W: Dangerous default value _MARKER (__builtin__.list) as argument (dangerous-default-value)

Since BusinessProcess.py was moved to portal_components, the cache in mixin/composition.py,
which was kept at the level of the zope process, was causing issue when portal_components
was reset : the classes kept in __class_cache would not get resetted, and later calls
to asComposedDocument would reuse them. In consequence, these classes would kept pointers
to objects in the old class BusinessProcess, causing bugs as these objects would have
been cleaned by the reset.

The first intuition was to remove this cache, unfortunately the reason of this introduction
(fixing a memory leak) is still valid nowadays, so I've decided to move this cache within
BusinessProcess, as it has the advantage of emptying it whenever portal_components is resetted.

For more information of the bug mentionned above, and the trials about removing the cache can
be found in the discussion of this MR :
nexedi/erp5!1032

ZODB Components: Migration dialog: Make sure that paths given in Working Copy (Preference) are valid.

Besides migrating Tools .py, this also moves the Tools themselves (portal_*):
 * Some of them were created before erp5_core being installed and this
   would not work anymore as their .py is now in erp5_core.
 * Some others were created after (via setupLastTool()) and this would not
   work neither as sub-objects may be installed by erp5_core.

Add them to template_keep_path_list to prevent them from being re-created like
ERP5Site.addERP5Tool() used to do (depend on 3180424b, 91393be2).

Not migrated:
  * AlarmTool: Needed for upgrader.
  * CategoryTool, IdTool: Bootstrap.
  * TemplateTool, TrashTool: Business Template installation.
  * SolverTool: TypeProvider.
  * ContributionTool: Imported by Products.ERP5.Document.Document used in many places, will be done later.
  * NotificationTool: Imported by ERP5.Document.EmailDocument, will be done later.

On re-create, all the sub-objects will be copied back and this should not happen
for Tools (example: portal_simulation currently created by addERP5Tool() but
later going to be migrated to ZODB Components and moved to bt5).

When force is selected, it should be updated regardless of user choice (for
example when `force_install` is set). But usually when installing a bt5, user
selects what should be updated (`object_to_update dict` later filtered to
`update_dict` (to not contain unselected objects)).

pylint: Fix false positive error `No name 'OOBTree' in module 'BTrees.OOBTree' (no-name-in-module)`.

We were using many different terms every for "username", this is an attempt to use a consistent term every where for login and a few other terms used for logging in and resetting password

See merge request nexedi/erp5!1164

As a follow up of nexedi/erp5!1158 , now that stock browser supports inventory conversion, the old converted inventory report is not longer needed.

The reason for removing instead of fixing is that this report was using approaches that have no equivalent in ERP5JS interface (select in the module and run a report based on selection, listbox domain tree), so we could repair it for old ui only, but we thought it would be better to remove it and improve stock browser further.

One reason this was broken is that it was using `SimulationTool.getAllInventory` which does not support brain attributes from `SimulationTool.getInventoryList`, so dynamic attributes like `node_title` were not displayed correctly. This adds a warning on `getAllInventory` docstring. This method has another user so we don't remove it.

See merge request nexedi/erp5!1162

Make stock browser a dialog where user can select the node category, for compatibility with ERP5JS

Introduce a preference to select the node category by default.

Allow to select quantity unit / metric type in the dialog to get converted inventories.

See merge request nexedi/erp5!1158

Uploaded the content of
https://github.com/nicolaskruchten/pivottable/tree/v2.23.0/dist/
using web dav and ran Alarm_checkSkinCache

This version allow sorting and fixes an XSS vulnerability

Browsers (at least chrome) saves the textarea content and restore it
when going back or resurecting a closed tab, which prevents loosing
unsaved work.
This browser feature does not work in "rich" javascript
editors, but this trick of syncronizing the rich editor content to a
hidden textarea seem to makes it work.

This is an emergency commit to repair failing tests

use "Slideshow" for now. "Slide Show" would also be OK, the discussion
about what's the best term is still ongoing, this is an emergency commit
to repair tests.

We fixed login forms to use consistently "Username", some translation
did not have message for "Username", but only for "user name".
Add missing "Username" messages, with same translation as existing one
for "User name"

with erp5_credential it's a bit more than password reset, because
there's also a feature to recover the username for a given email, but
all this can be named "Recover your Account". That was the term already
used in ERP5JS

Use consistent term as in PreferenceTool_viewChangePasswordDialog. In
dialogs the field label show "what information is in the field" not
"what the user is supposed to do with this field"

Other login related dialogs ( PasswordTool_viewEmailPassword ,
ERP5Site_viewCredentialRecoveryLoginDialog ,
PasswordTool_viewResetPassword ) all use Username.

Login_view/my_reference uses "User Login", but that's maybe OK, because
the context is different and there's also "User Password" on the same
form.

this has been merged in erp5_officejs

This gadget only works in ERP5JS, as it depends on multiple ERP5JS only gadgets.

Some functionnalities have not been backported for now:
* image upload, which seems broken on xhtml style
* test template (I was not able to trigger something from the UI)

Normally, it could be possible to create an OfficeJS app with this editor.

This functionality is enabled only if 'trusted-proxies' configuration
value is ('0.0.0.0',) by putting the following in zope.conf :

trusted-proxy 0.0.0.0

This report have been broken for years and does not work in new UI, now
that stock browser support inventory conversion this report is no longer
needed.

formulator does not seem to like generator for listfield items

getConvertedInventoryList is deprectated and use mergeResults which
does not support brain classes, so the brain.node_title was no longer
correct.
Use a simpler approach, same as Resource_getInventoryList ("stock
browser"), ie: getFutureInventoryList to get all future stocks +
getCurrentInventory and getAvailableInventory for each brain. It's a bit
less efficient because we now do 2 getInventory for each line instead of
3 getInventoryList merged, but that should be enough.

Also:
 - use float fields to format numbers properly
 - activate domain tree for site (instead of showing all nodes)

this 'relative_url | ascending' sort can lead to killer queries and it is of no use

Convert the amounts in inventory list view, the movement history list
views still show in movement unit for simplicity.

At the same time:
 - don't cast to int() in macro, but expect caller to pass already
   correctly formatted values for *_inventory
 - document macro like it was done for erp5_pdm_ui_test

This was never counting movement, but the number of inventory brain that
the getInventoryList should return.
Document this macro and rename argument everywhere.

init macro cannot open the stock browser report, because some other
tests are assuming that this macro remain in view mode and they can
select another view.
move the "select stock browser report" in a stock-browser-test specific
macro

make the node_category dialog pre-select the preferred node category

Previously, these tests were displaying the listbox with two lines, one
for source one for destination and where checking both lines at once.
Now that it's possible to filter by node category (site) in renderjs as
well, adjust the test to select the node categories one by one and check
that the results for each line category is what we expect.

Also fix a few typos

This allows to filter the node by site in new and old interface in
a consistent way.

The domain tree is also removed for consistency (so that we don't have
different features on new and old interface)

Old UI tests have also been updated to use more meaningful macro names.