Also change places where xml_marshaller were used directly to always use
the wrapper from utils (except in tests for simplicity)

On python3, deleting a chmoded folder causes this error:

    Traceback (most recent call last):
      File "/usr/lib/python3.7/shutil.py", line 423, in _rmtree_safe_fd
        dirfd = os.open(entry.name, os.O_RDONLY, dir_fd=topfd)
    PermissionError: [Errno 13] Permission denied: 'directory'

    During handling of the above exception, another exception occurred:

    Traceback (most recent call last):
      File "slapos.core/slapos/grid/slapgrid.py", line 607, in processSoftwareReleaseList
        software.destroy()
      File "slapos.core/slapos/grid/SlapObject.py", line 360, in destroy
        rmtree(self.software_path)
      File "slapos.core/slapos/util.py", line 267, in rmtree
        shutil.rmtree(path, onerror=chmod_retry)
      File "/usr/lib/python3.7/shutil.py", line 491, in rmtree
        _rmtree_safe_fd(fd, path, onerror)
      File "/usr/lib/python3.7/shutil.py", line 425, in _rmtree_safe_fd
        onerror(os.open, fullname, sys.exc_info())
      File "slapos.core/slapos/util.py", line 264, in chmod_retry
        func(failed_path)
    TypeError: open() missing required argument 'flags' (pos 2)

Our error handler was assuming that on python 3 the EACCESS for the
directory would happen first on an os.lstat call [1], which is true when the
top level folder is not listable, but it did not support the case where
the top level folder is listable but contain another folder that is not listable.
In that case, the first error is an os.open [2]

Tests from erp5.util (where this function came from) only covered the 1 case but
new tests in slapos.core covered the 2.

Anyway, do the "chmod folder and retry from beginning" for both 1 or 2 cases.

[1]: https://github.com/python/cpython/blob/3.6/Lib/shutil.py#L475

[2]: https://github.com/python/cpython/blob/3.6/Lib/shutil.py#L418

Since bea2620a, the controller uses parent_relative_url instead of
portal_type (because the portal_type canbe translated). For "Sale
Invoice Transaction", the module is accounting, so rename the view page
accordingly.

    getId is invoked before the updateLocalRolesOnSecurityGroups is invoked, so Shadow user has no permission on the context.

This workflow only makes it everything slow for no benefit.

Added a new movement group
Delivery Builder was reconfigured to perform accordinally

See merge request nexedi/slapos.core!229

See merge request nexedi/slapos.core!227

including the edge case where software permissions have been changed

This is usefull for display warning related to mantainence or changes of policies.

slapos_erp5: Accelarate Login by ignoring security_uids when scripts Login_is** scripts are called as manager

This avoid extra checks when login is related to Computer or Software Instances.

slapos_jio: Test password before reset it and give feedback when password don't comply with the policy

Just update to follow up fixes already done on the english version

  Test if the user adds a password that isn't complaint with the policy
  Try to logout and login with the new password
  Add test for the use removes the login he is using
  Add test for the user changes password he is using
  Include translated version of the tests to manage password

Mostly handle the response a bit better, and log out user if he changes the current login. This prevents his account get blocked.

This gadget aims to centralise entire "helper" test on various contexts, in this way translate becomes easy and update information is quite simple.

In order to use the gadget, you should include a template on rjs_gadget_slapos_annotated_helper_html:

<script id="add-new-example-header-text" type="text/x-handlebars-template">
Some text or <details> <summary>Some test<summary> <p>Content...</p></details>
</script>

and include the gadget on the target HTML:

<div data-gadget-url="gadget_slapos_annotated_helper.html"
data-gadget-scope="annotated_helper"
data-template-id="add-new-example-header-text"
data-gadget-sandbox="public">
</div>

This is also supposed to be compatible with GadgetField or usable from getDeclaredGadget(...)