Otherwise, method draw enter an infinite loop if users selected some elements

https://github.com/methodofaction/Method-Draw/tree/463e526dbc6ec48561822fde5dec02e085e86497

replaces nexedi/erp5!1691 

See merge request nexedi/erp5!1709

See merge request nexedi/erp5!1715

There is no good reason to prevent editing in cancelled state and
it can be needed, for example to remove a cancelled payment
transaction from a payment transaction group.

This is different from the total of durations of test result lines
because they are executed in parralel on different testnodes. The
actual duration is just the difference between start and stop, so
that we don't have to calculate.

This makes it easier to understand the duration in minutes and hours

Ran https://github.com/python/cpython/blob/3.7/Lib/lib2to3/fixes/fix_asserts.py after adding the missing methods (assertRegex and assertRaisesRegex) on ERP5TypeTestCase.

Also ran a custom fixer rewriting
 
```py
self.assertTrue(a in b)
```

into

```py
self.assertIn(a, b)
```

See merge request nexedi/erp5!1708

Prepare for https://github.com/zopefoundation/Zope/pull/1075 ( Zope 4.8.7 )

See merge request nexedi/erp5!1707

and fix some missing cache

See merge request nexedi/erp5!1674

automaticLayout can not be used with firefox 68 (the browser from test
nodes) because it lacks support for ResizeObserver

triggered manually with "Format Document" of "Format Selection"
actions

This enables usage of @ in quick pick to search for symbols and make
sticky scroll show the current context

   Probably there are missing files but it is a start.

With tweaks to support firefox 68 from test nodes:
 - build with babel, see jerome/monaco-editor-erp5@cf10f2d5
 - special handling of ResizeObserver that does not exist in Firefox
   68: it is used only for automaticLayout, which is not strictly
   needed. So this feature will not work for firefox 68.

Enable sticky scroll

Set caching manager to must_revalidate_http_cache

Unindexation happens in a transation after the corresponding document was
deleted. So there is no context available to check the type of the deleted
document. So these filter expressions are meaningless.

Unindexation happens in a transation after the corresponding document was
deleted. So there is no context available to check the type of the deleted
document. So these filter expressions are meaningless.

Unindexation happens in a transation after the corresponding document was
deleted. So there is no context available to check the type of the deleted
document. So these filter expressions are meaningless.

The add-hoc assertin filter:

--

from typing import List

import lib2to3

from lib2to3.fixer_base import BaseFix
from lib2to3.fixer_util import Comma, Name

class FixAssertIn(BaseFix):

  PATTERN = """
      power< any+ trailer< '.' meth=("assertTrue" | "assertFalse")>
      trailer< '('
        comparison< (needle=any ( comp_op<'not' 'in'> | 'in' ) haystack=any) >
      ')' > >
  """

  def transform(self, node: lib2to3.pytree.Node,
                results: List[lib2to3.pytree.Base]):

    needle = results['needle']
    haystack = results['haystack']
    meth = results["meth"][0]

    method_map = {True: 'assertIn', False: 'assertNotIn'}
    method_in = meth.value == 'assertTrue'
    if 'not' in str(needle.parent.children[1]):
      method_in = not method_in
    meth.replace(Name(method_map[method_in], prefix=meth.prefix))

    needle.parent.children = [needle, Comma(), haystack]

so that we can run 2to3.fix_assert

This updates the change to 32cead86 (xhtml_style: set a bottom
margin for dialog action, 2019-05-31) that got impacted by a rule
added on html and body in 5dde17a5 (xhtml_style: make editors in
bottom group use full height, 2022-08-22)

this code was latin 1, causing issues for refactoring scripts

Form.__call__ replaces page template call, so it is also responsible
to set the content type in response.

With Page Templates, ZopePageTemplate._exec (called by
Shared.DC.Scripts.Bindings.__call__ and _bindAndExec) sets the
response content type and calls pt_render. Form.__call__ also calls
pt_render, so to behave like page template, it should also sets the
response content type.

In erp5_xhtml_style, we don't notice this, because global_definitions
contains a setHeader call to set content type to html.

Check that it correctly includes the web page and sets the content
type header.