Before this commit, when obtaining link URL suggestions, both the
specific and the matching generic page were suggested.
After this commit, only the most specific ones are kept in the suggested
list.
This commit also adapts the sitemap in the same way.
In stable, a condition on a dedicated context key is used in case those
methods were called with the goal of obtaining both generic and specific
pages.
In 16.0, those methods will always filter duplicates pages as it was
supposed at first.
Steps to reproduce:
- Edit Contact Us page (to create a specific view)
- Edit the Contact Us menu
- Type "/" in the URL
=> "/contactus" appeared twice.
task-2968292
closesodoo/odoo#112746
X-original-commit: c5a50362ef55ced373c53c7af069fd5534043ae2
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
Signed-off-by: Benoit Socias (bso) <bso@odoo.com>
Commit [1] improved the url dependencies search behavior to include more
models to search in but also the multi record capability (needed for
multi delete in list view now that website is in the backend).
But a line of code was badly designed, making the perfs horrible.
That line code shouldn't have been part of the loop as it doesn't
depend of the loop.
This was drastically more impactful on the `/` page.
Benchmark: for the `/` page, searching in 1000 product template website
description will go from 29.74 seconds to 0.29 seconds.
See the speedscope result on the PR description.
For ~10.000 products, it will go from ~7 minutes to 1.35s.
[1]: https://github.com/odoo/odoo/commit/6ac17b93437868cbefbe13448a6fcbb29953f221
task-3169378
closesodoo/odoo#111933
X-original-commit: 9816b2ba6ee85dcba7b2f85e70c617994b7748c4
Signed-off-by: Christophe Monniez (moc) <moc@odoo.com>
Before this commit conditions based on `_handle_visibility` and
`_get_cached_visibility` did work only by relying on the cache of the
`menu.page_id` being populated when accessing `is_visible` in sudo.
This does not work if the cache is cleared between the calls.
This commit makes sure all 3 conditions have access the record.
The actual issue has not been reproduced locally yet.
The various workers, crons, websocket work on distinct envs - even
through code they cannot impact the cache of another local env outside
the `check_signaling` system which is only used between requests.
For the problem to occur, some intra-request multithreading is needed
but it could not be located so far.
task-3149270
closesodoo/odoo#111882
X-original-commit: a864192ecd270848fede23d3eb053318d07ae8e8
Signed-off-by: Benoit Socias (bso) <bso@odoo.com>
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Before this commit, if a link to a page was not correct because of a
case mismatch, it would simply land on a 404 page.
While it's correct, as URL are case sensitive, it leads to a few bad UX
flow at the admin/editor level:
- Create a link in your page (on a text or a button eg), type an URL
which does not exists (to create it after) like /Page
- Click on the link/button you just made, you are redirected to /Page
which display a 404 with the "Create page" option (correct)
- When you click on that button, it will actually create a page with
/page URL, leading to a mismatch between the URL you created and the
page URL.
Your link/button will still lead to a 404 URL as it points to /Page.
Since it's just a fallback when an exact URL match is not found, it
should not break anything and should not have bad impact at any level
(seo/speed etc).
Indeed:
- It's done through a 302 redirect
- `_serve_page()` is already a fallback case, so it will only make
the `website.redirect` and 404 cases a bit slower due to the extra
search query.
The only possible scenario seems to be if the user (mind the uppercase):
- Created a /Page page
- Created a redirect from /page to /another-page
In this case, /page won't land on /another-page but on /Page.
This flow seems unlikely and is not actually wrong either way.
At least, it certainly is less important than ensuring a case
insensitive fallback.
Finally, note that another solution would have been to either:
- Force page URL to lower case.
-> This is not stable friendly, people might be relying on this to
create pages with different casing:
`/Batman-VII-The-Dark-Knight-Whatevers`, while not recommended,
doesn't sounds idiot.
On top of not being stable friendly, we probably want to keep
offering this possibility
- Redirect all URLs to lowercase endpoints.
-> This is obviously not stable and not Odoo's jobs. It should be
something decided by the sysadmin and done at nginx (etc) level.
task-3110294
opw-3104030
closesodoo/odoo#111736
X-original-commit: f05491105f93939490cbeb078cb7653c38685644
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
Since [1] when the translations becames stored as JSONB, the
translations of website-specific views were lost when adding a new
language.
[2] did fix a similar problem when installing an App, but did not
solve this.
This commit makes sure to only change the requested translations
without impacting the existing ones when a language is added.
Steps to reproduce:
- Add French language to the website
- Add the number snippet on the homepage
- Translate "Useful options" -> "Options utiles"
- Go to setting -> Languages -> Activate any language (not even on a
website)
=> French translation disappeared.
[1]: https://github.com/odoo/odoo/commit/ef00294e7189359c47638c4a71626f1937395edb
[2]: https://github.com/odoo/odoo/commit/91a9c870eede96040ac49d2375fe0e18342342e1
opw-3120079
closesodoo/odoo#111378
X-original-commit: 52fd577de8e270f84e8ca23c9350f2fb3ab5a7d8
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
- Ignore anchors, those are not sent to the server anyway, no way to
compare even if we wanted to
- Ensure query string (qs) are the same to be considered equals
On top of that, it also fixes the case when the user inserted an
absolute URL instead of a relative one, it will now match.
task-3096367
opw-3091427
closesodoo/odoo#107782
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
- `clean_url` method was wrongly flag as `api.model`
- the `unslug_url` was not called for the URL comparison in the dropdown
case, meaning that `/shop/prod-1` would not match `/shop/product-1` as
it should (and as it does for regular non dropdown menu)
- the `active` class was actually never working for the dropdown case,
as the class was added on the wrong element (`li` instead of `a`)
The code was hard to read (mainly because huge python conditions in XML)
and kinda redundant, going through an util method should be clearer and
help reading the template XML.
task-3096367
opw-3091427
Part-of: odoo/odoo#107782
Since the refactoring of website.visitor with [1] (upsert), the
`access_token` is supposed to be holding the same value as the
`partner_id` when the visitor is linked to a partner:
- Anonymous visitor: no `partner_id`, `access_token` is a hash value
- Partner visitor: `partner_id` set, `access_token` should be sync with
`partner_id`.
`partner_id` is just a stored computed field holding the `access_token`
value if it is an integer value.
There should never be a case where there is a `partner_id` set and the
`access_token` is not equal to the `partner_id`.
For instance, having a visitor with `partner_id` = 4 and `access_token`
= `e4r3ejkj4` is supposed to be impossible.
It would lead to crash, because the visitor is only searched based on
his `access_token`, meaning that when searching for the visitor of
partner 4, none would be found and a new one would try to be created,
raising the `uniq_access_token_id` SQL constraint.
