Made the main Boost Build test runner script correctly report unhandled and unexpected exceptions raised from specific test scripts. Before, such an exception would terminate the whole test run while now it gets cleanly reported as just another test failure. This makes Boost library regression test results correctly report any Python errors (e.g. using a Python feature not supported by the used Python version) which previously developers had no built-in way of getting feedback about.
(No files)
Note:
See TracChangeset
for help on using the changeset viewer.