Replace `httprequest.stream.read()` by `httprequest.get_data()` so the
payload content remains available: get_data stores the request body (by
default) so it remains available for alternative processing or checks
(MAC checks for webhook validations for instance). With `stream.read()`,
once the data is read if it's not stored separately it is lost.
d28f8f704f carefully clears only the
relevant bit of the cache (only the top-level object which we're
exporting), however the commit did not considert that _export_rows
calls itself recursively for relational fields, and thus when
exporting relational fields the relation would become the new
top-level and get cache-cleared on every iteration, significantly
slowing down cases where such records are shared and would need to be
re-fetched every iteration, even more so if they're somewhat/somehow
expensive.
OPW-1833545
OPW-1836361
While c059e7b93c64b9dd9bc96eeb6ee5cadf significantly lowers the cost
of generating xids, a safety invalidate cache can regress cases
significantly e.g. 18s -> 430s (7mn) for some recursive exports of
xids as the cache would be completely cleared *after each record*
requiring complete fetching & recomputation of the following records.
Fix by only clearing the cache if we've actually had to create
xids *and* only for ir.model.data records.
It's possible that clearing the cache isn't even necessary at all (?),
check with @rco-odoo
OPW-1835226
Purpose
=======
The method compute is annoying for the following reasons:
- 'compute' is difficult to grep
- Developers don't understand that they can specify the company and
the date on which we want to convert the amount. This is quite
error prone because if these parameters aren't specified, they are
set by default to the company of the current user and the current
date (today).
- Passing parameters is the context is not a correct way to call a method.
Furthermore, using with_context is not really good for the performance too,
as the record as to be rebrowsed, the computed fields have to be
re-evaluated
Specification
=============
- Introduce a new method '_convert'. This method takes 2 more
parameters: 'company' and 'date'.
- Deprecate 'compute' and '_compute'. Log a warning to notify that we
should use '_convert'.
In 79038055 amount_to_text was refactored and improved, but there is a
drawback.
ie. if we have .29 cents with two decimals, `Twenty-Eight` was displayed
because `int(0.29 * 100) == 28` due to floating-point representation.
This commit keeps the improvements and fix this issue by using string
formatting of float as was done before.
opw-1829924
closes#24161
The goal of this commit is to allow fast entry of the same model using
the keyboard for navigating through the form once, by blocking the user
from advancing in the form if there is a required field or once they
went through the form once to get them to a primary action (button)
This commit includes the following changes
1) Enable moving forward from field to field using the TAB key
2) Entering the one to many and many to many using the TAB key
a. When entering it, set the focus on the "add new line" link or
button
b. When adding a new line, set the focus on the field visible
editable field of the new line
c. Discard adding a new line with ESC key
d. If the user leaves the first field empty and uses TAB, we
will cancel the adding of a new line and move to the next
field of the form
3) When a field is required and not filled in, do not allow the user
to move out of the field using the TAB key (the user is still
allowed to use the mouse though), mark the field as invalid
instead
4) After going though the form once, using the TAB key on the last
field will move the focus to the first primary button of the page
5) When the focus is on a primary button (EDIT/SAVE), the user cannot
move the focus forward using the TAB key. Hitting TAB again will
display a tooltip telling to hit ENTER to activate the button.
The mouse is still available to move the focus.
6) When the user saves, the focus is placed on the first primary
button of the form renderer (like VALIDATE for a new invoice)
7) On dialogs, primary buttons should stop the users from moving out
of them, and showing a popup if the user tries
8) When closing a dialog, the focus will be moved back to the widget
that opened it.
This commit does not include the following features
1) Navigation with the keyboard on a selection one to man
2) Navigate between tabs in a form using the keyboard
3) Cancelling the adding of a new line in a many to many using the
ESC key do not set the focus correctly
4) Enhancing the focusses fields (like blue underline)
Since Python 3.2, close_fds parameter is True by default on POSIX
platforms. On Windows, it's set to False when one of the standard descriptors
is set.
This commit will give the opportunity to run phantom js tests on
Windows.