While the `partner_id`/`access_token` sync might seems weird, it is done
to allow the `upsert` use in SQL to improve perfs of this low level
behavior.
It's actually not as weak as it seems as there is only a single entry
point to update the `partner_id` and `access_token`: the authenticate
override of website.
Those fields are not supposed to be changed elsewhere.
Note that modifying the `access_token` would not be an issue as the
`partner_id` is just a stored compute based on the `access_token`.
But it's only true when modifying through the ORM as if you do that in
raw SQL, it won't go through the `api.depends` which is supposed to
recompute the stored computed `partner_id` field.
But something was forgotten during the initial dev: the partner merge
behavior: it does (on top of other thing) auto discover the m2o field
relations that points to a `res.partner` and modify those values in raw
SQL to the new value.
This is obviously wrong regarding the `website.visitor`'s `partner_id`
field, the `access_token` should also be updated, or when possible
visitors should be merged too.
Note that for DB upgrated from previous version to Odoo 16, this is
ensured through the following upgrade script [2]:
```sql
UPDATE website_visitor
SET access_token = partner_id::text
WHERE partner_id IS NOT NULL
```
[1]: https://github.com/odoo/odoo/commit/d348bed1ad9d3d16b295f013f015706be6c07820
[2]: https://github.com/odoo/upgrade/commit/0cedcbf70494dfeddeb5a97c13bf875cb6a86886
task-3148111
closesodoo/odoo#111002
X-original-commit: a87b4142dd4a2c05e3e1885b2c54f5e0d3c7ac47
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Before this commit, this flow was broken:
- Install website
- Create a new website of your own (not using the one created
automatically from XML data)
- Choose another color palette for that website
- Uninstall the website app
- Reinstall the website app
- Try to choose another color palette for any website
=> It does not work
Indeed, after the uninstallation, the DB is left in an invalid state:
the SCSS customizations attachments of the website that was created by
the user are not removed, they just have their website_id field emptied.
Some code made at [1] was already there to remove those attachments. The
problem is that it only worked for websites which were created by XML
data (at website installation), not by the user. Indeed, the `unlink`
method is not called during uninstallation to remove records that were
created by the user, thus the `unlink` override was not called either.
See [2] for some details.
This fixes the issues by moving this attachment cleaning code in a
dedicated method, called in `unlink` but also in the `uninstall_hook` of
the website app.
This also takes the opportunity to refactor the code involved, in
particular to not even consider customized attachments which do not have
a website_id.
[1]: https://github.com/odoo/odoo/commit/2f361bec36dff09181b96d140d62c477cdf013a1
[2]: https://github.com/odoo/odoo/pull/97852#pullrequestreview-1067851656
opw-3127531
closesodoo/odoo#110338
X-original-commit: 988eafa03b57be3b3a7f110650c61ce8fda88e31
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
RATIONALE
Improve usage of composer in comment or email mode: support batch-posting in
comment, support more configuration from templates, improve global model.
SPECIFICATIONS
Support a real res_ids field on mail.compose.message model. Instead of relying
on active_ids from context, store it once for all at composer level and use
it in code. Active_ids usage is still done at default_get level, using it to
populate the field.
Improve usage of domain, renamed to res_domain to match other document related
fields naming. Add support of a res_domain_user_id field allowing to set the
user from which the domain should be evaluated.
Composer now runs on a list of IDs. Mass mail mode and comment mode are now
distinct from running on a singleton or on more records. Rendered or raw
mode is not triggered by
* mass mailing mode: always display raw mode, whatever the number of records;
* comment mode: display rendered mode when having a single record (like the
previous comment mode). Display raw mode when having either no records
either at least two records.
Task-3035101 (Mail: Support batch-posting from composer)
Part-of: odoo/odoo#99482
Since it's first implementation at [1], a new cursor is created to then
build a new environment after the website's `authenticate` overide with
the new `uid` of the logged in user.
Since the new visitor SQL upsert refactoring done at [2], it seems to be
introducing a SQL deadlock sometimes.
It has been detected on odoo.com while monitoring the logs. Despite
being quite rare, it still happens too often due to our heavy trafic.
Step to reproduce (among others):
- Install website_livechat
- Login as admin on 127.0.0.1
- On 127.0.0.2, load the website
- Open the livechat
- Type something in the discussion
- Directly try to login as admin
- It will load for a certain time then crash on a 502 timeout due to the
SQL deadlock
- You may need to repeat the process to actually face the bug
[1]: https://github.com/odoo/odoo/commit/6bec0e4d29e6b33b74962b2893a7a405667ef58c
[2]: https://github.com/odoo/odoo/commit/d348bed1ad9d3d16b295f013f015706be6c07820closesodoo/odoo#109830
X-original-commit: b241cf7de9329af1410b9dd45b161aa41926effb
Signed-off-by: Olivier Dony (odo) <odo@odoo.com>
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Before this commit, the behavior disabling the unused snippet assets
performed two checks:
1. It was looking for occurrences for its use in the snippet template
2. It was looking for occurrences for its use in the HTML fields
3. Checked on the occurrences of step 1 and 2 and return result.
In many cases there are already coincidences in the first step, making
the second step unnecessary since this second one is the slowest.
Matches are now checked between steps 1 and 2 to skip the second if
matches are already found.
closesodoo/odoo#109488
X-original-commit: 56100532dbcbdeb5487662d0cbf35ba6d62b15b5
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
When editing a website form, a `search_read` request is fired on
`ir.model` model to get the list of possible form actions (create
ticket, create opportunity, etc.).
But since commit [1] in Odoo 15, non-admin users don't have read access
anymore on this model, leading to a traceback when clicking on a form in
edit mode.
Steps to reproduce (as designer):
- Install only website and login as admin
- Make "portal" user an internal user and give him "Editor and Designer"
rights
- Login as "portal" user
- Enter edit mode on any page and drag & drop the form snippet, or
simply go to /contactus page which already has one
- Click on the form -> Traceback
Steps to reproduce (as publisher/restricted editor):
- Install website_hr_recruitment
- Make "portal" user an internal user and give him the following rights:
- Website: Restricted Editor
- Recruitment: Administrator
- Login as "portal" user
- Go to a job page like /jobs/detail/experienced-developer-4
- Enter edit mode and drag & drop the form snippet
- It will crash
[1]: https://github.com/odoo/odoo/commit/5dc4cff60a557e14b08440c227423291c407899b
opw-3098097
opw-3101884
closesodoo/odoo#109339
X-original-commit: 054c216e93b818794316bfd4b8ce56d6853a61d3
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
Steps to reproduce the bug:
- Install the "Surveys" application.
- Go to the "Website" application.
