The old-api model._all_columns contains information about model._columns and
inherited columns. This dictionary is missing new-api computed non-stored
fields, and the new field objects provide a more readable api...
This commit contains the following changes:
- adapt several methods of BaseModel to use fields instead of columns and
_all_columns
- copy all semantic-free attributes of related fields from their source
- add attribute 'group_operator' on integer and float fields
- base, base_action_rule, crm, edi, hr, mail, mass_mailing, pad,
payment_acquirer, share, website, website_crm, website_mail: simply use
_fields instead of _all_columns
- base, decimal_precision, website: adapt qweb rendering methods to use fields
instead of columns
Clarify the semantics of field attributes:
- field.store is True when the field is actually stored in the database;
- field.column is the column corresponding to field or None.
The various field definitions correspond to:
- new-style stored field: field.store and field.column
- new-style non-stored field: not field.store and not field.column
- old-style regular field: field.store and field.column
- old-style function field: not field.store and field.column
This is a memory optimization: instead of setting all attributes on all
instances, set them with their default value on the class, and only set
specific ones on instances. This reduces the memory footprint of around 14Mb
per registry with modules crm, sale, purchase and stock installed.
This is a memory optimization: it reduces the memory footprint of each
registry. We have observed a reduction of 10Mb on a database with modules crm,
sale, purchase, stock.
This fixes issue #2146. The inverse of a one2many field can be an inherited
field (_inherits). In that case, we cannot read its value with a simple
database query. Instead, we let the related field read it, but for performance
considerations we disable the prefetching of other fields.
When reading a one2many field, the inverse mapping of the lines (matching m2o -> lines of corresponding record) was instantiating each line and then triggering the prefect of fields.
To improve the performances, the inverse mapping is done in sql to avoid triggering the prefetching.
When the context contains 'recompute': False, the recomputation was not even
prepared. Now both create() and write() prepare the recomputation by invoking
method modified(). The flag only controls whether method recompute() is invoked.
In addintion, the former flag 'no_store_function' was converted to the flag
'recompute', so that both create() and write() use the same flag.
Fixes#1456
Due to the multi-company record rule on gamification.goal,
each access to the Goals menu and each opening of the
Messaging menu (thus calling get_serialised_gamification_summary())
is extremely slow (with several thousands goals/users).
Adding auto_join to the user_id FK on goals makes it much
faster. However it causes crashes when reading the table
because the _order of gamification.goal uses `create_date`,
which becomes ambiguous after the auto_join with res_users.
Solving this can be done by re-implementing _read_flat()
in the ORM using the internal Query object, as in search(),
which takes care of fully-qualifying all column names.
Until this is fixed, a simple workaround is to use
start_date in the _order instead of collision-prone `create_date`.
In combination with f28be81, this should help speed up
initialization of new boolean columns. psycopg2 handles
bool parameters values just fine inside cr.mogrify()
If any missing or partially incorrect values cause
an exception other than a psycopg2 error, we should
still catch it, rollback that record and report
the error, rather than letting bubble and fail
without any feedback to the user.
Fixes#1485
When a record is created, the magic fields (id, create_date,...) are first removed from the vals as the user should not set a value for these.
However if a value for this is given in default value (e.g. defined in an ir.value), the creation would crash (sql error : column specified more than once) as the magic column would be added again.
These parameters are (or should be) irrelevant for
a search_count(), and they could actually break the
result or make it significantly slower (e.g applying
`order` on large tables).
This fixes a performance regression introduced by
0f43032b.
We could also raise an error offset/limit are
passed in combination with count, but that seems
unnecessary.
Also switched to "SELECT count(1)" for the count
query, as it is simpler and just as fast.
We'd get the same perf with * or any constant value,
as in "SELECT count('me in')", but let's keep it
simple ;-)
At rev 84e9a67cdf a check to avoid the creation of ir.model.relation for custom modules was added. The condition is not correct as based on the string instead of the field name. We do not have access to column name at this level but the the m2m relation table do start with x_ for custom fields (see __init__ method).