See merge request nexedi/erp5!1251

1. set 0 as default value, otherwise leaflet_display_side will always be true
2. show message instead of crash
3. correct parameters
4. clean code

Since 206a8e25 (erp5_officejs: new OfficeJS Slideshow Editor application,
2020-07-06) this test runs with erp5_run_my_doc installed, and the first export
action for Web Page is no longer "Export as SlideShow" but "Export as Chapter".
This is an action which is here since a long time in erp5_run_my_doc, but since
erp5_run_my_doc was not installed in this test it was not impacting this test.

Adjust the tests so that they don't assume that the first export action is
"Export as SlideShow" but are more independent on which is the first action.

Also, all actions have a new index, so places where we select the action by
index no longer work. Instead of updating the indexes, select by label which
should be more future proof.

Mathematically, 1/(x^y) is the same as x^-y, and despite float caculation the results is usually precise enough :
> Math.pow(10, -5)
0.00001
I tested this on few platforms : firefox 68.12, firefox 77.0, node v10.19.0

Unfortunately on (recent ?) chromes :
> Math.pow(10, -5)
0.000009999999999999999
Which is a horrible value to use as step (and it prevents the form submission as most floats the user will enter won't match the init value * x * the step.

For an unknown reason, I get a more consistent result using the formula "1 / Math.pow(10, 5)", which returns 0.00001 on all tested platforms.

If we find more issues in the future, we maybe should try building the step using strings. Using strings for manipulating floats is in reality widespread, and many languages do so to round floats (ie: https://github.com/python/cpython/blob/4a97b1517a6b5ff22e2984b677a680b07ff0ce11/Objects/floatobject.c#L925)

The precision of 5 is not random-picked, it is the minimum precision needed to manipulate prices for currencies with 2 digits, like euros.

When portal_contributions.PUT() is called, Document was not converted (and `base_data`
was not set) by document_conversion_interaction_workflow as this was only done on
_set{Data,File}() and OFS.Image.update_data() sets `data` property directly.

This fixes failures on testWebDavSupport.test_GET_on_{contributionTool,document}
with erp5_full_text_*_catalog bt5 installed: SQLCatalog_deferFullTextIndexActivity
Activity failing with `NotConvertedError` when trying to convert to txt as
`base_data` is not set.

ZODB Components: erp5_trade: Migrate TestOrder{Mixin} and all Unit Tests inheriting from these classes.

Also, Split testInvoice to a Module Component in erp5_invoicing for Mixins used
by several other tests and Test Component in erp5_simplified_invoicing (initially
everything was migrated to erp5_invoicing but this does not work as Alarm_buildInvoice
is different in simplified and advanced invoicing).

efd71291 added on erp5_invoicing a test dependency on erp5_advanced_invoicing.
However, testRule for example uses erp5_invoicing but not erp5_advanced_invoicing
so this would not work.

Moreover, erp5_{simplified,advanced}_invoicing enhance erp5_invoicing by providing
a container for Invoice Line so to match real world use cases it is better to
have two CodingStyle tests: one checking erp5_invoicing+erp5_simplified_invoicing
and the other one erp5_invoicing+erp5_advanced_invoicing.

Also, fix CodingStyle issues on erp5_advanced_invoicing.

This reverts commit bc67c2c4.

This is seem to stall testFunctionalAdvancedECommerce with this error:

      Module script, line 5, in Workflow_ensureUserId
       - <PythonScript at /erp5_portal_7f1517681f85de9695ca475d69c4d66f/portal_workflow/login_interaction_workflow/scripts/Workflow_ensureUserId>
       - Line 5
        user = state_change['object'].getParentValue()
      Module AccessControl.ZopeGuards, line 87, in guarded_getitem
        if getSecurityManager().validate(object, object, None, v):
    Unauthorized: You are not allowed to access '2' in this context

let's revert for now, we'll re-do this with more tests.

New dependencies were added in 206a8e25

See jio@24b938cb

Should fix most of the cases where reloading components cause breakage.
There is still a race condition between a transaction's cache being
actually flushed (which can only happen at transaction boundaries) and
another transaction doing the actual class reload, which will
immediately affect all the started transactions in the same process.