Commit a2ae219d4e fixed the customize_template_get
function(s) which had been modified for the 10.0 web_studio release.
While the fix also reorganized the structure to make it more efficient
and understandable, it promised that the function would be refactored
in master. This is what this commit does.
The customize_template_get route of the website module was also removed
with commit 5a030db3eb as it was an useless
duplicate of the web_editor customize_template_get route.
The old route handled two totally different features of odoo: getting
all the page-related views to allow their edition with the HTML editor
and getting the "toggable views" which appear in the "Customize"
website menu. The parameter "full" allowed to determine which one the
route was supposed to do and was useless until website was installed to
add its feature (by extending the route behavior). In both cases, views
informations are returned, only the number of views differs. The
above-mentionned fix already did part of the refactoring by returning
the same info structure for each view in both cases.
Now, the function has been split. The web_editor customize_template_get
route is renamed as get_assets_editor_resources and returns the info the
HTML editor needs to work (the "assets" term is in prevision of the
new LESS editor).
A new website route "get_switchable_related_views" has been created to
return the info the "Customize" website menu needs.
Both use the renamed ir.ui.view function "get_related_views" (old
ir.ui.view customize_template_get function) which now returns a
recordset and not a list of dictionnaries.
Specification
=============
- no boxes around fields
- use regular form style for input fields
- don't split page/URL into two fields, just do one URL field, with autocomplete
- Add alpha, beta, gamma, delta and epsilon color in color selection because they are used in mass mailing and themes also.
- Add a selection option with label "L" inside circle for link.
- Add Color selector for choosing button color
- Remov select2 page selector and used jquery autocomplete on same input
- Fix issue of page search in ir.http
The GeoIP resolver keeps a permanent open file descriptor on the
GeoIP database file, for example /usr/share/GeoIP/GeoIPCity.dat
Because we initialize a resolver for each registry, servers with
many databases may easily blow up the open file descriptor limit
of multi-threaded workers (gevent or no workers multithread mode).
The Maxmind GeoIP API appears to be thread-safe since v1.1.4, so
it is much less wasteful to use a shared resolver for each process.
This patch keeps the _geoip_resolver class attribute on `ir.http`
for compatibility, but it will be removed for the next version.
The forward-port of the above revision missed
the conversion to the new api and that
`context` had to be replaced by `self.env.context`
Fixes#14922Closes#14923
The revision
8f48baeda6
broke the ability to search pages in the dialog
allowing to add a menu to the website from the website
(From the website, Content > Edit menu)
as `request.website` was not defined in this case
opw-702020
The website `default_lang_id` was not required,
while the algorithm choosing the website language
expects it to be required.
This solves the below issue, when your browser language is en_US:
- add a second language to the website (e.g. fr_BE or es_VE),
- in the website settings, remove English from the languages
- in the website settings, unset the default language
- try to go on the website
You will have a crash:
```
File "/home/odoo/src/odoo/saas-11/addons/website/models/ir_http.py", line 193, in _dispatch
resp.set_cookie('website_lang', request.lang)
File "/usr/lib/python2.7/dist-packages/werkzeug/wrappers.py", line 992, in set_cookie
self.charset))
File "/usr/lib/python2.7/dist-packages/werkzeug/http.py", line 905, in dump_cookie
value = to_bytes(value, charset)
File "/usr/lib/python2.7/dist-packages/werkzeug/_compat.py", line 106, in to_bytes
raise TypeError('Expected bytes')
TypeError: Expected bytes
```
This is because the `request.lang` is set to `False`
because of the below algorithm in `ir_http.py`:
```
nearest_lang = not func and self.get_nearest_lang(path[1])
url_lang = nearest_lang and path[1]
preferred_lang = ((cook_lang if cook_lang in langs else False)
or self.get_nearest_lang(request.lang)
or request.website.default_lang_code)
is_a_bot = self.is_a_bot()
request.lang = request.context['lang'] = nearest_lang or preferred_lang
```
`nearest_lang` is `False`, because there is no nearest language
available for the browser lang (`en_US`)
`cook_lang` is `False` for users who never went on the website
`request.website.default_lang_code`, which is the last feedback,
is also `False` because you removed the default language
in the website settings.
opw-695621
Previous implementation did not allow the user to search for pages 'name'
but only on the slugified name.
Now we slugify the needles before to find a match.
1. Create a new page: 'The new'
2. add a link on your website and try to find this page...
- Before this commit, the only ways was to type 'the-new'
- After this commit 'the new', 'The new', 'The-new', ... will match
This commit closes#10771
Currently when using a language on the website, the following is done in
this order to choose which one will be used:
1) use a language code if present in the url (eg. /fr_BE/),
2) use a language similar to the one in the url,
3) use a previously choosed language (saved in cookie),
4) use the browser language if available on website,
5) use a language similar to the browser language,
6) use the default website language
When a bot is requesting a page, he could either not specify a browser
language (en_US is set as default) or sometimes specify english
erroneously.
In these instance, we will then never use the default website language
if an english language is installed.
This fix changes this behaviour so if a bot is visiting a page, the 4th
and 5th steps are skipped and the default website language is used
instead of using the browser language.
closes#14033
opw-690910
Commit e11f23b701 split the website
customize_template_get into web_editor and website ones. This split
was not perfect and commit 20861b482a
tried to make it work... but did not made it totally and added an
indirect dependency to website module in web_editor...
As the final result is wrong and inconsistent and that the previous
9.0 implementation was already ugly, this was not possible to make
a consistent full-stable fix.
This commit tries to make a fix as stable as possible (the main method
name did not change, the arguments did not change, their purpose is
the same, ...) which makes the code work and be less ugly. However,
the function returns are now slightly different (and JS depends on it).
It returns only valid view information (not header stuff which is not
related to anything and the inherit_id value is now a combination of
the inherit view id and name). Also, a new private method has been
created.
Also, the method contained code which was deprecated since saas-6 but
was never removed (and never broke anything "by chance").
In master these methods will be refactored to make more sense.
The commit e59a7fe3ff has moved
the ACE editor from website to web_editor. As some code on ir_ui_view
was used for this widget, the code has also been moved to web_editor.
Several modules defines records with the external ID `base.foo_bar` while it is
created inside this module (typically menus and groups).
While there is no technical reasons to do so but this may introduce issues:
- these records will not be deleted during uninstall
- if a language is loaded before the installation of the module, it won't be
translated
The uninstallation will only remove the records with an external id linked to
this module (these would only be removed when removing base).
Installing a language before the module will drop the translations not linked
to an existing external id (as it can not be resolved).
This commit correct all the external ids tagged as from base or other incorrect
modules.
qsm-odoo