and when the websocket connection should be refreshed.
This is not enough to ensure the websocket connection is still
relevant: the user id could be the same but the db could have
changed. The websocket would then be linked to the previous
database.
This commit fixes the issue by sending the db name alongside the
user id. This will ensure the connection is refreshed when it is
outdated.
Steps to reproduce the issue:
- Log in on db1 on 2 tabs
- Log out
- Log in on db2
- Try to star a message in the discuss app
- Nothing happens, the messages are sent from db2 but the socket
listens to db1
closesodoo/odoo#141263
Signed-off-by: Alexandre Kühn (aku) <aku@odoo.com>
This patch provides the possibility to implement a custom security layer on top
of Odoo's imbus notification system and ir_cron live triggering system.
The key addition is the `ODOO_NOTIFY_FUNCTION` environment variable (opt-in),
which can now define a custom postgresql function to be called instead of the
default `pg_notify` function. This allows for greater flexibility and control
over the notification and triggering mechanisms within Odoo.
closesodoo/odoo#140089
Signed-off-by: Fabien Meghazi (fme) <fme@odoo.com>
Previously proposed formating was trying to normalize extra in one part
of the path. This means that no extra needed a placeholder.
The implementation was meant to be more generic and extendible since a
part of the logic has to be in website.
A suggestion was made to make it more restricted but explicite by
keeping the url simple in web/controllers/binary.py but adding a
controller in website to add this extra part.
The base extra direction is now in the extension, as the min part.
Initial urls:
/web/assets/{unique}/[{website_id}/][rtl/]{bundle_name}[.min].{extension}
New urls:
/web/assets/[{website_id}]/{unique}/{bundle_name}[.rtl][.min].{extension}
Managed by two routes:
/web/assets/<string:unique>/<string:filename>
/web/assets/<int:website_id>/<string:unique>/<string:filename>
Where filename is in the format {bundle_name}[.rtl][.min].{extension}
Multiple possibilities where proposed
- /web/assets/website/<int:website_id>/<string:unique>/<string:filename>
More explicit but prefixing by /website was considered
- /website/assets/<int:website_id>/<string:unique>/<string:filename>
This one is a litle painfull to match similar attachement, where
website is ignored.
- /website/<int:website_id>/assets/<string:unique>/<string:filename>
Almost accepted but subjective, and anyway two previous solution breaks
the cdn mecanism and would need a migration
- /web/assets/<int:website_id>/<string:unique>/<string:filename>
Almost accepted but subjective, and anyway two previous solution breaks
the cdn mecanism and would need a migration
This last solution was not ideal to match without unique
/web/assets/%/<string:filename> can match both
/web/assets/123456/<string:filename>
and
/web/assets/1/123456/<string:filename>
Anyway, matching without unique shouldn't be supported for al (even if
it is kind of supported with any right now) but it will work by changing
unique wildcard to a more specific one (_ * 7)
closesodoo/odoo#131353
Related: odoo/enterprise#47313
Related: odoo/design-themes#730
Signed-off-by: Xavier Dollé (xdo) <xdo@odoo.com>
The generation inside the rendering has some drawbacks:
- `commit_assetsbundle` is needed for reports rendering because the
template rendering may generate some assets that will be accessed by
another transaction before the transaction is committed. But this
solution is not ideal since the transaction is committed in the middle
of the request
- when the first rendered page is a 404, the assets are not committed
and the page is broken.
- when starting, deleting an attachment can create a concurrent update
error and the request is retried. This will occur once per attachment
and for all worker trying to access the same resource. The whole
transaction is rollbacked, even the previously created assets bundle.
- The cold page load is a slower since there is more work to do.
- Implementing a readonly request is difficult because it could be
transformed to read write and re-executed if the assets bundle does not
exist.
Generating assets when needed solves those issues. The concurrency
when deleting an assets could still occur but only once per bundle, and
in a smaller transaction. This could be solved with a lock now that we
have more control on the transaction. The commit_assetsbundle can be
removed and 404 page should have a correct layout. The cold page load
could be a little faster because the assets bundle can be generated in
parallel requests instead of sequentially when rendering the page.
