Christian Neumüller

Seems complicated to fix as some converters are stateful and rely on e.g. `match` being called before `apply` (see for example https://github.com/Oberon00/luabind/blob/master/luabind/detail/policy.hpp#L217). However, to instantiate the appropriate converter, the out...

Have you read my above comment about stateful converters? The current test is insufficient to catch this error. Try using a class-type as out-value parameter and I bet you get...

Since C++17 will probably remove `auto_ptr` entirely and [MSVC 14 (2015) already allows users to disable it](http://blogs.msdn.com/b/vcblog/archive/2015/06/19/c-11-14-17-features-in-vs-2015-rtm.aspx), I consider this pull request a quite reasonable idea. However, your approach of...

I won't merge the pull request until the many #ifdefs are replaced with a `typedef` & wrapper `move` function, as I commented above. Also, because it seems that this pull...

> Here, I disagree, removing deprecation warnings is no cosmetic thing, especially when the c++14 norm will be the standard. Good point, I haven't thought about that. luabind is compiled...

Hi @Bertram25! It seems this PR comes from your master branch? Do you still intend to merge this? If so, please see the comments above. I think you address some...

These tests do expect luabind::error to be thrown and should not be changed (maybe they need some ifdefs though). It is worrying that they succeed with LUABIND_NO_EXCEPTIONS defined.

Hi, this looks indeed like a bug, and the warning about optimizing overflow could be related. If you can a segfaulting test executable with gdb and post a stack trace,...

Actually I'm not sure if `nil` is needed altogether; is there a case where it behaves differently from an uninitialized (i.e. default constructed) object?

A uninitialized object can also be used to assign nil, but only because it ultimately does `lua_rawgeti(LUA_REGISTRYINDEX, LUA_NOREF)` and there is nothing at `LUA_NOREF` (-2) in the registry. The [Lua...