Please see 06e8f728aafa5c36d91975a2bedf8e6902d913f2

The original implementation were passing the object's url. So
we need to get the object's value again through portal_catalog.
But if we create a new Support Request just now, and then create
a post to use this SR as follow up. We can't get the SR value due
to the Vincent's reindex change: https://lab.nexedi.com/vpelletier/erp5/commit/2e35fa174ebafd2a6c9af84c116c35c29a0261c1
So pass the object's value directly.

After some debugs in this file and Products/DCWorkflow/DCWorkflow.py.
I found after creation, the Knowledge state is already 'visible'
Maybe related to https://lab.nexedi.com/vpelletier/erp5/commit/59e7ed2946b612076ad21623f80c4e2f50f64199#diff-3
But not 100% sure. Wait Vincent confirm it.

Replase it by add reindex_on_commit=True in newContent

Calling these can break catalog by not caring about concurrent indexation
activities. Even non-restricted code should not call these, but:
- they cannot be reliably renamed, as there exist activity dependencies
  looking for these exact names (which is a bad practice on its own)
- there is no simple access control mechanism for non-restricted code
So rely on developer discipline instead.

Also, simplify scripts while doing so.

These methods must not be called synchronously:
- they can break catalog by not being careful enough about other
  reindexations which may happen in parallel. See the serialization_tag
  mechanism for more.
- indexation gets executed in the security context of the user causing the
  call, which may lead to an indexation result different from what happens
  when indexation happens with an all-accesses user.

To remove the need to directly call immediate reindexation methods,
allowing to prevent further misuses of these.

Do not check portal's isIndexable, it is checked again in ZSQLCatalog.
Directly bind superclass callable instead of doing an extra getattr +
call.
Trivial python simplifications.
Assorted coding style improvements.
Drop unhelpful comments and docstrings.

298b59a0 (comment 52160)

This prepares for the removal of --firefox_bin and --xvfb_bin options. The
API between the testnode and the tested software must be exclusively generic.

generateMovementListForStockOptimisation was raising error when there was inventories.
The method was assuming that doing getObject() on an inventory line (from
getFutureInventoryList) was always returning a movement. But sometimes we get
the inventory itself. So do not use getVariationCategoryList, but use the variation_text
stored in the stock table.

- Allow superuser to get RSS ;
- Remove useless portal_skin=RSS paremeter.

Allow user to select which columns to display on module.
A new panel has been added for the UI.
The selectable column list is calculated from the columns and all_columns parameters.
Configuration is for now stored in the URL (as sort).

[erp5_pdm, erp5_trade] Fix bad field library field id. Rename ledger field library proxy field from my_ledger to my_view_mode_ledger.

[ERP5/tests/testERP5Commerce.py] quantity_unit/unit/piece category is used but it does not exist in this test. Add quantity_unit/unit/piece in afterSetUp.

To use non-existing category may lead an unexpected result. This time, getAcquiredCategoryList did not work correctly, then category values of products were not correctly indexed and test_17_getProductListFromWebSection failed.

Revert "[ERP5/tests/testERP5Commerce.py] Add erp5_configurator_standard_categories business template to provide quantity_unit/unit/piece."

This reverts commit 93c6b048.

erp5_configurator_standard_categories depends on too many business templates. I gave up using it.

[ERP5/tests/testERP5Commerce.py] Add erp5_configurator_standard_categories business template to provide quantity_unit/unit/piece.

To use non-existing category may lead an unexpected result. This time, getAcquiredCategoryList did not work correctly, then category values of products were not correctly indexed and test_17_getProductListFromWebSection failed.

/reviewed-on nexedi/erp5!538