Part-of: odoo/odoo#131353
This commit removes the promise extension which added the function
`guardedCatch`. It was used to filter server and connection errors from
javascript errors. Instead of using guardedCatch, we should use catch
and check if the reason is a server or connection error if needed and
re-throw the error if it's not handled.
closesodoo/odoo#137702
Task: 3439226
Related: odoo/enterprise#48451
Signed-off-by: Aaron Bohy (aab) <aab@odoo.com>
As a step to simplify model insert from python. All data are
formatted in a way to make data insert trivial in JS models.
closesodoo/odoo#138760
Signed-off-by: Sébastien Theys (seb) <seb@odoo.com>
`_authenticate` is updating the `env`, but it updated the `env` of
`ir_websocket` instead of `self`. There was then a mismatch between
the `update_context` and the current user.
This was highlighted in another PR that removed sudo from these methods
with the hope on relying on ACL, which depends on current user and other
context values at the same time.
closesodoo/odoo#139436
Signed-off-by: Matthieu Stockbauer (tsm) <tsm@odoo.com>
[1] tried to remove nondeterminism from bus tests. However, several
issues remain:
- the asset watchdog bundle is added to the registry during bus tests.
Websocket tests wait for the `connect` event after starting the environment.
This `connect` event can be triggered sooner than expected since the asset
watchdog services starts the bus.
- a the random delay is added to the reconnection attempts.
- `nextTick` is kept when waiting for a client to leave the worker.
This commit fixes those issues.
[1]: https://github.com/odoo/odoo/pull/138221closesodoo/odoo#138639
Signed-off-by: Alexandre Kühn (aku) <aku@odoo.com>
Before this commit, websocket tests were sometimes failing with the
`no result to fetch` error. This is due to a race between the
`subscribe` event being received which itself leads to dispatching bus
notifications interferring with the rest of the test (the connection
is shared).
In order to solve this issue, the `WebsocketCase` now exposes a
`subscribe` method that waits for this dispatching to be done in order
to ensure it will not interfere with the rest of the test.
Some tests were also failing because the subscription was done with 0
as the last notification id. This is an issue since this could lead to
notifications being sent while they are not expected.
In order to solve this issue, `last_notification_id` is now correctly
passed when subscribing to channel notifications.
fixes runbot-24226
closesodoo/odoo#138649
X-original-commit: 3256885fd08d33a67d7165fa33d373fd039ef7af
Signed-off-by: Alexandre Kühn (aku) <aku@odoo.com>
Signed-off-by: Matthieu Stockbauer (tsm) <tsm@odoo.com>
This commit makes bus tests more robust by implementing
functions that wait for some events instead of relying
on `nextTick` that might or not be sufficient.
fixes runbot-23097
closesodoo/odoo#138221
Signed-off-by: Matthieu Stockbauer (tsm) <tsm@odoo.com>
This commit cleans the bus tests by:
- flattening the tests structure by putting tests outside
of the `QUnit.module` function.
- removing unnecessary `assert.expect` calls.
- adding an helper to add common bus service to the registry
Part-of: odoo/odoo#138221
Before this commit, diagnosing websocket handshake error was
cumbersome since the only information available was the return
code of the initial HTTP request (400).
The HTTP stack does not log `werkzeug.exceptions.HTTPException`
and derivated classes since they are valid responses. This commit
adds a log in order to provide more insight about what went wrong.
closesodoo/odoo#138083
X-original-commit: 50af40302c2849b99e8e501bc170cf887ca8e6fa
Signed-off-by: Julien Castiaux (juc) <juc@odoo.com>
Signed-off-by: Matthieu Stockbauer (tsm) <tsm@odoo.com>
Before this commit, owl was in the linter's accepted global variables.
This allowed direct access to owl global object.
