- 17 Aug, 2015 9 commits
-
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
Fix `has_name`
-
ufownl authored
Make `has_name` support scalar types.
-
- 09 Aug, 2015 2 commits
-
-
Dominik Charousset authored
-
Matthias Vallentin authored
This used to be CAF_FAIL, but in the process of streamlining the semantics of this macro to be compatible with Boost.Test, we've lost the ability to express a check which is known to fail. The macro CAF_CHECK_FAIL brings back this capability. The suffix _FAIL after CHECK hints that this is a variation of CAF_CHECK.
-
- 08 Aug, 2015 3 commits
-
-
Matthias Vallentin authored
-
Matthias Vallentin authored
Resolves #332.
-
Matthias Vallentin authored
The macro CAF_REQUIRE and CAF_FAIL now both terminate test execution. The only difference between the two is that CAF_REQUIRE takes a condition and CAF_FAIL a termination erro message. This behavior is similar to Boost.Test. Resolves #330.
-
- 06 Aug, 2015 3 commits
-
-
Marian Triebe authored
GCC can compile code which use `thread_local` on OSX, but generates faulty code. The generated code leads to a heap-use-after-free.
-
Joseph Noir authored
-
Dominik Charousset authored
Fix build on MSVC
-
- 05 Aug, 2015 1 commit
-
-
Marian Triebe authored
This commit includes a fix for `accept` of the socket API with MSVC. Examples are now working, but use the `trivial match case`, the `advanced match case` is not available on MSVC due to internal compiler errors. The `thread_local` keyword is used on MSVC and GCC, while Clang does not support it for now. This commit also fixes minor warnings with the MSVC compiler and removes the `pragma comment` for linking libraries with MSVC.
-
- 03 Aug, 2015 4 commits
-
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
- 30 Jul, 2015 2 commits
-
-
Dominik Charousset authored
Since the future development of BASP is unclear at this time, it makes no sense to introduce breaking changes to the protocol that are likely to get broken again soon.
-
Dominik Charousset authored
-
- 28 Jul, 2015 5 commits
-
-
Matthias Vallentin authored
-
Marian Triebe authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
- 27 Jul, 2015 10 commits
-
-
Dominik Charousset authored
When building CAF with logging enabled, some brokers can be destroyed after the logging singleton has been destroyed, resulting in the (lazy) construction of another logger which then causes a memory leak.
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
Dominik Charousset authored
-
- 25 Jul, 2015 1 commit
-
-
f0rmiga authored
-