- Go in the page manager ("Site" > "Pages").
- Select a page (typically "Home").
- Try to delete it.
- Error of type "'survey.survey' object has no attributes 'name'".
The error comes from the fact that records do not necessarily have the
attribute called "name". To fix this error, this commit uses
"display_name" that is defined in all models.
task-3082345
opw-3100483
opw-3111671
X-original-commit: d5aa5d7ebf9cd17a6df054252a1313ec8b85e83e
Part-of: odoo/odoo#109329
request.geoip is no more a dictionnary cached in the session. It is now
a full blown object with lazy and smart geolocalisation capabilities.
Among other things, the previous dictionnary API is now deprecated. The
changes are:
* `request.geoip['country_name']` -> `request.geoip.country_name`
* `request.geoip['country_code']` -> `request.geoip.country_code`
* `request.geoip['city']` -> `request.geoip.city.name`
* `request.geoip['latitude']` -> `request.geoip.location.latitude`
* `request.geoip['longitude']` -> `request.geoip.location.longitude`
* `request.geoip['region']` -> `(request.geoip.subdivisions[0].iso_code if request.geoip.subdivisions else None)`
* `request.geoip['time_zone']` -> `request.geoip.location.time_zone`
It is safe to access all the attributes. Doing `request.geoip.city.name`
when the geolocalization failed (missing db, invalid address, ...)
evaluates to None. It does not raise an AttributeError.
Task: 2848206
Part-of: odoo/odoo#91337
Before this commit, we have redundant loc in sitemap.
When we iter the rules from the routing map, the comparison between
two same endpoints are not equal since httpocalypse.
After this commit we compare the function of the endpoint to know if we
have already processed it or not.
X-original-commit: c408de398c6b4e59cd9a233df251147d8750b9f2
Part-of: odoo/odoo#108670
Prior to this commit, adding an anchor using auto-complete would not
properly prefix the url with the URL of the current page.
Steps to reproduce:
- Create a page (e.g. /test), add blocks and create an anchor (e.g.
- Save the page
- Click on "Edit Menu" in "Pages"
- Add a new menu entry
- Type '#' in the URL field and select the anchor
- Save
=> The menu entry is wrongly set as it does not contain the original
page's URL inside the `href` attribute. This means that clicking on the
menu from any other place will not properly redirect to the anchor on
the /test page
On top of that, editing a menu entry that was previously linked to a
page would change the URL of the page with an anchor, making it
impossible to access the page anymore until you change the URL back in
the page manager.
Steps to reproduce:
- Create a new page (e.g. test)
- Add a new menu entry that has this page as its URL
- Save the new menu entry (close the dialogs)
- Re-open the Menu dialog and edit the newly created menu entry
- In the URL field, replace it with the page + an anchor
(e.g. /test#anchor)
- Save
=> The page's URL is now /test#anchor and is no longer accessible
The is due to the `save` method of the `website.menu` model not
containing code to properly manage adding an anchor to the menu.
This commit fixes that.
opw-3067750
closesodoo/odoo#108502
X-original-commit: 192bf08da191e6fad27dd2913be0801817256a3a
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
To reproduce the issue:
- Go to website > Add language > Add blocks to your homepage and
translate it.
- Make translations on the "Contact us" page (or a new custom page).
- Install a new app > The translation of the homepage is lost (only on
homepage).
Explanation:
After [1], The `_load_module_terms()` method was updated on website
module to copy translations from base to specific views when module
translation is loaded. And since [2], the translated field's column is
either NULL or a JSON dict mapping language codes to text (the field's
value in the corresponding language).
To explain what happens exactly when new app is installed, let's suppose
we want to load module translation for the config {'en_US', 'fr_BE'} in
the following situations:
S1 (e.g. "Contact us" page):
-=-=-=-=-=-=--=-=-=-=-=-=-=-
```
generic_arch_db = {
'en_US': '<div>Generic (EN)</div>',
'fr_BE': '<div>Generic (FR)</div>'
}
- specific_arch_db = {
'en_US': '<div>Specific (EN)</div>',
'fr_BE': '<div>Specific (FR)</div>'
}
```
`_load_module_terms()` will copy translations for 'fr_BE' from generic
to specific `arch_db` (using generic translation dictionary).
result:
```
new_specific_arch_db = {
'en_US': '<div>Specific (EN)</div>',
'fr_BE': '<div>Updated Specific (FR)</div>'
}
```
S2 (new custom page):
-=-=-=-=-=-=--=-=-=-=
```
generic_arch_db = NULL
specific_arch_db = {
'en_US': '<div>Specific (EN)</div>',
'fr_BE': '<div>Specific (FR)</div>'
}
```
result:
Nothing to do here (only specific version), the specific translation
will not be updated.
S3 (e.g. website homepage):
-=-=-=-=-=-=--=-=-=-=-=-=-=
The generic view has no translated `arch_db` (only the 'en_US' version):
```
generic_arch_db = {'en_US': '<div>Generic (EN)</div>'}
specific_arch_db = {
'en_US': '<div>Specific (EN)</div>',
'fr_BE': '<div>Specific (FR)</div>'
}
```
`_load_module_terms()` will copy generic translations only for languages
available on "generic_arch_db" and the 'fr_BE' version will be lost.
result:
`new_specific_arch_db = {'en_US': '<div>Specific (EN)</div>'}`
The goal of this commit is to prevent this behaviour by keeping the
specific translated `arch_db` even when the generic value has no content
for the translation language.
Remark: This behaviour occurred while loading module translations,
as a consequence, the specific translations are also lost when the
website module is updated.
[1]: https://github.com/odoo/odoo/commit/94db81d8d28c4cbc7b51ae1688c9362042ee3619
[2]: https://github.com/odoo/odoo/commit/ef00294e7189359c47638c4a71626f1937395edb
opw-3083480
closesodoo/odoo#108477
X-original-commit: 91a9c870eede96040ac49d2375fe0e18342342e1
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
It was reported the following error:
```
psycopg2.errors.DatatypeMismatch: column "country_id" is of type integer
but expression is of type boolean
LINE 9: false, 1, NULL, 4, 4,
```
The root cause is unknown, it's not yet sure how this is happening
exactly.
This commit is then fixing naively the consequence.
Somehow, GEOIP is returning a country code which does not exists in Odoo
for this error to appear. It's normally not possible.
There is multiple possible cases that would lead to this error:
- The user modified the `code` field of a country in his DB (unlikely)
- The user deleted a country in his DB (unlikely)
- The GEOIP database is returning a country which does not exists in
Odoo. This is the most probably case, even if the latest up to date
`GeoLite 2 City` seems to have the exact same country code list as in
Odoo 16.0 (250 exact same ones).