For instance, to use xml from owl, you could do :
`const { xml } = owl;`
or you could use it directly:
`owl.xml`
Now, owl is not accepted on linter's global variables anymore, so to
import xml, now you need to use a proper import:
`import { xml } from "@odoo/owl";`
task-id 3498859
closesodoo/odoo#137517
Related: odoo/enterprise#48364
Related: odoo/design-themes#709
Signed-off-by: Samuel Degueldre (sad) <sad@odoo.com>
This commit removes the legacy service mappers which are useless now.
closesodoo/odoo#136807
Task: 3439226
Related: odoo/enterprise#48015
Signed-off-by: Mathieu Duckerts-Antoine (dam) <dam@odoo.com>
- `Response.charset` is deprecated since 2.3, `Response.set_cookie`
accesses the currently-extent internal `_charset` directly until
this too gets removed in Werkzeug 3.0. Add a `_charset` to
`FutureResponse` so this does not crash.
- Bytes response headers are deprecated since 2.3, and will get
removed in 3.0, passing bytes in websocket is completely unnecessary
happenstance which is trivially fixed.
closesodoo/odoo#137145
X-original-commit: 66e3040d4ad9e03aabefebd4799f699ad30fca45
Signed-off-by: Julien Castiaux (juc) <juc@odoo.com>
Signed-off-by: Xavier Morel (xmo) <xmo@odoo.com>
* = bus, calendar, im_livechat, hr, project_todo, sms, snailmail,
test_mail, web, website_livechat
The current implementation often led to tests relying on DOM structure
to properly target the correct element with the text.
It is now easier to simply check if a parent contains some text.
closesodoo/odoo#136295
Related: odoo/enterprise#47770
Signed-off-by: Alexandre Kühn (aku) <aku@odoo.com>
* = account, base_iban, bus, calendar, crm_livechat, hr, hr_holidays,
im_livechat, mrp, project, sms, snailmail, test_mail,
test_mail_full, web, website_livechat, website_slides
Add support in `contains` for most operations that we use in tests.
Remove return value from `contains`.
Move into `web` module.
Remove import/export chains, directly import from correct module.
closesodoo/odoo#134652
Related: odoo/enterprise#47064
Signed-off-by: Sébastien Theys (seb) <seb@odoo.com>
This commit reworks a little bit the backend assets to remove a
bundle and thus save a call at webclient startup. The bundle
"assets_backend_prod_only" existed only to allow to add files in
production, but not in the tests (typically, the file that spawns
the webclient).
This commit introduces a new bundle "web.assets_web" that contains
"assets_backend" and the few files that we only want in production.
In the /web page, we now load "assets_web" instead of
"assets_backend" and "assets_backend_prod_only". In the /web/tests
page, we keep loading "assets_backend", which is now directly
included into "web.tests_assets".
For the sake of consistency, this commit also renames the dark
mode bundle "dark_mode_assets_backend" into "assets_web_dark".
closesodoo/odoo#135204
Related: odoo/enterprise#47316
Signed-off-by: Samuel Degueldre (sad) <sad@odoo.com>
This commit rewrite the module loader to simplify it and to improve error handling.
before this commit, you could have silent errors or "lucky" imports (import a module based on the fact that his asset will surely be load and if not, the error would be silent anyway).
e.g.: The tours cannot import modules extern to tours' bundle anymore. Or load module before your libs in assets.
TASK ID: 3266021
closesodoo/odoo#128502
Related: odoo/enterprise#44100
Signed-off-by: Bastien Pierre (ipb) <ipb@odoo.com>
Before this commit, the "subscribe to single notification"
test failed in an nondeterministic fashion. This commit
solves this issue.
fixes runbot-24033
closesodoo/odoo#134243
Signed-off-by: Didier Debondt (did) <did@odoo.com>
Before this commit, avatar url of author of message always had
style `.o_object_fit_cover`, which makes a nice crop when
ratio of image does not match ratio of the `img`.
When the author of message is a company, it shows the logo of the
company, and it doesn't look nice to crop it.
