This patch will always perform conversion against given encoding, in order to check if this codec is valid or not.

Check that declared encoding is supported by python before using it.

There is no reason not to support sql_catalog_raise_error_on_uid_check for both cases of UID inconsistency. UID change happens for example with inconsistent catalog and ZODB after disaster recovery. It can be useful to prevent raise during a double catalogin of the database after a disaster recovery. Else, it is just impossible to recover.

add new field for new coordinate_text property
improve proxyfication (from deprecated field library)

thanks to css class horizontal_align_form_box, sub fields will be displayed horizontally

increment version for 22baeab9c43113428ccf9472798732dc212d305f

Detailed properties take precedence on coordinate_text (raw string) if at least one of them is defined.
Introduce isDetailed method in coordinate API, which return True if at least one detailed property is defined.

Acquire more properties from coordinates

raw input is available with getData.
Backward compatibility is supported.

Associate it to all Coordinate Classes.
It defines data property to hold user input value

fromText was used to parse and store in different placeholders slices of coordinate strings.
This assumption is wrong. Parsing will be dedicated to asText renderer.

It was involving twisted conversion back and forth between
erp5.util.testbrowser which expected the URL and ERP5 site id whereas
erp5.util.benchmark which only expected the full URL.

Moreover, it was not working if there is no "real" ERP5 site ID, for example
with: http://foobar.org/.

After changing state make sure reindex happens.

Also, word-wrap and remove "captain obvious"-style comment.

This reverts commit e7e8f515.
only_group_columns doesn't belong to this API level, only to the level
immediately below.
Acked-by: Łukasz Nowak <luke@nexedi.com>

and check.

Just force only_group_columns=True in kw, as caller can already use this
parameter.

and check.

mapply needs to know which parameters it has to extract from the request.
that's what fill_args_from_request decorator does.

If Image is a Text, then conversion to image format is performed

This fix is required because of a lack of conversion server ability to perform
successive transformations.
Add a comment in docstring to address explicitely this issue.

display and quality parameters are not handled explicitely by OOoDocument._convert
as they are just passed as they are to Image._convert .

Add a test to prove the fix

If user omit format means he asked for original data
Guess format from mimetypes_registry later (if any) only to add attachment inside odf archive

progresively slower with data grow.