Purpose of this commit is to avoid errors for some performance tests that
have a slightly variable number of queries. Indeed some tests in test_mail
module vary of 1 or 2 queries due to a behavior still not found that does
not seem deterministic.
This should not be considered as a real error, but should not be considered
as completely valid. This is why a margin is accepted in some tests that
will raise a warning instead of an error. This warning tells people to
check if they have increased the query counter or if it is due to a random
effect in query count.
This is a manual forward-port of saas-11.2 commit 3fe0ee5ae6 .
Purpose of this commit is to avoid errors for some performance tests that
have a slightly variable number of queries. Indeed some tests in test_mail
module vary of 1 or 2 queries due to a behavior still not found that does
not seem deterministic.
This should not be considered as a real error, but should not be considered
as completely valid. This is why a margin is accepted in some tests that
will raise a warning instead of an error. This warning tells people to
check if they have increased the query counter or if it is due to a random
effect in query count.
This reverts commit d565596868. Indeed whole purpose of having tests
based on query counters is to fail every time query count is greater than
expected. Query count could be greater than expected because
* either it is required by a bug fix or a feature and tests should be
updated accordingly in the same commit;
* either implementation can be improved to avoid or lessen query count
impact and code should be updated accordingly;
* either we have some random queries due to nondeterministic behavior;
It seems purpose of the commit was to solve the third case. However we think
hiding first two cases can be dangerous compared to solving third case.
Moreover having a threshold of 10% does not make sense. It means some complex
test case could take 30 queries more without being noticed. Some tests could
fail after several commits adding queries and finding them afterwards could
be complicated.
Another solution based on flagging some tests that have a small random
behavior will come soon.
Install a database in some other language than en_US
In debug mode and in the app list viewer,
you'll see that the apps field are translated into your language
Then do an "Update App List"
Reload your page
Before this commit, the Apps' field lost their translation
This was because when checking for any updated data for known modules,
the ORM fetched the translation automatically, and compared the old (translated)
value with the new (English) value from the manifest
After this commit, This operation doesn't make the translations to be lost,
because only english values are compared
OPW 1832796
closes#23910closes#24091
Since report refactoring the function to override is get_report_values
that only returns a dict of value that are used in rendering.
In order to render a pdf with a landscape orientation, a context key could
be added however since it's not possible to return a report object in
get_report_values, it's difficult to add a contexti without using some tricks.
This commit add a report option to use in the template that will be used
in order to modify the pdf orientation. The process used was already existing
for other options like dpi,...
self is a rml_parse instance, not a BaseModel
create new environment to call ORM methods
prevent crash when calling formatLang on a RML report
Closes#24068
Before this rev:
* Take 2 domains that contain at least two leaves, and at least one of
them must use the implicit `&` operator
e.g.:
d1 = [('so_line', 'in', [91]), ('amount', '<=', 0.0)]
d2 = ['&', ('so_line', 'in', []), ('project_id', '!=', False)]
* Perform osv.expression.OR() between both domains
Expected result:
d3 = ['|', d1, d2]
Actual result (after normalization):
d3 = ['&', '|', d1, d2]
This is because, since the `&` is implicit for the first domain, when we
OR it, we give it an explicit `|` operator, so when we pass this domain
through the normalize_domain function, d1 no longer contains an implicit
`&` operator but instead the implicit operator is the one between d1 and
d2, therefore giving us a completely wrong domain.
The `combine` function states that it only accepts normalized domains,
however neither the OR nor AND functions do, this leads to a lot of
developers putting non-normalized domains into these functions, and
there's no error checking or anything that obviously indicates that the
domain is incorrect, so we might as well normalize all domains being
passed since it's already pretty optimized.
Before this patch the registry and cache signaling was only activated
for PreforkServer. In case Odoo was deployed in a multi process/multi
threaded architecture the signaling was not ensured, causing registry
de-synchronisation amongst threaded servers.
This fix will be backported in v10.0 and v11.0 as soon as it has been
battle tested.
