I'm not really sure how Cap'n Proto has worked at all on GCC 5. For some reason this only ended up being tickled after e6e29122.

Fixes #442.

Since this header is included by everyone, and units.h has lots of templates, this seems like it could significantly improve build times.

The problem with kj::range() is that if the inputs are not recognized as the same type, the typechecker can get annoyed.

MSVC refuses the following code:

struct Foo { Foo(Foo&) {} };
Foo foo;
[foo] {}();

because it only seems to want to consider const copy constructors when
capturing objects by value in lambda capture lists.

MSVC cannot compile the overloaded member function SFINAE technique used
in Kind_ and FromAny_. It has an easier time with std::void_t-style
SFINAE.

Fix JSON handler for DynamicStruct. The handler needs to be told the struct's schema to construct an orphan.

Also improve error message when addTypeHandler() is given a handler for a dynamic type but no specific schema is specified. We could support this eventually but we don't presently.

Maybe::emplace() should return the newly-constructed value so that you don't have to immediately KJ_ASSERT_NONNULL it.

- Reduce noise.
- Trace through async callbacks.

https://llvm.org/bugs/show_bug.cgi?id=22354

At the time DisallowConstCopy was introduced, GCC 4.7 and Clang 3.2 refused to allow the non-const copy constructors to be defaulted inline. This must have been concluded to be a compiler bug, as GCC 4.8 and Clang 3.4 seem fine with it. So, we can clean this up.

Meanwhile, the above-linked Clang bug triggered when a DisallowConstCopy derivative is captured by value in a C++14 generic lambda. capnp::CallContext in particular tended to be hit by this a lot. The bug only affects non-default (or out-of-line default) copy constructors, and thus is conveniently sidestepped by this change.

Work around incompatibility between various C++ standard libs in C++14 mode on some Linux distros, including Ubuntu 14.04.

Distinguishing between "local bugs" and "preconditions" was proving difficult in practice, because a precondition failure in one function may very well indicate a bug in a calling function, but the exception may be thrown through that function, thus when caught the classification is nonsensical. The distinction also was not as useful as imagined. So, I eliminated this distinction.

Add asBytes() and asChars() methods to array classes to reinterpret-cast to bytes / chars, since this happens all the time and is otherwise a huge pain.

Use the new methods in a bunch of places.

The project file still only compiles a test binary, but it should be easy to separate out a library project from here.

Thanks again to Bryan Boreham <bjboreham@gmail.com> for much help getting this working.

Thanks to Bryan Boreham <bjboreham@gmail.com> for much help getting this started.

This prevents the compiler from reporting warnings in these headers while compiling application code. Hopefully this will stem the never-ending stream of complaints from people who enable pedantic warnings.

Add __forceinline for MSVC

For portions currently copyright by Kenton (most of it), transfer copyright to Sandstorm Development Group, Inc. (Kenton's company).

The license change is practically meaningless, as MIT and BSD 2-clause are legally equivalent. However, the BSD 2-clause license is sometimes confused for its ugly siblings, BSD 3-clause and BSD 4-clause. The MIT license is more immediately recognizeable for what it is.

Rémy Blank and Jason Choy (the two non-trivial contributors) are on record as approving this change:

https://groups.google.com/d/msg/capnproto/xXDd2HUOCcc/gbe_COIuXKYJ