Changelog for package launch

0.6.0 (2018-08-20)

• Added a way to include other Python launch files (#122)
  • Signed-off-by: William Woodall <william@osrfoundation.org>
• Implemented the concept of Action conditions (#121)
  • Signed-off-by: William Woodall <william@osrfoundation.org>
• Added IncludeLaunchDescription action (#120)
  • fixes #115
  • Signed-off-by: William Woodall <william@osrfoundation.org>
• Contributors: William Woodall

0.5.2 (2018-07-17)

• Made a change to avoid reentrancy of signal handlers (#99)
• Ignored warning for builtins A003 (#100)
• Fixed exception when launch process with environment variables (#96)
• Contributors: Shane Loretz, William Woodall, dhood

0.5.1 (2018-06-27)

• Changed the behavior when signaling SIGINT to subprocesses on Windows, where it now does SIGTERM instead, because SIGINT causes a ValueError about SIGINT being an unsupported signal number. (#94)
• Fixed a bug by avoiding reentrancy in the SIGINT signal handler. (#92)
• Various Windows fixes. (#87)
  • LaunchService.run() now returns non-0 when there are exceptions in coroutines.
  • Updated launch_counters.py example for Windows.
  • Fixed a bug that would cause mismatched asyncio loops in some futures.
  • Addressed the fact that signal.SIGKILL doesn't exist on Windows, so emulate it in our Event.
  • Fixed an issue that resulted in spurious asyncio errors in LaunchService test.
• Contributors: William Woodall, dhood

0.5.0 (2018-06-19)

• Fixed a bug where unclosed asyncio loops caused a traceback on the terminal on exit, but only in Python 3.5 (#85)
• Changed to use variable typing in comments to support python 3.5 (#81)
• New launch API (#74)
  • See pull request for more details and links to architecture documentation and the design doc.
• Moved launch source files into launch.legacy namespace (#73)
  • This was in preparation for the new launch API.
• [for launch.legacy] fixed a flake8 warning (#72)
• [for launch.legacy] set zip_safe to avoid warning during installation (#71)
• [for launch.legacy] Fix hang on keyboard interrupt (#69)
  • When keyboard interrupt exception occurs loop.run_forever is called. But there is no loop.stop call. This causes a hang.
• Contributors: Devin, Dirk Thomas, William Woodall, dhood