When exporting to a tarfile, we first export each module's translations
to a file on the FS, then create a tarball from all of them.
With the P3 bytes/string split, translation export was defined to export
to a bytes IO object in order to work correctly with CSV (which work on
bytes IO for better cross-version compatibility). However the tgz file
export was left opening module's files in `w` mode, which in P3 means
text, which blows up as POFile expects to ultimately write to a bytes IO
object: it uses text, and a StreamReaderWriter is used to translate that
into an underlying bytes IO.
Fix by explicitly opening the files in binary mode.
Fixes#22700
The result of the following:
```
float_round(2.5, precision_rounding=0.05, rounding_method='DOWN')
```
is 2.45. Instead, 2.5 is expected.
In case of the `DOWN` method, epsilon should be added, not subtracted.
In the meantime, we introduce a slight reorganization of the method to
make it clearer.
opw-805008
Fixes#20046
Simply reverting e5ef1c5917 looks to
work *but* since I didn't note the repro case for the issue and don't
remember what it was it may re-introduce previous breakage. Instead,
ensure all "sizes" for the image are reset if any of the sizes is
present *and* Falsy by adding a final case to the conditional.
If any of the image fields is present and non-falsy one of the
previous branches will be taken, so we can just check that any of the
image fields is present at all (this implies it's falsy).
- Install French language
- Set language of Customer A to French
- Create an invoice due in the past
- Open the Customer Statements
The customer statement of A is translated in French, but the amounts are
formatted in the company's user language (English by default).
We add the possiblity to take into account the lang in the context when
calling `formatLang`.
Before this commit, if comments were found inside the body of a mail
template sent through mass_mailing, the regex in append_content_to_html
that converts possible uppercase tags to lowercase tags would replace
comments by broken html elements, this would break the in-browser parser
and thus would show partial content or no content at all, depending on
the positioning of the comments.
i.e. from <!-- This is a comment --> to <this -->
The regex in question is r'(</?)\W*(\w+)([ >])' which basically matches
opening and closing tags (this was a bad idea from the beginning btw),
the \W* will match any non-alphanumeric chars (so !-- in the
case of comments) and this would be completely discarded since it's not
a match group and when replacing we only use match groups.
This is fixed by simply stripping the \W* since it's discarded anyway,
an alternative fix would be to put it in a match group and add it when
replacing as well, but I don't know if we purposefully discard it or
not.
Neither fix is the good one, the good fix is to replace all the re.subs
by lxml which is much easier and more concise.
Following the 3.6 wordcode change, EXTENDED_ARGS gets used for jumps
offset by more than 256 bytes (128 opcodes), most commonly
absolute (so many script of more than 128 opcodes with conditionals
arer going to trigger the issue).
In bytecoded Python versions, EXTENDED_ARGS was only necessary above
65k opcodes: the opcode would get a separate 16-bits argument, while
wordcode opcodes get 8-bit intrinsic.
With Odoo v11 running on Python3:
1. Go to Settings -> Translation -> Import Translation
2. Choose a valid .po file (e.g. generated with the menu Export Translation)
3. Click Import
There is a server error.
There are two bugs behind this error:
1. we open a temporary file that cannot insert bytes ('w+')
2. the name of the temporary file is an integer (its file descriptor)
Python2 processed some implicit conversion between str and bytes, so that the 1st error doesn't occur. The name of a temporary file in Python2 is '<fopen>'.
We made the following corrections to the code:
1. open the temporary file with w+b access (write bytes)
2. enhanced a conditional statement so that we skip instructions in case where the field name of the tempfile is a string
Related Ticket:
opw-775445
The class PoFile was already using it.
TextIOWrapper closes its underlying buffer causing the "Synchronize Terms" action
to fail (trying to read on a closed buffer).
Fixes#19911
- The `--no-database-list` option will now also block access to database
management functions and screens.
Presumably this flag should only be used in production when all
databases have been provisioned, so the admin should like to block
access to the db manager at the same time.
- If no `--database` or `-d` parameter is provided, the system will be
unable to fetch a list of databases at all, so users will be blocked
with an error message.
- Hide the link on the login screen to the DB manager when it is
disabled, to prevent sending users to an error page.
- Weak attempt at updating the documentation
Note: the security check for RPC methods could have been done in the RPC
dispatcher, however that would not have protected service methods when
called directly, e.g. by a controller (e.g. the dump method).
- Add support for hashed master passwords (super-admin password) using a
strong scheme (PBKDF2_SHA512).
- Replace the password with a hash in memory (tools.config map), after
verifying it
- Automatically replace the plaintext master password with a hash when
saving it after a password change
- Preserve support for setting/using plaintext passwords when necessary
(e.g. as a temporary deployment thing)
The 'xmlrpc'-based configuration parameters have been a misnomer since
the introduction of the generic HTTP service, years ago.
Hide these options from the server parameters, and replace them with
more appropriate 'http' ones:
--xmlrpc-interface -> --http-interface
--xmlrpc-port -> --http-port
--no-xmlrpc -> --no-http
The config entries for these have been adapted as well.
The old parameter names are still silently supported in both
command-line arguments and config files. However they are
stored with the new names in the `tools.config` dict,
and when saving config files (with the -s option).
Also clarified and cleaned up the descriptions of the HTTP/WEB server
parameters.
And finally, added a short version `-p`, for the `--http-port` option.
Credits to @dreispt for this (via #19518)
Closes#19518Closes#19778
Reading them as bytes would not crash but could fill the corresponding
field with pretty much garbage. Turns out Html fields inherit from
Text and XML content generally goes into text fields, so handling
these items as text makes much more sense than as binary data.
By using the thread's repr() we get its type, name, daemon status and
ID, all at one time. No need to include them separately.
Fall back to the threadId in case the repr() is empty, which can
happen for exotic threads (the gevent main thread seems to be like that)
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/...)
Split big universal RNG into view-specific schemas and moved rng based view validation
into view_validation.py file.
Adapt old views, and remove deprecated tag `version="7.0"` from views.
Fields sent from the client are pretty much always *text*. This can be
problematic for image fields sent as base64 *text* as in Python 3 the
base64 routines only work between bytes.
Alter the core field-resizing function (image_get_resized_images) such
that it ensures text base64 sources are encoded back to bytes before
being forwarded to the main resizing routines (which I don't think
should deal with *that* issue).
Fixes#18985: error when trying to update a company's logo
Purpose
=======
It makes no sense to crop an image from a ration. We don't say "I don't know my image size, but divide it by 2"
Specification
=============
Pass a maximum size argument instead of a thumbnail ratio
xmo-odoo