Moving the country_id search directly in the query will prevent the
issue and is better anyway to avoid an useless query everytime when
GEOIP is enabled.
Perf wise, it doesn't seems to alter at all the query speed.
closesodoo/odoo#108232
X-original-commit: 42fe2540620b7529d9a848c4570290cbf64a4f48
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
One can very well select a "auth=user route" as homepage for his
website, like /my.
There would then be an issue with such an URL being set as homepage:
- As the user landed in the homepage controller which is auth=public,
the system will add the public user as env user (see
`_auth_method_public()`).
```
@http.route('/', type='http', auth="public", website=True, sitemap=True)
def index(self, **kw):
```
- Then, that controller will reroute to the homepage url (/my). The
request.httprequest.path will now be /my
- Then, that controller will recall the dispatcher stack:
`request._serve_ir_http()`
- From there, this call won't fail as it should because the user is
considered as logged in as it went already through the
`_auth_method_public()`, adding the public user as env.user.
`_auth_method_user()` won't raise its error.
- The /my page will be rendered despite not being logged in.
Once the user land on that page, the system will actually detect him as
logged out on a page supposed to be accessed when logged in.
It will then:
1. Show a toaster to inform the user
2. Auto reload the page
As the current URL is still `/` (due to the reroute and not a redirect),
the same flow will happen again, and again, looping forever.
--------
Note that accessing /my directly won't be an issue, as it won't go
through the homepage controller (which is auth=public), so there won't
be a user_id set on the env (the public user), meaning that the dispatch
layer will reject the access and raise an access error (through
`_auth_method_user()`.
Also note that the same behavior will occur when going through the first
menu fallback mechanism:
- if the user didn't setup any homepage_url
- and deleted his / website.page
- and his first website menu is /my
In that case, it will go through the first menu redirect fallback, which
will be working fine (as it's a redirect and not a reroute).
With both those 2 flows (going through redirect), the user will
correctly land on the login page (which will redirect to /my once logged
in).
------
Finally, the other solution would be to prevent such a configuration
(/my as homepage_url) but it was not easily doable (if doable at all),
see https://github.com/odoo/odoo/pull/99100#discussion_r963055251
------
A test is also added and over the more complexe case (which should cover
everything):
- With /my as homepage_url
- With the / website.page deleted
- With /my as first menu URL
-> Accessing / as public user should:
1. Reroute to /my (because of the homepage_url set to it) which
should fail now thanks to this commit
2. Since the reroute / re-serve failed, it should reach the
"first menu fallback" mechanism, which is also /my
3. That fallback should be a redirect, not a reroute, so the user
should actually land on the login page
4. Once logged in on that page, it should properly redirect to /my
opw-3077339
X-original-commit: 4f5899f769266edc27a0b2d812df092b95f43c71
Part-of: odoo/odoo#107759
Prior to this commit, when using a background video on the cover of the
website_slides homepage (/slides), after saving, a traceback would
occur and the video would not be displayed.
Steps to reproduce:
- Install website_slide
- Go to /slides
- Enter edit mode
- Change the cover of the page by a video
- Save
- Traceback
The reason for that is that the cover snippet used on the homepage of
elearning is the root of an XPath instead of its own oe_structure. (This
was changed in 16.0)
[1] made it so that the style and class attributes are saved but did not
save any data attributes needed for the background-video widget to work.
This commit adds those data attributes.
[1]: https://github.com/odoo/odoo/commit/5b252215dc375526af87c31a32256dd1a9582f02
opw-3063043
closesodoo/odoo#107577
X-original-commit: 86dc4d4349d033c95c0e04e8c173f63e5e2044ed
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
The main goal of this commit is to reduce the size of the registry by
removing the (almost) useless __last_update field.
Statistics # of fields with all modules installed:
before 30184 fields, 1299x last_update (4.30%)
Before this commit, the computed field __last_update was added on every model.
The idea behind this field was to have a computed field that had either
the write_date or the create_date if the write_date was empty. However,
the write_date is always written, even on creation, making it useless
to have the computed field __last_update
After this update, we completely remove from BaseModel:
* __last_update
* CONCURRENCY_CHECK_FIELD that was always defined as "__last_update"
* _compute_concurrency_field that was the compute function for __last_update
closesodoo/odoo#105739
Task-id: 3062140 (part of 3062137 improve registry load time)
Related: odoo/upgrade#4038
Related: odoo/enterprise#33939
Signed-off-by: Raphael Collet <rco@odoo.com>
*: website
Commit [1] introduced the possibility to configure "website filters"
used in the website to display some records dynamically via the "dynamic
snippet" and its variants. One of their field is `limit` which is a max
number of records the filter allows to show (so that the client side
cannot ask for thousands of records). Users can configure the limit used
by a snippet via the editor: they are allowed to choose up to 16, still
safe-guarded by the internal limit configured on the website filter on
the python side.
The problem here was that [2] introduced website filters for the blog
posts but set up a max limit of 6. Thus breaking the editor option if
the user choose a limit between 7 and 16. This commit fixes the issue,
for newly configured snippets (as a stable fix) or for users who would
-u their blog application. EDIT: as the master version of this fix, all
blocks will now use their set limit. We consider that is an acceptable
change after migration that the user can change by himself if he wants
to.
Steps to reproduce:
- Install blog application
- Add a "Blog Posts" snippet
- Set "Fetched Elements" to 10 (there are 7 records in demo data)
=> Only 6 are still shown
[1]: https://github.com/odoo/odoo/commit/0e7640b5f22d2bea04bbe22d3189cff7e03af545
[2]: https://github.com/odoo/odoo/commit/3c0d98bcd8adf9325ee3497eb8d25ec7f904d6a5
opw-2885948
closesodoo/odoo#107179
X-original-commit: 3a3dc8729aa2beec49602fe4c141043f60effdf1
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
This commit ensures that when searching on a `inherits` field of
a model (like `name` of `website.page`), it works properly when
`pg_trgm` is activated.
Indeed, `name` is a field of `website.page` record but only at the ORM
level, not in SQL, due to how `inherits` works.
So, when the `pg_trgm` extension is enabled, it will switch from ORM
queries to raw SQL query (to use the native SQL similarity feature and
not our custom python/orm one, as obviously the SQL one is better, more
powerfull/accurate and faster).
But this will actually make the code fail when searching on fields from
a model which has `inherits` and that you search on that `inherits`
model fields.
Note that in 15.2, the `pg_trgm` extension is auto installed when
possible thanks to [1] and [2].
