When a module is updated, the sql constraints were checked against postgres pg_get_constraintdef() to see if they have diverged from the one defined in python code.
However this check was failing as the syntax returned by pg_get_constraintdef() (e.g. "CHECK ((credit * debit) = 0::numeric)") differs from the one returned by `unify_cons_text` method (e.g. "check(debit*credit = 0")) so the constraints were always dropped and recomputed.
This adds a definition column on ir.model.constraint (storing original defintion) to detect changes.
The pg_constraint is still inspected to make sure we detect constraint without ir_model_constraint but with postgres constraint defined (e.g. name_uniq defined in base.sql).
This changes also the date_update field when the definition changes.
Recently, the graph view was changed to prevent changing groupbys
for active custom filters. Unfortunately, I did not take into account
the fact that pivot table can be drawn in part when expanding rows.
The "frozen" parameter was then undefined and caused display problems,
this patch should fix the issue.
The method was expecting that name_get() returns complete and in-order values.
Because of this, some records in the recordset could end up without a value.
[FIX] sale: sale report view, wrong domain
Quotations are sale.order with state draft and sent
Sale orders are sale.order with state other than draft, sent and cancel
During the update of a module, the existing foreign keys are dropped if they have a different ondelete_rule than the one specified on the field.
The foreign keys for many2one transiant -> non-transiant are created with cascade rule by default (see `m2o_add_foreign_key_checked` method) so the check needs to be realised in the same conditions.
The problem was that when the user manipulates the graph view (in pivot
table mode), the graph view resetted the group by facet in the search
view. It was not a problem unless a custom filter with a groupby was
already there, in which case, the group bys were duplicated.
The search view is now smarter, it only resets the additional groupbys
(and col_groupbys). Also, to prevent usability problems, the graph
view disable the '+/-' groupbys added by a custom filters.
Note that this fix is only temporary: a revamp of custom filters, facets,
search view is coming in the next months. (at least, that's the idea). Right
now, too much 'search logic' is in the graph view.
Another note: this fix is somewhat fragile: it makes some assumptions
about the search query (mainly that the custom filter is the first facet,
also, that no other filters add groupbys/col_groupbys)
Fabien Meghazi