We edited the HTML, but did not include the accompanying increase of modification date.

See nexedi/erp5!662 (comment 65953) for more on this issue.

/reviewed-on nexedi/erp5!790

When a workflow transition was not used by any state,
calling some_module.Module_listWorkflowTransitionItemList was raising:
    [..]
      Module script, line 49, in Module_listWorkflowTransitionItemList
       - <PythonScript at /tsxx/Module_listWorkflowTransitionItemList used for /tsxx/some_module>
       - Line 49
        result['listbox_parameter_dict'][transition_id] = [(state_variable, allowed_state_dict[transition_id])]
      Module AccessControl.ZopeGuards, line 83, in guarded_getitem
        v = object[index]
    KeyError: 'some_action'

Now, pick transition among the allowed_state_dict instead of workflow.transitions.

This also fixes mass change state dialog "??? ()" issue.
This commit adds a transition that allows to reproduce that issue.
And this line covers the case :
bt5/erp5_web_renderjs_ui_test/PathTemplateItem/portal_tests/renderjs_ui_zuite/testMassWorkflowTransition.zpt (line 61)

/reviewed-on nexedi/erp5!788

Some things in support request app where working apparently a bit by chance and they no longer work with new erp5_hal_json (
09d00cae )

This aims at using erp5_hal_json as intended and simplifying the customisations by customising "only" the header gadget and not the page gadget.

/reviewed-on nexedi/erp5!787

erp5_real_time_inventory_accounting: Accounting Transaction state should be 'Posted to General Ledger' (stopped), not draft.

erp5_hal_json does not seem to allow using gadgets by page name,
anymore, so instead of using a custom "page template" gadget (
disclaimer: I have not really understood this fully at this point).

Instead of using a custom page gadget, just use a header gadget that
"extends" the default gadget but removes the add_url

Because the fast input should be used.

Also simplify a bit the test, there's no need to add a SR in this case.

This reverts commit 463855b6.

unlike RSVP.all, RSVP.hash is not patched to support cancellation, so
it's not safe to use it at the moment.

Such as:
  - getStartTransitionDateList() and getStartTransitionDate().
  - getDeliverTransitionDateList() and getDeliveryTransitionDate().
  - ...

Where the non-List accessors returned the last element.

Use Case: Set AT.{start,stop}_date from PPL stop transition date.

Until this commit, loading a ZODB Component would add it to the current Request
object to prevent its reference counter reaching 0 (thus its global being reset to
None and then being later GC'ed) when a reset happens in another Request.

This problem was found when investigating testFunctionalConfigurator failure
when installing bt5s as requested by the Configurator:
  1. Request R1 calls Zuite_waitForActivities importing extension.erp5.ERP5Zuite (M1).
     => Add M1 to R1._module_cache_set.
     => M1 ref counter equals to 2 (sys.modules and R1._module_cache_set references).
  2. R1 terminates and is GC'ed.
     => M1 ref counter equals to 1 (sys.modules).
  3. Request R2 runs Configurator configuring the Site.
  4. testFunctionalConfigurator calls Zuite_waitForActivities to wait for the
     Configurator to finish (request R3 which may take ~15 minutes). This calls
     time.sleep() in a loop where 'time' module is imported at top-level.
  5. R2 installs bt5 triggering reset.
     => M1 ref counter equals to 0.
        ===> M1 global variables are reset to None and thus 'time' is set to None
             raising an Exception in the next call of time.sleep() in the loop.

The easiest way would be to have a hook on sys.modules dict lookup and thus add M1
to R2._module_cache_set when being imported, but this is not supported... Instead
create a global cache on erp5.component package.

There is no action with referecne `list` on support request module.

This test alone does not define site-dependent security groups, so this
change is not required as such.
But in real configuration, ERP5 instance define security groups, which
sometimes (often ?) depend on which site a document or user is linked to.
When multiple groups apply (ex: having a certain function at a certain
site), the final group is composed by joining group identifiers using
underscores. This forbids using underscores in group identifiers.
And group identifiers are generated from document's (here, the Category)
codification if set, otherwise the reference if set, otherwise the document
identifier.
So set a underscore-less codification on these categories so this test can
be run on tests where security is configured to include site in security
groups.

as the comment (which is removed now) said:

```
        XXX: This method is broken, since "total_size" field is not
        present in the listbox rendering any longer. It should be
        removed.
```

Also:
- make sure that Listbox always knows how many lines are to be displayed.
- add testPagerWithCustomListMethod in listbox_zuite of erp5_ui_test,
which checks that listbox jumping to previous/next/first/last works correctly
for a list method that returns only the elements that the listbox is to render.


/reviewed-on !776

Default method seem to work just fine and is compatible with the recent
improvements in erp5_hal_jon_style (
09d00cae )

set.__iadd__ does not like non-set values, but the intent of this API is
to allow just any iterable (producing strings) to be returned, so use
the more tolerant set.update method instead.

This is a mix of features to help time tracking and reporting with small unrelated bug fixes and improvements.

To help time tracking we show the time on support request with hour:minutes everywhere and set a stop date automatically. We also add interaction so that "user doing" is automatically added to the support request so that we can see "who did" and make reports based on this.

Bug fixes/unrelated features:
 * when project does not have supply lines, show all services members of the support request use. Also change to pass relative URL of services, usually we don't pass IDs, but relative URLs. This is a visible/breaking change, because it's now required to configure services use. Because it's also required in erp5_crm, I expect it would not break too much. 
 * momentjs' related time was not refreshed, so they kept displaying "a few second ago" even hours later
 * don't set HTML source as post title
 * prevent double submit of comments in support request app.


/reviewed-on nexedi/erp5!780

predecessor, successor and trade_phase ListBox fields didn't have their
own fields and thus only the title was displayed (eg 'Ordered' for
'Manufacturing/Ordered').

This allows breaking a circular dependency of catalog on itself: to
index a local role, catalog must tell whether it's granted to a group or
to a user, which used to happen using a catalog lookup (via PAS API),
preventing catalog from being rebuilt from scratch.

With this change, ERP5Site_filterUserIdSet must be implemented when
custom role columns are used (viewable_owner is not concerned by this
change).
If it is implemented in a way which does not rely on catalog,
catalog become rebuildable from a lone ZODB, without double-indexation
of users nor security_uid explosion.
Nothing prevents relying on catalog (which mayu be desired for
backward-compatibility purposes), but such script will not be provided
by generic ERP5, as the impossibility of recreating catalog from ZODB
is considered a bug, and introducing such script amounts to introducing
this bug.

Create ERP5Site_filterUserIdList in tests which add custom viewable_*
columns.

Use Owner role to help filling user_set: owners must be users, never
groups.
Only look for users among group_ids when they are candidate for indexation
into a catalog_role column. This should very significantly reduce the
number of user lookups, even bringing it to zero on instance with default
catalog setting of catalog_role_set = set(["viewable_owner"]) .
Also, only look for users if there are candidates for lookup.
Also, rename local variable to better describe its content.
Also, use set.__iadd__ instead of set.update (shorter).

Test loop invariant outside of loop.
Use clearer and more consistent variable names.
Avoid iterating twice when once is enough, allowing removal of
"local_role_dict" local and simplification of corresponding dict
structure.
Avoid creating new instances when the existing one can be mutated.
Evaluate dict as boolean instead of constructing an empty dict to compare
against.
Avoid using %-operator just to concatenate strings (allows en-passant
typechecking).
Drop unused locals.
Do not use a dict when a set would be sufficient.

So that it can be reused outside of group manager without having to
duplicate the backward-compatibility order.
Also, simplify code.