This allows functions which use KJ_EXPECT and other test macros to be built without linking libkj-test, which pulls in a possibly-unwanted main() function.

In particular, afl-testcase.c++ could not link in -fno-exceptions mode without this change.

This is necessary to resolve an MSVC compile error in capability.h:

struct RawSchema;
extern const RawSchema NULL_INTERFACE_SCHEMA;
constexpr RawSchema const* schema = &NULL_INTERFACE_SCHEMA;

causes MSVC to complain that `schema` is not initialized by a constant
expression. Providing the full definition of RawSchema resolves this
spurious error, so I moved it out of generated-header-support.h and into
its own raw-schema.h.

Rename async-io.c++ to async-io-unix.c++, in preparation for factoring out common bits from async-io-win32.c++.

sed -r is a GNU extension, apparently, and due to stars aligning, this didn't break the build, but led to a totally bogus soname on non-GNU systems. Argh.

Apparently most systems (including GNU, undocumentedly) support -E instead, but perhaps the safest thing is to use a basic regular expression. Unfortunately on many systems ? and + are not available in basic regexes, but using * instead happens to work in our case.

Thanks to @neverpanic for catching this.

Fixes #210

Fixes a build error on GNU/Hurd.

A compatibility layer is provided to ease migration.

Failed to notice this because only the cmake build refers to it, and this only makes cmake fail on installation, which was never quite tested.

I've added -Wextra as well as removed some of the -Wno-* flags and fixed issues in the code.

Also fixed the cmake build to put user-defined flags after default flags so that they can be overridden.

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.

kernel32.dll and msvcrt.dll are the only remaining dependencies.

The command-line tools aren't built in LITE_MODE, so there's nothing
to create a symlink to.

To use, pass --disable-reflection to the configure script.

This produces a smaller runtime library. However, using it for this purpose is not recommended. The main purpose of lite mode is to define a subset of Cap'n Proto which might plausibly compile under MSVC. MSVC still lacks full support for constexpr and expression SFINAE; luckily, most of our use of these things relates to reflection, and not all users need reflection.

Cap'n Proto lite mode inherits its name from Protocol Buffers' lite mode. However, there are some key differences:

- Protobuf generated code included global constructors related to registering descriptors and extensions. For many people, this was the main reason to use lite mode: to get rid of these global constructors and achieve faster startup times. Cap'n Proto, on the other hand, never had global constructors in the first place.

- Schemas are actually still available in lite mode, though only in their raw (Cap'n Proto structure) form. Only the schema API (which wraps the raw schemas in a more convenient interface) and reflection API (which offers a convenient way to use the schemas) are unavailable.

- Lite mode is enabled in an application by defining CAPNP_LITE rather than by specifying an annotation in the schema file. This better-reflects real-world usage patterns, where you typically want to enable lite mode application-wide anyway.

- We do not build the lite mode library by default. You must request it by passing --disable-reflection to the configure script. Before you can do that, you must have a prebuilt Cap'n Proto compiler binary available, since the compiler can't be built without reflection.

- Relatedly, the lite mode library is built with the same name as the full library. This library is not intended to be installed. If anything it should be statically linked. But, mostly the option only exists on non-MSVC platform to give us a way to test that we haven't broken lite mode.

1) Apparently mixing member defaults and initializer lists is a C++14 thing.
2) My test script compiles with -Werror which errors out on calls to now-deprecated getDependency().

Protected by an ifdef to avoid tripping up non-autotools builds