This commit fixes the issue by using `.o_object_fit_contain` for
avatar of company, so that the logo is fully visible.
When displaying the avatar of a partner whose `is_company` is
undefined, the data is group-fetched, in order for models to
eventually know the value of `is_company` and use the correct
desirable showing.
Task-3381748
closesodoo/odoo#133311
X-original-commit: 48194366f2364e2eb9d99dc4fb71e2de83d8f7a2
Signed-off-by: Alexandre Kühn (aku) <aku@odoo.com>
This commit removes almost all exports of the module services/core.js
(only the bus is left) and adapts the module that imported it.
task 3439226
Part-of: odoo/odoo#133153
*: auth_password_policy, bus, event, im_livechat, mail, mass_mailing,
mrp_subcontracting, point_of_sale, pos_self_order, project, stock,
survey, web, web_editor, web_tour, website, website_event,
website_forum, website_sale, website_slides, base
Historically, the web.assets_common bundle was used to contain assets
that were needed by both the frontend and the backend. In practice, this
caused a bunch of issues where people would add things in assets common
that were not needed by both, and it was also abused as a way to get
bootstrap working in unrelated places by only using that bundle's css.
Because of this, as a first step, the assets_common stop being used in
the frontend, but was left everywhere else.
This commit removes the bundle completely, and moves the files that used
to be in that bundle in the other bundles that need them, this will
allow those bundles to evolve independently going forward.
in im_livechat and mail, some of the unneeded legacy code was removed, this
allows us to avoind including all of the legacy code from web in the
livechat embed bundle and in the dicuss public bundle respectively.
closesodoo/odoo#132190
Related: odoo/enterprise#45884
Signed-off-by: Aaron Bohy (aab) <aab@odoo.com>
Before this commit, the current user was cleared after every
test. This is an issue since we can still interract with the
page after the test is done when in debug mode. This commit
fixes the issue by clearing the current user at the beginning
of each test instead of the end.
closesodoo/odoo#132801
Signed-off-by: Alexandre Kühn (aku) <aku@odoo.com>
In this commit, all usages of env._t() are replaced by _t().
In templates files, env._t() didn't work because terms used
in attributes where not extracted into the translation files.
Only string are exported from .xml files to translation files.
So, to make it works, we set a variable that is then used
in attributes.
For example :
<t t-set="string_to_translate">String to translate</t>
<Dialog title="string_to_translate>...</Dialog>
task-3292454
closesodoo/odoo#131390
Related: odoo/enterprise#45631
Signed-off-by: Michaël Mattiello (mcm) <mcm@odoo.com>
*: bus, calendar, im_livechat, web, website_livechat.
This PR adds the possibility to set the current user during tests as
well as the mail guest. This will be used to test mail guest page and
livechat once the visitors will be treated as guests.
part of task-3332628
closesodoo/odoo#130811
Related: odoo/enterprise#45331
Signed-off-by: Matthieu Stockbauer (tsm) <tsm@odoo.com>
This PR adds the `add_guest_to_context` decorator in order to provide a generic
way to extract the guest from a request. It will be used to unified guest
extraction from cookie/param based on its provenance (external livechat/public
page).
This is better than the `pre_dispatch` method since it can be applied to
specific routes instead of adding this logic to every request.
part of task-3332628
closesodoo/odoo#130052
Signed-off-by: Matthieu Stockbauer (tsm) <tsm@odoo.com>
Before this PR, the user during livechat embed tests was the admin partner while
it should be the public partner. This commit fixes the issue.
part of task-3332628
closesodoo/odoo#130429
Signed-off-by: Matthieu Stockbauer (tsm) <tsm@odoo.com>
In the commit [1], the patch has been refactored to support the
native keyword `super`. The current commit just adapts the codebase
to that change.
task 3410198
[1]: 19ea1ac08043e22a811630968e44715cc3bfc495
Part-of: odoo/odoo#125716
*: im_livechat, web.
Since [1] the messaging service is added to the livechat embed
bundle. As a consequence, the bus service is started even when
it is not needed. This is wrong since it would lead to lots of
useless connections to be made.