[1]: https://github.com/odoo/odoo/commit/eedf37d6e286b995c47b946be1a6b66817094eff
[2]: https://github.com/odoo/odoo/commit/75e6b645acdc95aece506ce0249fd0760838281c
opw-3063592
closesodoo/odoo#107178
X-original-commit: 3d978caad9de2010603581687ee39a5240b74074
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
With commit [1] we refactored the cookies bar to let the user decide if
he wants to accept the cookies (and/or only part of it).
Before that commit, the `website_cookies_bar` cookie could only hold
`true` as value, which now is holding an object like
`{"required": true, "optional": false}`.
This creates an issue if a user is coming from a previous version with
`true` as cookie value because since the refactoring it will crash both
in JS and PY because `in` instruction with a boolean value will fail in
both languages.
The decision taken here is simply to remove the cookie if we face such a
case so the user can decide again what he wants (since there is more
choices now).
It also means that we won't be holding an outdated value in the cookie
any longer.
[1]: https://github.com/odoo/odoo/commit/2cbda6c98ee947cea1d06c09880eee8c758304a8
opw-3074303
closesodoo/odoo#106621
X-original-commit: 22392755ed4c504e6ba2a5a9c9eb6643f2c95994
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Before this commit, `view_hierarchy` extended the QWeb view to see
the hierarchy of views.
With the evolution of the javascript framework, this use case has
now more sense in a client action rather than an extension of a
QWeb view.
This commit refactors `view_hierarchy` into a client action.
Part-of: odoo/odoo#103477
*: test_website, website, website_blog, website_event, website_forum,
website_sale, website_slides
Reverts this refactoring because it breaks multi-tenancy: if several
databases run on the same server and different modules are installed
in each database, the global variable becomes wrongly populated.
Steps to reproduce:
On runbot:
- Go to the `-base` URL or simply select the `base` DB
- Drag & Drop the search snippet (and let it set to "Search in
Everything")
- Try to use the snippet -> Crash on unknown model
On Odoo.com saas DB:
- Drag & Drop the search snippet (and let it set to "Search in
Everything")
- Try to use the snippet -> Crash on unknown model (unless you installed
each and every module related to the search snippet)
Locally:
- odoo-bin -d first -i website --stop-after-init
- odoo-bin -d second -i website_blog --stop-after-init
- odoo-bin -d first,second
- Try to use a search bar on "first" database (using "Search in
Everything") => it tries to access the blog models.
Revert of PR: #98423
opw-3045092
closesodoo/odoo#105077
X-original-commit: 5806f71af1b6f44aac063f7e504be736300f699d
Related: odoo/enterprise#33669
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Signed-off-by: Benoit Socias (bso) <bso@odoo.com>
This fix allows generating the sitemap (via `website.search_pages`) via
RPC. The changes are necessary as since #99667 the `request` object is
no more available to RPC-executed functions.
See also: the `test_search` test case of `website.tests.test_page`.
closesodoo/odoo#104567
X-original-commit: 8f4e214417cfb1aed7e595b8d6ab71f33a347b29
Signed-off-by: Julien Castiaux <juc@odoo.com>
`is_view_active` relies on the `website` specified in the context.
This commit turns `is_view_active` to a model method, to avoid that its
callers think the `website` on which it is called is the one used.
closesodoo/odoo#103126
X-original-commit: 52d436706bf650bb14f2bc2a398d107ff99f81ea
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Previous commit adapted the URL dependencies screen following the
website frontend > backend merge done at [1] and [2].
It allowed to pass other records than website.page and also added the
multi record capability.
This commit is going a step further, by searching for the URL in all
HTML fields and not only views + pages + menu + blog.
It also let the XML take care of the wording instead of the python.
closesodoo/odoo#103136
X-original-commit: 6ac17b93437868cbefbe13448a6fcbb29953f221
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
This is following the website frontend > backend merge done at [1] and
[2].
Before that improvement, the old page manager had a button to delete a
page which was behaving as the one in the page properties dialog: it was
showing the list of (possible) dependencies as a confirm step.
But since [2], that delete button was removed as the action menu of the
list view already has a delete button, which is better as:
- It is hidden and take no space, deleting a page is rare
- It is known by odoo users as all list views have that button
- It handles multi delete
So this commit basically just restore that delete warning step for that
list view delete button, and also make it possible to use that
dependencies warning dialog for multiple pages, not only one.
It also now handles records in a generic way, not only website pages.
It's needed because now, all the main Odoo frontend records are sharing
a list view mixin (see `js_class="PageListController"`).
It also fixes the fact that the text of the collapse were inside a font
awesome class, basically using a weird font.
[1]: https://github.com/odoo/odoo/commit/31cc10b91dc7762e23b4bde9b945be0c4ce3fe3b
[2]: https://github.com/odoo/odoo/commit/940f4ee875332dafa1f379970a7683be6b3ee606
X-original-commit: 11db2f6ed81419ac724ff27ac95a4438d670cbd5
Part-of: odoo/odoo#103136
This is basically reverting commit [1] which was done a few years ago
as improvement to [2], but it was a bit overkill IMHO.
Indeed, there is a sort of hack when creating a page with special
extension like "my-page.js" that will actually create a special page
with special arch. The goal is for such pages to be `t-call`ed later.
That's commit [2].
On top of that, when deleting such a special page, we introduced a
mechanism to show the views and pages that would possibly `t-call` the
that page which is requested to be deleted.
That's commit [1].
It basically is mimicking what was already done when you change the URL
of a normal page, we tell the user where that url is actually possibly
used.
But since the recent improvement in Odoo 16 done at [3], the page
manager is now a backend list view. It means that multi delete is now a
thing.
Thus, the page dependencies behaviors (key and url) need to be
refactored to handle multiple given URL/Key and not just one.
We choose to remove the key dependencies part (only used for those
special pages) instead of adapting it:
- That 'hack' is probably almost never used
- It's some code to maintain, eg now we need to refactor it
- We would need some extra code to make it only triggered for website
pages and not all records, unlike the url dependencies screen which
concerns all records
- That's an advanced feature (special pages), if you are using it, you
probably knows how to handle a website and you don't need us to remind
you where this special page's view is used.
[1]: https://github.com/odoo/odoo/commit/a91a3a563338e746d6d23cd57551af30e322e367
[2]: https://github.com/odoo/odoo/commit/727d461e1d2bcec4571665b90b6d1630f671a0a3
[3]: https://github.com/odoo/odoo/commit/31cc10b91dc7762e23b4bde9b945be0c4ce3fe3b
X-original-commit: 77f03abc56b147171eeba66bc55df5ee92c8af1a
Part-of: odoo/odoo#103136
make searching translated field language dependent
add an extra filter using trigram index to speed up '=' and 'like' search
X-original-commit: 5b6f0a9e5f9d60735e306709a3e7d7ac46a586be
Part-of: odoo/odoo#103031
*: website_forum, website_hr_recruitment, website_sale
Before this commit, most of the backend redirections to the
WebsitePreview, introduced in [1], were using ir.actions.url with the
get_client_action_url util, which was not optimal.
