/!\ Partial backport of https://github.com/odoo/odoo/commit/0a31cb9b4ae461119ecafdd05c5b979621ffb039
/!\ Should not be forward-ported
Before the websitepocalypse introduced in 11.0 (saas-17), when saving a
page in the editor, every `cleanForSave` method of snippet options were
called, even for snippet options which were not initialized (so if a
snippet was left unchanged before saving, the related `cleanForSave`
methods were called anyway). This behavior was bad code and did not make
much sense and was so removed with the websitepocalypse
(see https://github.com/odoo/odoo/commit/2972976962617d4b8a0113bae58c640ab41cdff8#diff-4c580abda3220e45c6b3f6bdcf77c733L415)
Unfortunatly, the behavior was necessary given the states of our snippet
options. Indeed, for the "latest posts" snippet, the `cleanForSave`
method removes the entire snippet content (as it is dynamically loaded).
The dynamic loading is performed by the snippet *animation*... and
removing the dynamic content should logically be done by the animation
too... which is the case. The problem is that animations are not stopped
before the editor is saved. This behavior was only implemented in
saas-11.1 (for future version 12.0) with commit https://github.com/odoo/odoo/commit/0a31cb9b4ae461119ecafdd05c5b979621ffb039
As a stable fix for 11.0, part of the mentioned commit is backported
here and some specific `cleanForSave` will be reviewed as animations if
necessary. Note: `cleanForSave` made useless by this commit are left
anyway as this is a stable fix.
See https://github.com/odoo/odoo/issues/22089
Before this fix, the chatter always showed dates of messages in military time format.
This was not the expected behavior, as the user could have had customize
the date and time format in the settings of the current language.
Steps to reproduce the issue:
1. Set Odoo in English
2. Settings > Translations > Languages > English
2.1. Set Date Format to %m/%d/%Y
2.2. Set Time Format to %I:%M:%S %p
3. CRM > (Select any opportunity with at least one message) > Mouse hover on elapsed time (e.g. "an hour ago")
Expected:
- It shows the customized time format (e.g. 11/07/2017 5:33:36 PM)
Results:
- It shows the military time format (e.g. Tue Nov 07 2017 17:33:36 GMT+0100)
opw-781647
When a snippet is dropped into a page, all its animations are started.
The problem was that all its children's animations have to be started
and that was not the case anymore since recent big website refactoring.
Fortunately, this allows to improve and optimize animations start
function: now all animations are (re)started for a portion of a
given DOM (default to #wrapwrap) while it was only possible to
restart animation of a given DOM *element* before.
Before this commit, when an user added a video in his page (an <iframe>)
it was then duplicated each time the snippet in which the video is was
edited and saved again, slowing the page loading.
As the code responsible for that is quite outdated, it should be
reviewed and refactored in master. This commit only fixes the problem
with more code. This commit will also fix existing databases, if the
user edit the video snippet once again.
(Note: even if this commit's code will be even more strange in master,
it *should* be forward-ported anyway).
Before this commit, clicking on 'Go to page' in a page properties dialog would
redirect user to an incorrect link (relative vs absolute) because the link
would not have a leading slash. (We removed it earlier to hide it in the input)
Now, the link is correct
qsm:
This is because of multiple animations during edition (which is not
occuring afterwards). This is not supported, but as it does not break
anything, it can be kept as it is. However, we cannot access the
animation via 'snippet-view' anymore, which is fine as I am about to
deprecate it in master.
Before the website refactoring, snippet animations were cleaned thanks
to the `stop` method. As those animations are now extensions of the
standard `Widget` class, the `stop` method was replaced by the standard
`destroy` method. This was tweaked so that the $el is not removed from
the DOM, but with proper cleaning of events. However, this changed the
way the `stop` method content has to be called. Indeed, we should be
allowed to call the `_super` method and then use the $el variable.
Before this commit, you was able to double click on the button when
are in registration flow. In this case, you subscribe 2 times and so
take 2x more seats, what can be annoying when you have a limited room.
In the same time, we fix the form in the form that generate
strange behaviour like some events not bubbled correctly.
The attendee form (into the modal) was inside the registration form.
$'attendee_form).on('submit') obviously failed due to this bad dom.
This commit closes opw-778191
Before this commit, if you had a simple link (just a href, not handled
in JS), you couldn't simply make it clickable only once.
This could be a problem in some cases. (eg: a btn creating a record, it would
create as many record as many times you clicked on it)
The same behavior could occur for link/button redirecting to a route creating a
record. There is a window of time between the click on the btn and the moment
the page is getting redirected.
You have time to click multiple time on it and it would call the URL multiple
time.
Now, there is a class 'js_disable_on_click' that will be handled in
website_root.js and will disabled the button after the first click.