This PR prevent mail services from starting the bus eagerly in
this context and adds a tour in order to prevent further errors
on this matter.
closesodoo/odoo#130059
X-original-commit: b2bcef3c3b57061ef0694a5fc3c326cc8babc818
Related: odoo/enterprise#44782
Signed-off-by: Matthieu Stockbauer (tsm) <tsm@odoo.com>
Signed-off-by: Alexandre Kühn (aku) <aku@odoo.com>
*: bus, crm_livechat, crm_mail_plugin, im_livechat, payment, payment_demo,
pos_online_payment, test_discuss_full, test_mail_full, tests, website_livechat,
website_sale.
Several python tests have their own way to make jsonrpc requests. This PR adds a
generic `make_jsonrpc_request` method to the `HttpCase` in order to provide a
generic way to do so.
At the same time, calls to `_open_livechat_channel` are removed in favor of
jsonrpc request to `get_session`. This makes the tests more realistics (some
incoherences were present like passing `country_id` to the open channel method
while the user country id is not set...)
Finally, this will ease the diff in the PR introducing livechat visitors as mail
guests.
part of task-3332628
closesodoo/odoo#130036
Related: odoo/enterprise#44840
Signed-off-by: Sébastien Theys (seb) <seb@odoo.com>
luxon and moment are both used in the solution,
but these two libraries facilitate the manipulation
of dates. It was decided to replace all uses of
moment with luxon so we can then remove
moment.js from the code and lighten the assets.
task-3391739
closesodoo/odoo#128752
Signed-off-by: Bruno Boi (boi) <boi@odoo.com>
One of the main issue with ormcache is that the invalidation clears
everything, meaning that some value, slow to compute but with a long
lifetime, can be removed from the cache because an easy to invalidate
value is cleared, like after writting or creating a product has an
example.
Most example in the code will try to invalidate the cache of the models
doing something like `env['ir.qweb'].clear_caches()` but it is
finally equivalent to `env.registry.clear_cache()`, and cross worker.
The idea is to have multiple cache, maybe with specific sizes for a
specific purpose.
Having one per model is maybe a bad idea because it will be difficult
to size the LRU correcly, and it is too dynamic. Checking invalidation
may be expensive.
The proposed solution is closed allow a limited number of named caches,
using onse sequence per cache. This is actually close to the
cache_longterm.
We want to discourage using a specific cache for one use case in
the buisness code. Adding a cache shouldn't be something easy, doable
in stable.
Note that we could also change the invalisation mecanism using an
insert only table. We an check the sequence of this table, but also
fetch all invalidation messages.
Another possible improvement, especially if we have more than x cache is
to have a global sequence, checking signaling would mean to check the
main sequence, and only the other ones if the main one changed.
Note that this poc is inspired from the long term cache but not all
use case where applie yet.
Part-of: odoo/odoo#119813
Purpose of this commit is to regroup some tests, or rename some too specific
files in order to avoid explosion of files, making tests hard to find when
looking for feature-specific tests.
Task-3346388
Part-of: odoo/odoo#123678
Websocket instances use `selectors.DefaultSelector` in order to
wait for the socket to be ready for I/O. Gevent replaces the
default selector standard implementation by its own, using
`select.select` by default. The issue is that `select.select`
can only handle fd between 0 and 1023 which leads to select
out of range errors when this cap is reached.
In order to solve this issue, this commit enforces the use of
the `PollSelector` when running in evented mode and when it
is available: `select.poll` does not suffer from the same limit
as `select.select` does.
closesodoo/odoo#127692
X-original-commit: 7e8c0dbe118494c94745ba0fde6290db66d610e2
Signed-off-by: Julien Castiaux (juc) <juc@odoo.com>
Signed-off-by: Matthieu Stockbauer (tsm) <tsm@odoo.com>
When user uninstalls the module from one system but in another system same
database was active then error occurs.
