1. 22 Dec, 2020 7 commits
  2. 21 Dec, 2020 1 commit
  3. 15 Dec, 2020 2 commits
  4. 12 Dec, 2020 2 commits
    • Koseng's avatar
      Complete client alarms and conditions example · b7431220
      Koseng authored
      Complete client alarms and conditions example for usage with OPCFoundation UA-.NETStandard-Samples Quickstart AlarmConditionServer
      b7431220
    • Koseng's avatar
      Add variables to generated event filter and event object · 8ba94e58
      Koseng authored
      Event types can not only contain properties, but also variables. For example the ConditionType, which is the foundation of Alarms and Conditions, contains variables. Add the variables to the generated event filter and event object.
      8ba94e58
  5. 09 Dec, 2020 2 commits
  6. 08 Dec, 2020 1 commit
  7. 07 Dec, 2020 6 commits
    • oroulet's avatar
    • oroulet's avatar
      update prosys crypto client · e8eaa97d
      oroulet authored
      e8eaa97d
    • Marco Panato's avatar
      User manager and permission ruleset fix (#352) · ac4373e0
      Marco Panato authored
      * User-defined PermissionRuleset is now propagated to all InternalServers Endpoinds (also with the one with SecurityPolicy#None)
      
      * Fixed check_user_token method of InternalServer and activate_session of InternalSession to provide an encrypted password to the user-defined UserManager
      
      * Fixed bug in InternalServer when password is not provided
      Co-authored-by: default avatarMarco Panato <marco.panato_01@univr.it>
      ac4373e0
    • Fabian Beitler's avatar
      Add check if BrowseName already exists in ParentNode (#306) · 6a913f89
      Fabian Beitler authored
      * Add check if BrowseName already exists in ParentNode
      
      * Replace try-except blocks
      Log error in check_method, not everywhere else
      
      * Remove checks, because they are not cought if UaClient sends addNode request
      
      * Add test to try identifing if BrowseName of new Node is child of ParentNode. If yes, raise error.If Node is root, ignore test.
      
      * Fix tests, because multiple identical BrowseNames with the same parent Node is not allowed.
      
      * Remove exception and ignore root Node directly
      
      * Fix test by change Browsename
      
      * Tidy up codeformat
      
      * Try to get all sorts of NodeId child classes
      Delete root Node parent check
      
      * Remove empty aspace arg
      Check if Reference is a HasChild Reference or a subtype of it
      
      * Change pytest fixtures to function to not interfere with other tests
      Add check if Nodes are in the same Namespace as well
      Add duplicated browsename test
      
      * Seperate between root Node or others
      As well pushing test again...
      
      * Add more Meta information about Node positions
      
      * Add more Meta information about Nodes which seems to fail
      
      * Change Browsename since both Nodes contain the same Browsename
      
      * Replace dirty print with log
      
      * Register Namespace and import Nodes, because Tests have been interfered eachother
      
      * Try if test interferes with others
      
      * Add check to catch faulty preperation
      
      * Revert "Try if test interferes with others"
      
      This reverts commit b0ac20dc2e898a70a7dedb433da9b88dfb2dc3fd.
      Revert "Add check to catch faulty preperation"
      
      This reverts commit c0e5ff6c628076b99d972778fef2dd33132a29a2.
      
      * Change parallel jobs to stages, to reduce the load from Travis
      
      * Check if just this test causes problems
      
      * Fix typo in decorator of pytest
      
      * Change object name
      Delete some nodes after test, to be reusable in other
      Undo scope back to module
      
      * Add more Node cleanups
      
      * Add cleanup if xml was imported
      
      * Add more Node cleanups to avoid duplicated browsenames while import xml in tests
      
      * Fix list issue in test
      
      * Add even more necessary node cleanups
      
      * Add more Node cleanups
      
      * Remove faulty await
      
      * Add Node cleanup for imported xml Nodes
      
      * Try to get rid of imported nodes from xml
      
      * Getting rid of imported nodes from xml
      
      * Revert Travis
      Some PEP8 cleanups
      
      * Fix some changes back to master branch state
      
      * Revert pytest changes
      
      * Avoid multiple Nodes with same Browsename in test server
      
      * Add some more Node cleanups
      
      * Revert wrong changes
      
      * Switch try/except with if ... not/else statement
      
      * Remove empty parent check
      Replace explicit in _nodes[] check with a general in _aspace[] check
      6a913f89
    • Fabian Beitler's avatar
      Replace lxml with xml.etree.ElementTree (#346) · 0261d3ed
      Fabian Beitler authored
      * Replace lxml with xml.etree.ElementTree
      Replace iter() with iterfind()
      Remove lxml from setup.py
      Remove done FIXME Graffitti
      Remove lxml from travis
      
      * Replace wildcard iter via iter over every element to be python3.7 compatible
      0261d3ed
    • Joey Faulkner's avatar
      fix bug when sending event with no message · 9e5bc78a
      Joey Faulkner authored
      9e5bc78a
  8. 02 Dec, 2020 1 commit
  9. 25 Nov, 2020 2 commits
  10. 24 Nov, 2020 4 commits
  11. 23 Nov, 2020 1 commit
  12. 18 Nov, 2020 5 commits
  13. 24 Oct, 2020 1 commit
    • Julien Prigent's avatar
      When disconnecting the client, we have to cancel the uaclient _publish_task. · 9925e7d0
      Julien Prigent authored
      If we don't, this can prevent reconnection since the task is still running, and the previous task instance could be hanging on a previous request that has never completed (i.e network issue).
      
      Yes, there's no timeout on request coming from the publish_loop, so we
      must ensure the publish_task is always cancelled. Hence, the
      uaclient.close_session is always called on disconnect.
      9925e7d0
  14. 17 Oct, 2020 3 commits
  15. 16 Oct, 2020 1 commit
  16. 12 Oct, 2020 1 commit
    • Julien Prigent's avatar
      [ Security ] Password encoding bug and typing adjustement · d88ead00
      Julien Prigent authored
      String are immutable, therefore the password is never encoded and it
      throws this error when it's a string:
      
      ```
      ...../cryptography/hazmat/backends/openssl/backend.py in _load_key(self, openssl_read_func, convert_func, data, password)
         1284         userdata = self._ffi.new("CRYPTOGRAPHY_PASSWORD_DATA *")
         1285         if password is not None:
      -> 1286             utils._check_byteslike("password", password)
         1287             password_ptr = self._ffi.from_buffer(password)
         1288             userdata.password = password_ptr
      
      ...../cryptography/utils.py in _check_byteslike(name, value)
           36         memoryview(value)
           37     except TypeError:
      ---> 38         raise TypeError("{} must be bytes-like".format(name))
           39
           40
      
      TypeError: password must be bytes-like
      ```
      
      This patch fixes this as well as some typing inaccuracy introduced by #281
      d88ead00