This is a simple framework to implement services in ERP5 based on an
OpenAPI document.

A new type "Open API Type" (similar to "Base Type") is introduced,
this is responsible for the definition of operations and types.
The Open API document is set as text content of the Open API Type
and can be edited from the Open API Type.

For each service, a new portal type will be created. The portal type
use OpenAPIService as class and this is responsible for serving
requests. The process of serving requests is:
 - find the matching operation from the request method and request
   path
 - extracting request parameters and request body using the parameter
   definitions from the Open API Document
 - validate parameters and request body according to the schema from
   the Open API document
 - finding the method, this is done by using _getTypeBasedMethod with
   the operationId
 - calling the method and formatting the result or handling error.
   The default handling of errors is to reply with rfc7807 json
   responses, but it can be customized by defining an
   `handleException` type based method.

Typically, the services will be created in portal_web_services. From
there, there is also a view using a new SwaggerUI gadget to try out
the API.

What's not supported:
 - OpenAPI document in YAML format is only partially supported and
   have some limitations over JSON:
    - On python2 the order of operations is lost, the lookup of
      operations is not made in the order of the operations from the
      document. Also the operations are not in order in the SwaggerUI
      gadget.
    - The text editor does not provide rich editing of YAML
 - "partial" parameters in path elements ( /users/{user_id} is
   supported, but /documents/report.{format} is not )
 - XML (decoding of request bodies and parsing of responses) is not
   supported.

We use to show ID in case reference is missing, but since many years
this is no longer the case.

See merge request nexedi/erp5!1766

See merge request nexedi/erp5!1767

The fix from 7597c300 (accounting_l10n_fr: don't export
lines with 0 quantity in FEC, 2023-03-31) was not 100% correct,
lines with a quantity but with an asset price of 0 were still
exported.

We used a wrong condition that was causing lines with a debit
and asset credit (ie. different sides) to be output wrongly
with debit and credit.

in 5ae0508d (accounting_l10n_fr: Workaround encoding bugs
in Test Compta Demat, 2023-03-31) we replaced a few characters
that were known to be problematic, but while trying on some
other production data, we found new problematic characters.
This shown that this approach was too "optimistic" and that we
the safest way to have a file that is compatible with Test
Compta Demat is to use only ascii characters, this use
unicodedata to try to retain the original characters (so that
"Jérôme" becomes "Jerome") and also special case € to replace
it by "EUR"

    If two open orders are validated on the same transaction, archivePrevious script will get the an already archived as catalog will be outdated.

we pass the schema already, but for now this only supports basic
syntax highlighting

programmatically calling script.setParameterSignature('') was
setting _params to None and making it causing errors when calling
the script

This change modifies the FEC export to conform to the test software from
https://www.economie.gouv.fr/dgfip/outil-test-des-fichiers-des-ecritures-comptables-fec 
/ https://github.com/DGFiP/Test-Compta-Demat.

A new `test_compta_demat_compatibility` to enable compatibility mode was introduced.
This will be True when rendering the report from the dialog, i.e., using `AccountingTransactionModule_viewFrenchAccountingTransactionFile` and False when
using `AccountingTransactionModule_viewFrenchAccountingTransactionFileActive` directly
so the behavior of reusing FEC export for accounting interfaces should be the same as before.


See merge request nexedi/erp5!1764

> IX. – Le fichier des écritures comptables est nommé selon la
> nomenclature suivante :
>
> SirenFECAAAAMMJJ, où " Siren " est le Siren du contribuable mentionné
> à l'article L. 47 A et AAAAMMJJ la date de clôture de l'exercice
> comptable.

according to https://bofip.impots.gouv.fr/bofip/9028-PGP.html/identifiant%3DBOI-CF-IOR-60-40-20-20170607#III._Donnees_devant_figurer_12

> Dans le cas des écritures pour lesquelles il n’existe pas de référence
> de pièce (par exemple, dans le cas des écritures d’à nouveau), ce
> champ doit néanmoins être rempli. Dans ce cas précis, la référence
> sera alors remplacée par une valeur conventionnelle définie par
> l’entreprise. Celle-ci sera précisée dans le descriptif remis au
> vérificateur en même temps que le fichier des écritures comptables
> (cf. VI § 390).

La valeur conventionnelle dans cet export est de donner à PieceRef la
même valeur que EcritureNum pour les écritures sans référence de pièce.

"Test Compta Demat" complains about these
https://github.com/DGFiP/Test-Compta-Demat/blob/1563a42a260e754a7543b2ee178d4300d1582db5/src/testeur/trt_txt.pl#L1023-L1033

This removes some characters causing 'Erreur Wide character in
subroutine entry' bugs in Test Compta Demat from the XML output

The test software from
https://www.economie.gouv.fr/dgfip/outil-test-des-fichiers-des-ecritures-comptables-fec
only validates if the XSD is referenced with xsi:noNamespaceSchemaLocation

See merge request nexedi/erp5!1763

XLTE(*) follows the same top-level packaging structure like
wendelin.core does.