Error: A traceback appears: "UndefinedColumn column res_partner.debit_limit does
not exist"
sentry-4057868584
closesodoo/odoo#127230
X-original-commit: ce7c7df7633155bf15d588fadc34901d96f250d5
Signed-off-by: Julien Castiaux (juc) <juc@odoo.com>
This commit improves samesite support for embedded livechats by
allowing the bus service to open anonymous connections.
opw-3336222
closesodoo/odoo#126919
X-original-commit: 2fb6427992514e0254e40e8c7c012f0fc2332e85
Signed-off-by: Olivier Dony (odo) <odo@odoo.com>
Signed-off-by: Matthieu Stockbauer (tsm) <tsm@odoo.com>
Before this commit, a new unread message could contribute to unread
counter of conversation twice.
Steps to reproduce:
1. Log in as Demo and start a conversation with Mitchell Admin
2. Log in as Admin: if this is the first message in the conversation,
counter is correct and shows (1)
3. Log out as Admin
4. Send another message to Admin
5. Log in as Admin and notice that the counter is doubled
6. If the conversation was still open, counter is correct!
This happens because some notification are pending dispatch before
fetching messaging state initially. The counter at init_messaging
fetch takes into account the notification. However, when handling
the message notifcation, it's wrongly considered as a new
notification, thus incrementing the counter when it shouldn't.
This commit fixes the issue by fetching the last bus notification
when fetching init_messaging. Doing so allow to determine whether
a notification is newer or older than init_messaging state.
This helps deciding whether the counter should be incremented or not.
Task-3357960
closesodoo/odoo#124767
X-original-commit: 081ab772dece86f1b5a5d4b3f6227b4b311d9206
Signed-off-by: Stockbauer Matthieu (tsm) <tsm@odoo.com>
1. move cache to _get_asset_paths
The `_get_asset_content` cache has many cache key that are related to a
posprocessing of the `_get_asset_paths` result, the heavy part of this
method. Moving the cache to _get_asset_content will have the benefit
to create less duplicates entries in the ormcache as well as less cache
miss.
To simplify even further, the css and js parameters are removed since
they only filter the output of get_paths, the heavy part of globing the
file will be done before that. Anyway, they are both true when called
from _get_asset_content, and the only other call, in
`_get_related_bundle` don't really need to filter them since it is not
a critical part regarding performance, and the funtional result will
stay the same.
The initial orm cache key was using `_get_template_cache_keys`, a little
overkill and possibly creating duplicates entries again. The only
context key needed is website_id for `_get_related_assets`.
Note that it is not really enough, the orm cache key should actually
contain `request.session.get('force_website_id')` as well has
`request.httprequest.host`. This will be addressed latter since a nicer
solution would be to have website_id as a unique parameter computed
earlier.
2. better _get_asset_paths cache key
The orm cache key was simplified in previous point but there is still
one concern, the website_id depends on more parameters than that:
- request.session.get('force_website_id')
- request.httprequest.host
- existing websites
The idea here is to call `get_current_website` instead of using all
parameters that could define the webiste.
In the same spirit of `_get_template_cache_keys` `_assets_path_params`
can be overriden to give extra params that are usefull to list assets
path. Those params are computed before entering the method
`_get_asset_paths`. This may latter put at a higher level latter, in
get_asset_node, to simplify the _generate_asset_nodes_cache key.
3. better assets_node caches key
The main purpose of this part is to improve ormcache containing assets
nodes. The ormcache key contains
- to much context key
- missing session/host/env info
- unwanted boolean options.
- keys leading to the same cache value
The main goal being to reduce the size of the cache keys, decrease the
number of cache entries and improve the cache hit.
This will also make the behaviour more coherent and hopefully less bug
prone because of mismatch in parameters.
The main reason of the orm cache is the slowness of the validation of
the assets. This includes:
- listing files (dedicated orm cache)
- computing version
The cache key was depending on
- `debug`
The only relevant value for debug is "contains assets"
We dont need to differ between debug='', debug='1', debug='test',
and 'debug=assets', 'debug=tests,assets', ...