This commit changes that to use a new get_client_action method, which
returns the ir.actions.client record. It will execute the action
directly, avoid to redirect the router before, and improve performances.
Also, it allows to solve bugs, as:
- The back navigation from the WebsitePreview to the ListViews.
Going through the /web controller would add an entry in the history, and
going back on it would reload the client action.
[1]: https://github.com/odoo/odoo/commit/31cc10b91dc7762e23b4bde9b945be0c4ce3fe3b
task-2687506
closesodoo/odoo#102991
X-original-commit: 30fb11e479fb7b3db3649e55fcbdb06bfcdb398c
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
The goal of this commit is to add some fixes on page properties dialog
after the Owl REF [1].
- XML: use default form view style (remove </group>).
- Move the '/' back into the non editable part of the URL.
- Prevent python code from creating a different URL (and optionally
setting "website.rewrite" record for it) when the new URL is the same
as the initial one after slugify.
- 'useAutofocus()' on the first page properties field.
[1]: https://github.com/odoo/odoo/commit/61a9d7bd2abc6081b321a51d329f9e85209215e5
task-2687506
closesodoo/odoo#102869
X-original-commit: 1165dfcd4fd5677d14f3d3329ad7c62be847b31a
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Before this commit the neutralize system introduced in v16 was using ORM
methods in order to change appropriate records. Although flexible, this approach
could lead to call some methods with side effects while neutralizing
(eg: overloads of write).
This patch converts the neutralize system to a safer "inert" SQL based approach
by migrating the generic method _neutralize to SQL files exposed in the
data folder.
Task id: 2961687closesodoo/odoo#102792
X-original-commit: e5dbded9bb363351feff7ca8a56c7f8a6860f492
Related: odoo/enterprise#32580
Signed-off-by: Fabien Meghazi <fme@odoo.com>
The /web loading was slow during test when some themes are installed.
This is because without any context and request, the assets are
generated for website 1 before this fix.
With this default website:
_get_active_addons_list returns only addons that are not themes
_get_asset_paths will return less assets
When loading /web, no website is found because the domain is not empty.
This means that the web.assets_backend will have less assets when
generated outside of a request than in a request not matching any
website.
There is normaly no assets for the backend in theme BUT a tour is added
in each theme
('/theme_anelusia/static/src/js/tour.js', 'theme_anelusia', 'website.assets_editor')
('/theme_artists/static/src/js/tour.js', 'theme_artists', 'website.assets_editor')
('/theme_avantgarde/static/src/js/tour.js', 'theme_avantgarde', 'website.assets_editor')
('/theme_aviato/static/src/js/tour.js', 'theme_aviato', 'website.assets_editor')
('/theme_beauty/static/src/js/tour.js', 'theme_beauty', 'website.assets_editor')
('/theme_bewise/static/src/js/tour.js', 'theme_bewise', 'website.assets_editor')
('/theme_bistro/static/src/js/tour.js', 'theme_bistro', 'website.assets_editor')
('/theme_bookstore/static/src/js/tour.js', 'theme_bookstore', 'website.assets_editor')
('/theme_buzzy/static/src/js/tour.js', 'theme_buzzy', 'website.assets_editor')
('/theme_clean/static/src/js/tour.js', 'theme_clean', 'website.assets_editor')
('/theme_cobalt/static/src/js/tour.js', 'theme_cobalt', 'website.assets_editor')
('/theme_enark/static/src/js/tour.js', 'theme_enark', 'website.assets_editor')
('/theme_graphene/static/src/js/tour.js', 'theme_graphene', 'website.assets_editor')
('/theme_kea/static/src/js/tour.js', 'theme_kea', 'website.assets_editor')
('/theme_kiddo/static/src/js/tour.js', 'theme_kiddo', 'website.assets_editor')
('/theme_loftspace/static/src/js/tour.js', 'theme_loftspace', 'website.assets_editor')
('/theme_monglia/static/src/js/tour.js', 'theme_monglia', 'website.assets_editor')
('/theme_nano/static/src/js/tour.js', 'theme_nano', 'website.assets_editor')
('/theme_notes/static/src/js/tour.js', 'theme_notes', 'website.assets_editor')
('/theme_odoo_experts/static/src/js/tour.js', 'theme_odoo_experts', 'website.assets_editor')
('/theme_orchid/static/src/js/tour.js', 'theme_orchid', 'website.assets_editor')
('/theme_paptic/static/src/js/tour.js', 'theme_paptic', 'website.assets_editor')
('/theme_real_estate/static/src/js/tour.js', 'theme_real_estate', 'website.assets_editor')
('/theme_treehouse/static/src/js/tour.js', 'theme_treehouse', 'website.assets_editor')
('/theme_vehicle/static/src/js/tour.js', 'theme_vehicle', 'website.assets_editor')
('/theme_yes/static/src/js/tour.js', 'theme_yes', 'website.assets_editor')
('/theme_zap/static/src/js/tour.js', 'theme_zap', 'website.assets_editor')
Making the bundles for /web different with or without theme, and with or
without website id.
The proposed fix wont return a website if not specified in context and
if not during a request and if not forced to fallback.
The side effect is that the frontend assets where generated with a
website id 1 before that and it won't be the case anymore. This may be
a problem that could slow down frontend call on website 1 when design
theme is installed.
closesodoo/odoo#102690
X-original-commit: a5ed957b37b966bc50ed1219b6af65ed818f04a0
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
When loading a registry from route with auth="none" and where module
'website' is set as 'to install', request.env will be None and install
will crash as follows:
```
Traceback (most recent call last):
...
File "/home/odoo/src/odoo/saas-15.3/odoo/modules/registry.py", line 88, in new
odoo.modules.load_modules(registry, force_demo, status, update_module)
File "/home/odoo/src/odoo/saas-15.3/odoo/modules/loading.py", line 482, in load_modules
processed_modules += load_marked_modules(cr, graph,
File "/home/odoo/src/odoo/saas-15.3/odoo/modules/loading.py", line 371, in load_marked_modules
loaded, processed = load_module_graph(
File "/home/odoo/src/odoo/saas-15.3/odoo/modules/loading.py", line 303, in load_module_graph
module.write({'state': 'installed', 'latest_version': ver})
File "/home/odoo/src/odoo/saas-15.3/addons/website/models/ir_module_module.py", line 78, in write
if request and request.context.get('apply_new_theme'):
File "/usr/lib/python3/dist-packages/werkzeug/local.py", line 348, in __getattr__
return getattr(self._get_current_object(), name)
File "/home/odoo/src/odoo/saas-15.3/odoo/http.py", line 1029, in context
return self.env.context
AttributeError: 'NoneType' object has no attribute 'context'
```
closesodoo/odoo#102343
X-original-commit: fcf6e462e116d13533014e31d4191763354811cf
Signed-off-by: Xavier Alt (xal) <xal@odoo.com>
*: im_livechat, survey, utm, website_crm_iap_reveal, website_forum,
website_livechat, website_sale, website_sale_comparison
Before this commit all cookies were considered essential.