The fix was originally aimed to fix multiple click on "Create Page" btn on 404
The idea is to warn user if he created a special file (website.page with
supported extension) and he wants to change the file's name, which will at the
end also change the key).
If this special file has been t-called elsewhere, they t-call wont work anymore
This is why warning user is needed.
Note that this is different than the already existing page_search_dependencies
that is just checking if the URL is present in menus, templates, blog or pages.
This is about checking if the key is present somewhere. (we assume that if the
key is present, it is surely about a t-call.
Technically, the flow is:
1. Get supported extension
2. Get key dependencies
3. Check if user modify name
4. If name changed and it is/was a supported extension (special file) and it
has dependencies (surely t-call) then warn user these call wont work anymore
--
search_page_dependencies is also improved by returning links to be clickable
(on delete dialog eg) and splitting template/page results.
it also refactor dependencies warning that was creating HTML in javascript into
using XML template
The deprecated-but-still-used website.prompt requires some static XML
to be loaded to work. In the case of creating a new website page, this
XML was implicitely loaded by the "Content" menu... but when the user
is on a translated page, this menu is not there. Anyway, this static
XML loading should be done by the website.prompt function itself if
necessary.
Before 11.0, website 'animations' were able to be started on any DOM
element. Since 11.0, this was reduced to any DOM element inside the
main website element (#wrapwrap). This commit allows this element
to be attached a website 'animation' too.
Before this commit, error thrown by scripts injected in DOM won't return a
clean error on window.onerror.
It would open the crash_manager's show_error's dialog that wouldn't display
any informations except "Script Error." and "Traceback:".
This is because of browser's security reason: it will just give you a
'Script error.' as message and nothing else for security issue.
Now, this kind of error will show a clearer message informing that it is
probably originating from a Javascript file served from a different origin.
There is now also the possibility for developers to handle/override the
behavior when this kind of error is catch to prevent the dialog to open for
example.
Might be of use:
https://blog.sentry.io/2016/05/17/what-is-script-error.htmlhttps://stackoverflow.com/a/7778424https://danlimerick.wordpress.com/2014/01/18/how-to-catch-javascript-errors-with-window-onerror-even-on-chrome-and-firefox/ (Search for Script Error)
Before this commit:
1. you could not edit a menu URL because the input was hidden.
This was due to radio options being switched (new page is now selected choice)
when you create a menu.
But when you edit, there is no choice and the input should be shown.
2. if you set a menu url to an email address, the display link would be prefixed
by a / and so it would open mysite.com/mailto:XXX instead of just the mailto
Now:
1. Input is now hidden only when you add a new menu, not when editing a menu.
2. mailto link now works as expected: it will open your desktop mail software
Seen behavior: SEO keywords not saving
Expecte behavior: SEO should save
Reason: destroy event method from dialog.js is now hidding modal
programmatically. Which at the end call twice the keywords save
method.
The second time there were no keywords passed to that function
which is causing that behavior (saving empty array).
- labelling and screen redesign (update labels, hide labels, hide fields, ...)
- make UI more mobile friendly (icon biggest, color using less variables,...)
- Page & Redirection menu are for visible only in debug mode
- Add meta noindex for website.page not shown in sitemap
- Add mechanism to be able to edit in backend a broken view
(Needed for - Internal Error 500, css, js file, ...)
- website pagemanager dedicated to website_publisher only
Purpose: Add button to see the number of
abandoned carts and the ability to send
a recovery email to those customers.
On the abandoned carts menu, you can see
the SO from the website where no reminder
email has been sent.
By selecting them, you can send an email
to the customer to tell him there is still
a cart waiting.
Clicking on the link, he can reuse his old
cart or merge it with his current one.
An abandoned cart is a SO from a website
sales team, and created after a treshold
parametered by a 'delay' settings (number
of hours to be considered as abandoned).
website.page = old ir.ui.view with page=True
website.redirect is a new mechanism to replace in the futur the ir.attachment
mechanism of redirect.
From now, we don't have a specific /page controller to serve 'page'.
We use a new model website.page which is rendered if none route matches the url
and that the field 'url' on website.page matches the request.httprequest.path.
The order to serve a path is:
- Routes defines in controllers (/shop, /blog, ...)
- ir.attachment with name matching the path
- website.page with url matching the path
- website.redirect with url_from matching the path
- 404
To improve:
- allow regexp in website.redirect model
- allow to edit the view_arch from the page.management via redirect backend
(needed when traceback in the page, or when modifying a js/css/less/...)
The commit 8749410b10 corrected the way the iframe was rendered to display a video in v9.0.
In v10 though, the html sanitizer is a bit more severe and cleans data-src attributes, preventing the iframe to have a src
This commit works around that and allows the src of the video to be saved in db
OPW 761424
closes#19125
qsm-odoo