See merge request nexedi/erp5!1775

- add back the drone collision logic
- check drone out with map limits
- randomize seed library
- use seed to randomize values
- terrain texture change
- adapt default configuration
- refactor to allow map_size as parameter (instead of min-max lat-lon)
- new map function
- camera and size limits
- enemy team
- enemy AI script
- add flags and obstacles
- obstacle and flag collision
- allow dronemanager to take cartesian coordinates
- game finish rulls
- user score and result message
- refactor AI default script (drone dodges obstacles)
- new drone implementation for enemy drone (API)
- add getDroneViewInfo to API(s)
- restore drone color dot trace (improve color pick)
- allow random seed as url parameter
- handle fullscreen
- flag weight as parameter
- show default seed value, and make input mandatory
- do not show enemy drone logs on UI
- flag collision crashes the drone
- update score calculation
- drop flag weight (model and parameter)
- use geo coordinates in general
- refactor enemy drone to use cartesian
- add debug mode for test using console logs
- async obstacle detection (API changes)

- based on drone simulator

- babylon gadget msg htm
- babylonjs gadget fullscreen (includes mobile fix)
- bug fix in offscreen canvas
- scroll focus for usability

See merge request !1805

See merge request !1804

- drone simulator getContent returns console logs if debug mode

See merge request !1803

- rename `getSpeed` and `setSpeed` into respectively `getAirSpeed` and `setAirSpeed`
- add yaw, speed and climbRate in drones dictionary
- stop hardcoding leader's id
- remove setAltitude function
- use loiter with coordinates as parameters

- fix setAltitude:
  - Update `_last_target` when calling `setAltitude` otherwise `loiter` is broken.
  - Convert altitude parameter from absolute to relative value.

- fix getInitialAltitude
  - return absolute altitude

- fix getAltitudeAbs

See merge request nexedi/erp5!1801

This reverts commit 25cf4933.

Revert "patch : Caching Policy Manager should not override existing response headers."

This reverts commit 7187fb69.

Base, WebSection, OOoDocumentExtensibleTraversableMixin: do not compute the response body in 304 Not Modified response case.

Base, WebSection, Sound, OOoDocumentExtensibleTraversableMixin: set Cache-Control header in 304 Not Modified response case as well.

- Use portal_status_level when redirecting
- Check that password and confirmation match
- Fix  wrong argument name `password_confirmation` sometimes used

See merge request nexedi/erp5!1792

If mapping is not a dict, we store a message that would fail later when
being viewed, typically breaking History tab when used in workflow
history.

Instead of displaying the same ImportError without any detail for both
cases of a non existant module or of an existing module which cause
error during import, improve the second case by displaying a different
message containing the original traceback.

The pattern of using > will start by erasing the file, so if mysqldump
does not run to completion developer will loose the previous dump, but
in such case, that previous dump was better than nothing.

User is usually interested by last movement in stock, not by several
months/years old movement. Thus display first fresh movement, and user
could then go to next pages to see oldest movements if needed.

nexedi/erp5@21f85a6f
added zope2 compatibility, but nexedi/erp5@21f85a6f (comment 180420) noted
the given assumption isn't true, which is why fixup commit nexedi/erp5@7a2c616e was
added. But nexedi/erp5@21f85a6f (comment 184548) noted
that nexedi/erp5@7a2c616e missed one case,
which is why this fixup commit is added. Now everything should be ok.

Because these two plugins implement the refresh of the authentication
token (unlike the facebook plugin), we can use a longer cache here.
With the previous value of 1 day, when not using the site for more than
1 day, token would expire from cache and user have to authenticate again.
Because the scenario of "access is revoked from authorization server"
is handled by refreshing the access token, we can use a longer duration
here, so that user do not have to login again after not using for two
days.

After server restarts, mariadb uses an error 1927 which is mariadb
specific, with no corresponding constant in CR module when server
restart. This is the case at least with mariadb 10.3.38.

This fix the following problem:
 - start zope and use mariadb connection
 - restart mariadb
 - use mariadb connection again, see an error:
    Error Type: OperationalError
    Error Value: (1927, 'Connection was killed')

This change to treat 1927 as an error for which we can reconnect and
retry the query. Some other connectors do it, for example sqlalchemy in
https://github.com/sqlalchemy/sqlalchemy/issues/5493

This prevents tracebacks during shutdown:

  Unhandled exception in thread started by <bound method Thread.__bootstrap of <Thread(Thread-1, stopped daemon 139820471113472)>>
  Traceback (most recent call last):
    File "soft/lib/python2.7/threading.py", line 774, in __bootstrap
      self.__bootstrap_inner()
    File "soft/lib/python2.7/threading.py", line 814, in __bootstrap_inner
      (self.name, _format_exc()))
    File "soft/lib/python2.7/traceback.py", line 242, in format_exc
      return ''.join(format_exception(etype, value, tb, limit))
    File "soft/lib/python2.7/traceback.py", line 141, in format_exception
      list = list + format_tb(tb, limit)
    File "soft/lib/python2.7/traceback.py", line 76, in format_tb
      return format_list(extract_tb(tb, limit))
    File "soft/lib/python2.7/traceback.py", line 100, in extract_tb
      linecache.checkcache(filename)
  TypeError: 'NoneType' object is not callable

Also rename the method because this is not longer *Z*server.

This waitress close() API is not public part of API, so we still ignore
the case where the server thread does not stop after 5 seconds. This
happens if some HTTP connections are still open at the end of test.