on python2, we use have a [python patch][patch] to globally make
asyncore use poll, at that time we were using medusa and it was not
exposing an option to use something else than select.

[patch]: https://lab.nexedi.com/nexedi/slapos/-/blob/180f1553fa7774f05ae8e739dfef43fa2ae56c0a/component/python-2.7/asyncore_poll_insteadof_select.patch

also add the missing tests

Because the librairies used here were never ported to python3.

Notable changes:

  - ERP5Site_createGoogleUserToOAuth is dropped
  - internal API changed radically, so customizations made by overriding
    scripts are broken.
  - the core logic is now implemented in a connector class (still in
    portal_oauth for simplicity, but it would be simpler to move it to
    portal_web_services)

No changes required in the google console, the redirect uri is still
ERP5Site_receiveGoogleCallback

page templates unicode resolver already takes care of coercing when
needed and this is directly compatible with python2 and python3
Co-authored-by: Kazuhiko SHIOZAKI <kazuhiko@nexedi.com>

This only works on python2

for same reason as the xhtml_style version, there is no problem with
having more than one listbox, see d2dedca3 (core: remove useless log
in Base_callDialogMethod, 2020-07-03)

The test testCMFActivity.TestCMFActivity.test_onErrorCallback was
failing because some weak references were not dropped.
The reason was that _DequeueMessageException was declared as global
and it kept a traceback including the variables pointed by the weak
references, preventing their garbage collection.
This was fixed by using a class for the _DequeueMessageException
instead of a singleton, and creating a new instance on every raise.

This test assumed that activating a method with a string of length `n+x`
as argument would be serialized in `x` more bytes that activating the
same method with a string of length `n` as argument, which is only true
for protocol 0 for multiple reasons ( short strings and long strings
are serialized differently, "frames" are used with protocol 5) and
db.string_literal seems to introduce a difference.

Change the test to mock Message.dump to produce longer dumps instead of
relying on this assumption.

_resultForDoCleanups no longer exists on python3, getting the status
of a test in teardown is very different accross python versions ( see
https://stackoverflow.com/q/4414234 ) and just executing this code when
the test failed only has the problem that a test method may count as
two failures or errors, which is after all not a big problem.

The default manager user (ERP5TypeTestCase) now have a random password
generated at the beginning of test. The password is stored as a class
attribute of the test case instance.
Many tests have been updated to not generate manager user and use the
existing one instead. When tests need to create users (for example
users in the root acl_users), we try to delete the users afterward.

Another significant changes is that we no longer have the same user in
ERP5/acl_users and /acl_users, some tests were logging in as the root
user in ways that never seemed intentional.

For functional tests, the approach is that ERP5TypeFunctionalTestCase
sets cookies with manager username and password before running zelenium
tests, so that in case tests want to log in again as manager, they can
read the username and password from cookies.

This also revealed (through test_manager_actions_on_portal from
testERP5Core) that some tests were running with a user without all the
expected permissions and the "Manager Components" actions was not
visible for the default manager user. Fixing this also revealed that
two actions ("Manage Components" and "Manage Callables") had the same
priority, so the later was modified to use a different priority.

See merge request nexedi/erp5!1943

- setting configurator header link
- allow auto_sync when adding opml
- update form title
- fix landing page
- fix dispatch
- fix form naming
- fix form and fields naming (coding style tests)

   Otherwise, alarms to fetch what was paid (status) don't work.

  Mistake was probably there because copy and paste

Raise an exception instead of casting to string unhandled search values,
because it's more likely that caller accidentally passed a wrong value.

See merge request nexedi/erp5!1864

- site web section changes
- update app router
- update app panel
- adapt monitoring jio gadget and jio storage
- add form views configuration
- add new portal types, views and actions
- migration_version check
- update precache manifest
- sync checks and redirection
- refactor dispatch page according to new monitor urls
- add opml redirects to front page
- slapos portal types
- drop obsolete files
- drop auto-sync on add OPML page
- parameters gadget field for instance tree view
- custom views for software instance resources and processes
- options header gadget (custom header options)
- custom submit gadgets
- graph gadget field

- allow custom app configurator
- render listbox form field extra configuration
- fix date format rendering
- allow configuration form fields to have extra js code
- allow to customize listbox column styles
- allow to customize listbox jio_alldocs
- fix custom view handling
- fix gadget field rendering
- allow dev user to set custom header options
- truncate long string fields
- add custom submit gadget option
- improve renderjs_extra
- update graph gadget

Some general changes to test framework and shared test code

See merge request nexedi/erp5!1936

this happens with new ghostscript, but this does not seem a critical
problem.

The test was only testing with empty cache

This should not happens, but if this happens we should not just return
in the middle of the loop, because this returned value is not used,
so we would just silently edit partially.

The first commit of this merge request was introduced for zope4py3 branch but it revealed issues on py2 as well, i.e. we don't have roles in auto-generated tester methods.

```
======================================================================
FAIL: test_method_protection (testSecurity.TestSecurity)
This test will list all implicitly Public methods in any objects in ZODB.
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/(SR)/parts/erp5/Products/ERP5/tests/testSecurity.py", line 113, in test_method_protection
    self.fail(message)
AssertionError: 
The following 7 methods have a docstring but have no security assertions.
	/(SR)/parts/erp5/product/ERP5Form/PreferenceTool.py:65 isPreferredVcsPushMode
	/srv/slapgrid/slappart19/t/eiy/soft/8a7759fd7b65b20d9f87713605745d05/parts/erp5/product/ERP5Type/Accessor/AcquiredProperty.py:217 hasTelephoneValidationState
	/(SR)/parts/erp5/product/ERP5Type/Accessor/Base.py:219 hasViewFormIdList
	/(SR)/parts/erp5/product/ERP5Type/Accessor/Constant.py:94 isWebDocumentType
	/(SR)/parts/erp5/product/ERP5Type/Accessor/Content.py:224 hasImage
	/(SR)/parts/erp5/product/ERP5Type/Accessor/ContentProperty.py:283 hasImageWidthList
	/(SR)/parts/erp5/product/ERP5Type/Accessor/Translation.py:233 hasFrTranslatedTitle

----------------------------------------------------------------------
```

See merge request nexedi/erp5!1911

If a transaction is selected but instead of having accounts as
source/destination on lines it acquires the organisation set as
source/destination on the transaction, this causes an AttributeError
getGapList when trying to use this organisation as an account.

To fix this, adjust the getMovementHistoryList parameters to select only
lines using accounts, by passing node_uid, which is O(n) on the number
of accounts, but the number of accounts is not supposed to be too large
and other reports also do this assumption.

This prevents one failing test to make all subsequent tests to fail

this is removed from python3, it was not so useful here either

use sys.maxsize, see discussion on
nexedi/erp5!1751 (comment 203735)