This commit makes some of them optional. It also makes it possible for
the website visitor to only accept the essential cookies.
task-2800976
X-original-commit: 9a8a9463289a7446e9be0ef62ff895feb37a4de4
Part-of: odoo/odoo#101845
Co-authored-by: Benoit Socias <bso@odoo.com>
Before this commit, the only possibility when adding a google font was
to use google servers to serve the font.
This was not ideal as some people really want to serve it themselves
without the need of their visitors to reach google servers.
That's especially true since recently where it seems like German clients
are receiving letters about that to tell them it's illegal and this
should be changed, as it wouldn't respect the GDPR.
Somehow, it seems related to the fact that google knows you visited a
website by just downloading the font, because they very well know with
just your IP who you are exactly.
It's yet unsure if that issue is well-founded or not, but since German
courts seem to be sanctioning people about this, there is no reason to
not at least provide a workaround.
What is sure is that it makes a lot of noise and more and more people
seem to be impacted by this as many opw are getting opened, as well as
github messages.
Whether it is well-founded or not is thus not really our problem
anymore, we should just provide a way for our users to protect
themselves against this "German law problem" (or at least think they are
protecting, if Odoo thinks that's a non issue or the German court is
wrong or ambiguous).
Note that a cookies banner to inform users would not be enough for that
"problem", as the user would already have accessed your website and thus
the related problematic fonts.
Another solution which is not something we want (at all) would be to
serve local system fonts while the user did not consent about google
fonts, or having a blocking screen page telling people visiting the
website will fetch google fonts. Obviously those 2 possibilities are a
no go as it leads to terrible UX.
Finally, note that:
- in Odoo 16, the default fonts will be the system fonts, meaning there
won't be any call to google by default, regardless of this pr
- there is a work in progress to improve the current cookies bar to
differentiate essential and non essential cookies and to allow user
to accept only one or both (task-2800976).
Useful links:
- https://github.com/odoo/odoo/issues/83638#issuecomment-1054470699
ODO detailed point of view about this
- https://rewis.io/urteile/urteil/lhm-20-01-2022-3-o-1749320/
The German law about this
Closes#83638
task-2756486
opw-2970167
opw-2960466
opw-2960555
opw-2952427
opw-2800976
opw-2748647
(possibly many more)
Courtesy of @bso-odoo for the regex part which was inspired by another
of his google font fix attempt
closesodoo/odoo#101826
X-original-commit: b06ce21eba6388ce34bbffffadcb489f0e8557dd
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
If an user modifies a theme value through the "Theme" tab of the website
builder, a custo of the related SCSS file is made. To make a SCSS custo,
the related python methods need to know for which bundle the custo is
made. This is to ensure that if a file appears in multiple bundles, the
custo targets the right one (amongst other things). For those automatic
SCSS custo made via the website builder, the given bundle does not
matter as those are custo made in variables SCSS files, which appear in
all bundles, more precisely in a sub-bundle included in all bundles. The
code should be more robust to leverage that fact.
When the "assets_frontend" received all "assets_common" files in order
to only use one unique bundle for main frontend pages with [1], it was
though smart to leave that given bundle to "assets_common" although
"assets_common" is not used on the frontend anymore as it would allow
to not care about migration of those automatically edit SCSS files and
it was also "not wrong" as you effectively cannot edit those files
without impacting all bundles.
It was although very wrong as editing those variables files via the SCSS
editor automatically considers them as part of "assets_frontend" and not
"assets_common". So editing them via the SCSS editor would create/update
the files relying on the fact the related bundle is "assets_frontend"
but the website builder would create/update them relying on the fact the
related bundle is "assets_common". This would lead to creating two
ir.asset records trying to replace the same file in the sub-bundle which
is used by both those assets... and thus to make the database crash.
The work done at [1] actually forgot about other things related to all
this. Those will of course be fixed but they are less urgent. This
commit here just fixes the SCSS edition issue described above for now.
[1]: https://github.com/odoo/odoo/commit/08dcbc8f1def06476d025b110049ee596e906816
task-2994071
closesodoo/odoo#101312
X-original-commit: 15e503a841d757e622c8ef2070043b766b28bfde
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
The goal of this commit is to:
- Tweak the website filter (on website pages list added at [1]) to make
it work for all website content records (page, blog, ...).
- Update the 'New Page' dialog to be able to select the website_id for
the new page.
- Tweak the '_compute_is_homepage()' method to set 'is_homepage = True'
on website's '/' page when 'homepage_url' is not set in settings.
[1]: https://github.com/odoo/odoo/commit/940f4ee875332dafa1f379970a7683be6b3ee606
task-2889981
X-original-commit: d6014c60acc4231a5e56d492d2a39deaf789cbe8
Part-of: odoo/odoo#101215
*: auth_password_policy, bus, calendar, event, mass_mailing, stock,
survey, web_editor, web_tour, website, website_event, website_forum,
website_mass_mailing, website_sale
This commit is a first step towards a potential deletion of the
assets_common bundle, although that step would need more work, specs and
discussions as some layouts kinda only use the assets_common bundle
(some take the full assets_common but parts of the assets_backend one
for example).
The main goal of this commit is to have the assets_frontend bundle
directly include the "common" files we need. As a first step, this
commit only blindly duplicates them all into assets_frontend (without
removing the potentially useless ones). The goal is to have those
advantages:
- Reaching a frontend page only calls two main JS files (one normal and
one lazy-loaded) instead of 4 (two normals and two lazy-loaded). This
may help reach a better google page speed (which is becoming more and
more strict).
- The frontend CSS is built as one: the common SCSS which was using
bootstrap variables, or even Odoo-based SCSS added by mistake in
common instead of both backend and frontend is now computed with the
right bootstrap customizations. E.g. the tempusdominus datetimepickers
use bootstrap grays... after this PR, they use the right grays as
customized by the user on the website.