- Install Website
- Load the 'Norwegian Bokmål' translation, and choose to translate the
website
- Logout
When accessing the website, the language displayed is not consistent to
the browser language.
1. `no` shows homepage in `nb_NO` (Firefox only) => expected since
Bokmål is the main language in Norway
2. `nb` shows homepage in `en_US` => unexpected
3. `nn` shows homepage in `en_US` => expected since 'Norwegian Nynorsk'
is not a language available (this will be the topic of another PR)
The issues comes from Babel's side, since `nb` is not in the
`LOCALE_ALIASES` while `no` is. We monkey-patch the value to avoid this
while Babel is corrected.
opw-1827258
* add UUIDs to XID sections (4 bytes / 8 hex digits) so it's not
necessary to handle collisions & add a fallback generator
* use COPY for performances over executemany: executemany just
performs an implicit loop on all statements (in psycopg2)
* ~pure SQL was possible:
INSERT INTO ir_model_data (module, model, name, res_id)
SELECT
'__export__',
'{model}',
'{table}_' || A.id || '_' || uuid_generate_v4(),
A.id
FROM {table} A
LEFT JOIN ir_model_data B
ON A.id = B.res_id AND B.model = %s
WHERE B.res_id IS NULL;
but would have required installing pg extensions (problematic
especially in stable) and performances are about the same as
the COPY version
task 36343
Fixes#22493
* don't use FS to do export
* update translations export dialog to fix outdated links (launchpad
& 6.1 translations doc)
Note: can't use ``TarFile.gettarinfo`` to create the TarInfo object
because it wants an actual file object with a ``name`` attribute and
possibly even an fd.
closes#22798
When saving the settings, we perform a `write` on the groups and implied
groups. This will trigger a recomputation of the fields `share` on
`res.users` and `partner_share` on `res.partner`.
On a database with many users, this can lead to a non-negligible time
spent in these methods.
opw-1830145
In the normal course of saving-action, readonly fields are stripped
out when sending data to the server. The SSF would reuse the same code
when "saving" O2M lines to the parent record however that is not
correct and would lead to misbehaving views as readonly (non-stored)
fields used as "transients" (storing data within the extent of a
record's edition session) would not behave properly.
Fix by overriding the O2M Form's save so that readonly fields are
kept and stored in the parent record.
Closes#23620
The caller of `read_group` can now provide the aggregating function to use for
a given field:
# set aggregating operator for fields 'foo' and 'bar'
model.read_group(domain, ['foo:sum', 'bar:avg'], ...)
One can also aggregate the same field several times, by giving a specific
output name for each:
# aggregate 'foo' with both 'min' and 'max'
model.read_group(domain, ['foomin:min(foo)', 'foomax:max(foo)'], ...)
The commit f1e842bcd2 corrected the Khmer code language but the csv language file was left with a blank line inside.
The csv parsing (to populate the list of available languages) failed into tools.misc.scan_languages() because of the blank line
The code of khmer language was not the offical one. Installing it
produce an error in the Babel python library.
It's not possible to easily fix it by correcting the code km_KM to km_KH
because Odoo
prohibit changing language code. The solution to this is to remove the
Khmer language from the res_lang.csv file and add it to an xml data
file with the correct code. There, we can put on this record the
"noupdate" tag. This tag
will block any modification on the ir_model_data corresponding to the
xml_id=lang_km
and so avoid the error if we update the base module. Old database still
wont be able
to use khmer language (so we'll need to manually change it in database)
but new one created
after this commit will be able to use it without issue.
Size attributes (e.g. `max-width`) are sometimes set in css (i.e. with classes,
see `oe_avatar`).
They must however be overriden if they are specified on the widget.
Example:
<field name="image" widget="image" class="oe_avatar" options="{'size': [180, 180]}"/>
The `max-width: 90px` set on oe_avatar must be overriden in this case.
Note that the attribute `img_width` and `img_height` have been depreciated as
`width` and `height` are fulfilling the exact same purpose.
Christophe Simonis