- `defer_load`, `lazy_load`, `media`
Those values are only useful to generate html node, a leightweight
operations that does not really needs to be in cache. `media` was also
used in the generation but it looks useless if we have the media on the
node. THIS NEEDS TO BE VALIDATED but in any case, since media is not
used to generate the url, it doesn't make sence to use it in the
generation.
The main idea to remove them from the ormcache key is simply to generate
the nodes outide the ormcached values.
-`async_load`
This one is similar to `defer_load` and `lazy_load` but it looks like
it wasn't used anymore. This was simply removed
- context.get('lang')
The only information needed is the direction, rtl or ltr. This means
en and fr languages, despite sharing the same css assets, will duplicate
the ormcache entries.
-`_get_template_cache_keys`
Only the lang and webiste where really relevant in this flow. Other
keys are actually useless in this flow.
Some information used in the generation where not in the orm cache key
- `self.env.user.lang` if there is no lang in the context
- `request.session.get('force_website_id')`
- `request.httprequest.host`
- ...
The proposed solution is to:
- extract any informùation needed from thecontext, request, environment
before entering the ormcache, reduce it to the minimal possible set of
values needed
```
rtl = self.env['res.lang']._lang_get_direction(self.env.context.get('lang') or self.env.user.lang) == 'rtl'
assets_params = self.env['ir.asset']._get_assets_params() # website_id
debug_assets = debug and 'assets' in debug
```
and remove a leightweight part of the logic
```
def _get_asset_nodes(self, bundle, css=True, js=True, debug=False, defer_load=False, lazy_load=False, media=None):
links = self._get_asset_links(bundle, css=css, js=js, debug=debug)
return self._links_to_nodes(links, defer_load=defer_load, lazy_load=lazy_load, media=media)
```
Where _get_asset_links is the cached part, and _links_to_nodes is the
lightweight part generating the nodes based on the `defer_load`, ....
Additionnal notes:
- data-asset-version and data-asset-bundle are removed from the node
since they don't seem to be used anymore since 65d70acdbf
- async_load is removed since there is no occurence of this in the code.
- a small hack is still needed to pass javascript content instead of
links, this is only to manage css compile error and will hopefully be
removed in the future.
- a context key is still in use to generate the bundle, the
`commit_assetsbundle` but it has no impact on content and will hopefully
be removed in the future.
4. Add test for ormcache hit/miss
In this context, hit/miss is about having the same cache key for the
same result. This test demonstrates the current state, were entries are
create in the ormcache only if the key is really different and will lead
to a different result.
5. remove cache invalidation
This cache invalidation is quite agressive since everytime an
assetbundle is updated, all workers will clear their cache.
The concerned cache by this clear_cache is `_generate_asset_nodes_cache`
throug `_get_asset_nodes`.
The cache is ignored, both in dev=xml and debug=assets.
This clear cache was made conditionnal in 553ea82f81 but this does
not solve an issue we can have in production.
Lets imagine a clean solution
- all sources are updated
- all workers are restarted.
The orm caches are all empty, but since the sources
changed, all bundles will be recomputed. This means that every bundle
updated in database with save_attachement will invalidate the cache of
all workers. Rendering a pdf report of any kind using a specific bundle
will invalidate all cache. Starting a debug=assets for the first time
will invalidate all cache, even if the cache is not used in this case.
But for a regenerated bundle we would expect the ormcache to be:
- empty (did not generate the same bundle yet)
- have the same value (concurrent generation of the same bundle)
Having a different value would mean that the bundle was generated with
another version of the sources. In this case it is maybe even better not
to invalidate the cache since it could lead to an invalidation war
between two workers.
The only case where invalidating this cache is useful is when a bundle
changes, Usually if an ir_asset is created, modified, ...
There is still another rare but possible possibility to have a 404 if
the transaction is rollbacked after populating the assets node cache.
In this case, we only need to clear the cache locally in case of
rollback.
Part-of: odoo/odoo#121376
Odoo Translation Bot