Registration is done via new `register_xpkg` function instead of just
loop so that pkgname argument is properly captured in created lambda and
does not change after going to next iteration.

(*) https://lab.nexedi.com/kirr/xlte

/reviewed-by @jerome
/reviewed-on !1762

In 5796a17a (core_test: Add test to make sure that wendelin.core
basically works) I added wendelin_transform to pylint so that the
checker could recognize wendelin.core's special wendelin top-level
module/package and handle ok

	from wendelin.bigarray.array_zodb import ZBigArray

this works, but e.g. for plain

	import wendelin			or
	from wendelin import bigarray

pylint currenly fails with

	AssertionError: explicit relative import, but no context_file?

To better see what is going on let's conside the following test program

	"""ZZZ"""

	from wendelin.bigarray.array_zodb import ZBigArray
	from wendelin import bigarray

	def main():
	    """main"""
	    _ = ZBigArray

and run pylint on it. Here is what it gives:

          wendelin_transform Module(wendelin)
          -> Module()

    visit_from From()
      basename: wendelin.bigarray.array_zodb
      names:    [('ZBigArray', None)]
      modnode:  Module(test)
        get_imported_module From() wendelin.bigarray.array_zodb
      importedmodnode: Module(wendelin.bigarray.array_zodb)
    ('_add_imported_module', <From() l.3 [test] at 0x7ff214b4cb90>, 'wendelin.bigarray.array_zodb.ZBigArray')
          wendelin_transform Module(wendelin)
          -> Module()

    visit_from From()
      basename: wendelin
      names:    [('bigarray', None)]
      modnode:  Module(test)
        get_imported_module From() wendelin
          wendelin_transform Module(wendelin)
          -> Module()
      importedmodnode: Module()
    ************* Module test
    W:  4, 0: Relative import 'wendelin', should be '' (relative-import)
    ('_add_imported_module', <From() l.4 [test] at 0x7ff214b4cf50>, '.bigarray')
    Traceback (most recent call last):
      File "./x.py", line 22, in <module>
        Run(['test.py', '--reports=n'], exit=False)
      File "/home/kirr/src/tools/py/pylint/pylint/lint.py", line 1332, in __init__
        linter.check(args)
      File "/home/kirr/src/tools/py/pylint/pylint/lint.py", line 747, in check
        self._do_check(files_or_modules)
      File "/home/kirr/src/tools/py/pylint/pylint/lint.py", line 869, in _do_check
        self.check_astroid_module(ast_node, walker, rawcheckers, tokencheckers)
      File "/home/kirr/src/tools/py/pylint/pylint/lint.py", line 946, in check_astroid_module
        walker.walk(ast_node)
      File "/home/kirr/src/tools/py/pylint/pylint/utils.py", line 874, in walk
        self.walk(child)
      File "/home/kirr/src/tools/py/pylint/pylint/utils.py", line 871, in walk
        cb(astroid)
      File "/home/kirr/src/tools/py/pylint/pylint/checkers/imports.py", line 288, in visit_from
        self._add_imported_module(node, '%s.%s' % (importedmodnode.name, name))
      File "/home/kirr/src/tools/py/pylint/pylint/checkers/imports.py", line 328, in _add_imported_module
        importedmodname = get_module_part(importedmodname)
      File "/home/kirr/src/tools/py/astroid/astroid/modutils.py", line 359, in get_module_part
        'explicit relative import, but no context_file?'
    AssertionError: explicit relative import, but no context_file?

we see that the line

	from wendelin.bigarray.array_zodb import ZBigArray

corresponds to

	('_add_imported_module', <From() l.3 [test] at 0x7ff214b4cb90>, 'wendelin.bigarray.array_zodb.ZBigArray')

and the line

	from wendelin import bigarray

corresponds to

	('_add_imported_module', <From() l.4 [test] at 0x7ff214b4cf50>, '.bigarray')

notice that in the latter case there is no 'wendelin' prefix and the
import goes as just '.bigarray' instead of 'wendelin.bigarray' which
leads to further crash in get_module_part.

-> Fix it by initializing wendelin module yielded by wendelin_transform
with .name set. Not sure why it is working longer imports without that.

Without the fix added test breaks as

    .../instance/slappart5/bin$ ./runUnitTest -v -v -v  testDynamicClassGeneration.TestZodbModuleComponent
    ...
    Running Unit tests of <class 'testDynamicClassGeneration.TestZodbModuleComponent'>
    ok
    testImportVersionedComponentOnly (testDynamicClassGeneration.TestZodbModuleComponent) ... ok
    testInvalidId (testDynamicClassGeneration.TestZodbModuleComponent) ... ok
    testInvalidSourceCode (testDynamicClassGeneration.TestZodbModuleComponent) ... ok
    testModuleSecurityInfo (testDynamicClassGeneration.TestZodbModuleComponent) ... ok
    testPylint (testDynamicClassGeneration.TestZodbModuleComponent) ... FAIL
    testPylintAstroidModuleGeneratedOnce (testDynamicClassGeneration.TestZodbModuleComponent) ... ok
    testPylintNamedtupleUnicodeLiteralsRegression (testDynamicClassGeneration.TestZodbModuleComponent) ... ok
    testReferenceWithReservedKeywords (testDynamicClassGeneration.TestZodbModuleComponent) ... ok
    testValidateInvalidateDelete (testDynamicClassGeneration.TestZodbModuleComponent) ... ok
    testVersionPriority (testDynamicClassGeneration.TestZodbModuleComponent) ... ok
    testVersionWithReservedKeywords (testDynamicClassGeneration.TestZodbModuleComponent) ... ok
    testWorkflowErrorMessage (testDynamicClassGeneration.TestZodbModuleComponent)
    Check that validation error messages are stored in workflow ... ok