It was also chosen to not have a common "sub-asset" which is included in
assets_frontend. Making assets_frontend completely independent makes
sense (as it probably will for other "main" asset bundles): we can focus
on adding the files each layout needs without the need of worrying if it
impacts unrelated layouts. Sub-assets (when not strictly necessary) is
also a source of errors: extending the "main" bundle instead of the
right sub-asset it may use (like it was the case with the sub-assets of
assets_common: _assets_common_scripts and _assets_common_styles as
explained in the previous commit). So this is indeed a small drawback of
not factorizing the code for the inclusion of "common" files in bundles
but it seems more explicit and easier to maintain that way. Note that
adding "common" file is not the most common usecase anyway, apps
generally only need files in backend or frontend.
The __manifest__ declaration will also likely evolve in more and more
uses of wildcards to match entire directories. In the future, adding
"common" web-app files in both backend, frontend and other "main"
bundles could just be about one line duplicated into each bundle.
closesodoo/odoo#100314
Related: odoo/enterprise#31394
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
Before this commit, when one would remove the favicon of a website in
the settings it would crash.
This is because since [1] the introduction of `base64.b64encode()` in
the `_handle_favicon()` method make it crash with `False` value.
Steps to reproduce:
- Settings > website
- Delete favicon
[1]: https://github.com/odoo/odoo/commit/6b8752604898bf2b583b7f5334e35f6a1583595e
opw-2958836
closesodoo/odoo#100399
X-original-commit: 35ce152d7e5c50cb363c0fa7adecf174fede9e46
Signed-off-by: Romain Derie (rde) <rde@odoo.com>
When creating an account from the website (through the link "Don't have an
account?") while having been previously identified, the creation lasts a very
long time and ends-up with a http 502 error (actually a timeout). Same when
logging although being already logged.
Technical note: The problem was due to a database deadlock because of going
twice through authentication. A record was deleted in one transaction while
being written in another (_merge_visitor deletes it and _update_visitor_last
visit writes it but with different environment).
Task-2950241
closesodoo/odoo#99723
Signed-off-by: Thibault Delavallee (tde) <tde@openerp.com>
Website: the context install_filename='dummy' is used to prevent
arch_updated from becoming True while updating translations of
ir_ui_view.arch_db (if arch_update becomes True, test_inherit_specific
fails)
Fuzzy search for jsonb translated fields has been adapted in the case of
website. It may require some refactoring later.
Translated fields no longer use the model ir.translation. Instead they store
all their values as JSON, and store them into JSONB columns in the model's
table. The field's column value is either NULL or a JSON dict mapping language
codes to text (the field's value in the corresponding language), and must
contain an entry for key 'en_US' (as it is used as a fallback for all other
languages). Empty text is allowed in translation values, but not NULL.
Here are examples for a field with translate=True:
NULL
{"en_US": "Foo"}
{"en_US": "Foo", "fr_FR": "Bar", "nl_NL": "Baz"}
{"en_US": "Foo", "fr_FR": "", "nl_NL": "Baz"}
Like before, writing False to the field makes it NULL, i.e., False in all
languages. However, writing "" to the field makes its value empty in the
current language, but does not discard the values in the other languages.
Here are examples for a field with translate=xml_translate:
NULL
{"en_US": "<div>Foo<p>Bar</p></div>", "fr_FR": "<div>Fou<p>Barre</p></div>"}
Change for callable(translate) fields: one can now write any value in any
language on such a field. The new value will be adapted in all languages, based
on the mapping of terms between languages in the old values. Basically the
structure of the value must remain the same in all languages, like before.
Reading a translated field is now both simpler and faster than the former
implementation. We fetch the value of the field in the current language by
coalescing its value with the 'en_US' value of the field:
SELECT id, COALESCE(name->>'fr_FR', name->>'en_US') AS name ...
The raw cache of the field contains either None or a dict which is conceptually
a subset of the JSON value in database (except for missing languages). For the
sake of simplicity, most cache operations deal with the dict and return the text
value in the current language.
Trigram indexes have been adapted to the new storing strategy, and should enable
to search in any language. Before this change, only the source value of the
field ('en_US') could be indexed.
Computed stored translated fields are not supported by the framework, because of
the complexity of the computation itself: the field would need to be computed in
all active languages. We chose to not provide any hook to compute a field in
all languages at once, and the framework always invokes a compute method once to
recompute it.
Code translations are no longer stored into the database. They become static,
and are extracted from the PO files when needed. The worker simply uses a cache
with extracted code translations for performance. This is reasonable, since
fr_FR code translations for all modules takes around 2MB of memory, and the
cache can be shared among all registries in the worker. Changing code
translations requires to update the corresponding PO file and reloading the
worker(s).
Performance summary:
(+) reading 'model' translated fields is faster
(+) reading 'model_terms' translated fields is much faster (no need to inject
translations into the source value)
(+) searching translated fields with operator 'ilike' is much faster when the
field is indexed with 'trigram'
(+) updating translated fields requires less ORM flushing
(-) importing translations from PO files is 2x slower
Some extra fixes:
- make field 'name' of ir.actions.actions translated; because of the PG
inheritance, this is necessary to make the column definition consistent in
all models that inherit from ir.actions.actions.
- add some backend API for the web/website client for editing translations
- move methods get_field_string() to model ir.model.fields
- move _load_module_terms to model ir.module.module
- adapt tests in test_impex, test_new_api
- because env.lang is injected into SQL queries, its returned value is
now guaranteed to correspond to a valid active language or None
- remove wizard to insert missing translations (no longer makes sense)
task-id: 2081307
Co-authored-by: Fabien Pinckaers <fp@openerp.com>
Co-authored-by: Raphael Collet <rco@odoo.com>
Short summary:
- People can now use the snippets option to show/hide based on country
- We never use / test / maintain this feature, if it still works since
its introduction 4 years ago that's by luck
- It is probably barely used, we never got a single question or issue
about it
-----------
When we introduced the multi-website feature, it came with 2 main use
cases:
1. Multi website by domain -> Every website has its own domain and based
on that we serve the expected website
2. Multi website by geoip -> Multiple website can have the same domain
and based on GEOIP/country we serve the expected website
We never really supported, highlighted or promoted the geoip case.
We actually never use it ourselves when doing tests, and we don't
consider it when doing specs / improvements.
At most, only a very few people know about it in Odoo.
That GEOIP case is probably not useful at all, as displaying a whole new
website based on lang is not easy since nothing can be shared out of the
box -> Every website has its own COW records.
For instance, one could think of having its main website A and another
website B on which he just added 3 products specific to that website B
country.
But such a case won't even work properly, as any adaption on website A
won't be reflected on website B (design, theme etc).
closesodoo/odoo#99938
Related: odoo/upgrade#3892
Signed-off-by: Quentin Smetz (qsm) <qsm@odoo.com>
Xavier-Do