    ======================================================================
    FAIL: testPylint (testDynamicClassGeneration.TestZodbModuleComponent)
    ----------------------------------------------------------------------
    Traceback (most recent call last):
      File ".../parts/erp5/Products/ERP5Type/tests/testDynamicClassGeneration.py", line 2347, in testPylint
        component.checkSourceCode()
      File ".../parts/erp5/product/ERP5Type/mixin/component.py", line 329, in checkSourceCode
        return checkPythonSourceCode(self.getTextContent(), self.getPortalType())
      File ".../parts/erp5/product/ERP5Type/Utils.py", line 540, in checkPythonSourceCode
        Run(args, reporter=TextReporter(output_file), exit=False)
      File ".../develop-eggs/pylint-1.4.4+slapospatched002-py2.7.egg/pylint/lint.py", line 1332, in __init__
        linter.check(args)
      File ".../develop-eggs/pylint-1.4.4+slapospatched002-py2.7.egg/pylint/lint.py", line 747, in check
        self._do_check(files_or_modules)
      File ".../develop-eggs/pylint-1.4.4+slapospatched002-py2.7.egg/pylint/lint.py", line 869, in _do_check
        self.check_astroid_module(ast_node, walker, rawcheckers, tokencheckers)
      File ".../develop-eggs/pylint-1.4.4+slapospatched002-py2.7.egg/pylint/lint.py", line 946, in check_astroid_module
        walker.walk(ast_node)
      File ".../develop-eggs/pylint-1.4.4+slapospatched002-py2.7.egg/pylint/utils.py", line 874, in walk
        self.walk(child)
      File ".../develop-eggs/pylint-1.4.4+slapospatched002-py2.7.egg/pylint/utils.py", line 871, in walk
        cb(astroid)
      File ".../develop-eggs/pylint-1.4.4+slapospatched002-py2.7.egg/pylint/checkers/imports.py", line 253, in visit_import
        self._add_imported_module(node, importedmodnode.name)
      File ".../develop-eggs/pylint-1.4.4+slapospatched002-py2.7.egg/pylint/checkers/imports.py", line 319, in _add_imported_module
        importedmodname = get_module_part(importedmodname)
      File ".../develop-eggs/astroid-1.3.8+slapospatched001-py2.7.egg/astroid/modutils.py", line 359, in get_module_part
        'explicit relative import, but no context_file?'
    AssertionError: explicit relative import, but no context_file?

    ----------------------------------------------------------------------
    Ran 13 tests in 137.609s

    FAILED (failures=1)

/helped-and-reviewed-by @jerome
/reviewed-on nexedi/erp5!1762

See merge request nexedi/erp5!1759

See merge request nexedi/erp5!1538

  On the automatic upgrade some bt5 (erp5_core) can be upgraded
  before erp5_property_sheets. Since, it isnt a problem if this happens
  if no Mixin configuration is present, it is better directly check the
  existence of the property on the bt5.

  In general we expect that erp5_property_sheets is upgraded before any
  other bt5 with TypeMixin definition.

       Use the proper way to include Mixin into a portal type

   Use the proper way to include Mixin into a portal type

  This value isnt exported anymore

 Handle type_mixin like property_sheets and use Business Template properly

  Allow the developer associate a Mixin on Business template level
without require to overwrite the portal type itself

  Include mixins can be done in the same way, allowed content types or
property sheets are set.

After updating Products.CMFCore we had issues on sites where URLTool was registered as ISiteRoot.

In the case of ERP5, we don't really need portal_url.getPortalObject(), this changes to use ERP5Site.getPortalObject(), while at the same time doing something so that portal_url.getPortalObject() also works.

See https://github.com/zopefoundation/Products.CMFCore/pull/126 for more details

See merge request nexedi/erp5!1758

this is a "magic" local variable used to add more debugging information
in zope traceback, it's not supposed to be used directly so it was a
false positive in our case.

This extends the default value of --dummy-variable-rgx that was defined
to _$|dummy in pylint 1.4.4 (
https://github.com/PyCQA/pylint/blob/pylint-1.4.4/pylint/checkers/variables.py#L296
)

This can happen when request is used as a cache storage, especially
after eab00f14 (deferred_style: Properly keep request parameters,
2021-03-10)

  that it was created.

  squash me