By default, Better Enums will generate with an inaccessible (private or
deleted) default constructor. However, if the user defines
BETTER_ENUMS_DEFAULT_CONSTRUCTOR(Enum), the expansion of that macro
will be used instead. The macro approach was chosen because the
expansion can include access modifiers and fragments such as
"= default".
Resolves#10.
The user-supplied constant names are used to declare an internal enum
type. They end up sharing a namespace with some internal library
values. Those internal values had names not prefixed with underscores.
As a result, it was impossible to declare enum constants with these
names.
Prefixed the internal names with underscores.
VC2008 has two quirks. It generates linking errors if a copy
constructor is not explicitly defined on a Better Enum, and it has a
buggy interaction between the ternary operator and throw. This change
detects VC2008 and generates alternative code for that compiler.
Having an explicitly-defined copy constructor in a literal type appears
to cause an internal compiler error in g++4.7, and causes a spurious
compilation failure in g++4.8. For this reason, the copy constructor
generation is conditioned on the compiler.
The replacement code for the ternary operator is also generated
conditionally, because it uses an if-statement. The normal code has to
compile in a constexpr context, and so has to use the ternary operator
instead of the if-statement.
Resolves#6.
These are now only assumed in C++11 mode. long long is also assumed
only in C++11 mode for clang, which may make some programs that rely on
long long as an extension in C++98 fail with Better Enums. I will solve
that at a later date if it becomes a problem.
An alternative constant _size_constant is provided for use in C++98,
for example for declaring arrays.
Also renamed underlying_traits to integral_mapping.
Removed the function are_equal. Comparison is now done by converting
operands to their integral representation, and comparing those. Also
restored ordering of enum values along the same lines (according to
integral representation).
Before this change, in C++98 and C++11 "fast" mode, initializer
trimming was done "lazily" the first time _to_string or _names was
called. To make performance more "predictable", an object with static
storage is now used to force initializaton during program start-up,
when static object constructors are called.
The benefit of this change is very debatable. I had to give the static
object static linkage to avoid duplicate symbols, so there is a copy
now in each translation unit. I hope this does not increase code size
too much in realistic scenarios.
Lazy initialization checks are still performed and cannot be removed,
because other objects with static storage may try to use an enum from
their constructors before the enum's initialization is forced.
When compile-time stringized constant name trimming is disabled (off by
default), trimming happens "lazily" - the first time the user calls a
function such as _to_string, the function allocates space for trimmed
constant names and trims them there.
With this change, space is reserved statically in a writeable char
array, and trimming happens in that array instead.
To avoid paying the huge penalty of including iostream and string for
users that don't need those headers, and to avoid creating a second,
optional header file, I resorted to defining the operators as templates
to prevent type checking until the user tries to actually use them. The
stream types and strings are wrapped in a metafunction that depends on
the template parameter. This is basically a hack, but it seems to work.
With this change, the underlying type can be a non-integral type that
provides conversions to and from an integral type. See the test at
test/cxxtest/underlying.h for some examples - though they are more
verbose than strictly necessary, for testing needs.
Move constructors in underlying types are not supported. It has been
difficult so far to get constexpr code not to select the move
constructor, which is generally not constexpr, for various operations.
support to VC++.
The unit test is currently not being run on VC++ due to a problem with CxxTest,
Cygwin, and paths. However, the examples are being compiled and having their
output checked, and the multiple translation unit test is being run.
Running "(cd test ; ./test.py)" should now run tests only using the default
compiler, on a Unix-like system. test.py --all runs tests on the full array of
compilers that I have installed and symlinked on my development machines.
The documentation is now generated from markdown. Samples are generated from the
tutorial pages. Testing is done by a Python script which runs the tests for a
large number of compilers.
This version is not very developer-friendly - the Python scripts need ways of
limiting what compilers they try to run. If you don't have 15 compilers
installed, you won't be able to run the tests in this commit. Fix coming soon.
This patch contains several minor changes.
- Eliminated the use of a deleted constructor in C++11. C++98 private default
constructor is sufficient.
- Eliminated old namespace _enum and merged it with namespace better_enums.
- Prefixed size_t with std:: to comply with standards more strictly.
- Shortened feature control macros by deleting the word "FORCE".
Also moved make_macros.py.
This makes C++98 and C++11 Better Enums fully compatible by default. If the user
defines BETTER_ENUMS_FORCE_STRICT_CONVERSION before including enum.h, it is
necessary to prefix enum constants in switch cases with '+', but Better Enums
are not implicitly convertible to integers.
These modifications ensure enum.h can be used in a wider
selection of end user projects without triggering warnings.
GCC 4.9.2 was used with the following warning flags set:
-Wall -Wextra -Wshadow -Weffc++ -Wno-unused-parameter
-Wno-unused-local-typedefs -Wno-long-long -Wstrict-aliasing
-Werror -pedantic -std=c++1y -Wformat=2 -Wmissing-include-dirs
-Wsync-nand -Wuninitialized -Wconditionally-supported -Wconversion
-Wuseless-cast -Wzero-as-null-pointer-constant
This commit includes the modifications required to enable successful
use of enum.h via both the "test" and "example" directories.
enum.h tries to automatically detect whether it is running with C++11 support.
If not, it emits alternative code that is supposed to work on compilers
supporting C++98 and variadic macros. This code is largely interface-compatible
with the C++11 code, with the following semantic differences:
- No compile-time stringization. This is done upon first use of a function other
than to_integral.
- Implicit conversion to integral types. This is due to the lack of enum class
support.
- The values _name, _names, _values are replaced with functions _name_, _names_,
_values_.
Each Better Enum now has an internal enum class type to which it is convertible,
instead of being convertible to the regular enum that defines its constants.
switch statements are compiled at the enum class type. This comes at the price
of the user having to type +Enum::Constant instead of Enum::Constant in cases,
in order to trigger an explicit promotion of the pre-C++11 enum to Better Enum,
so it can then be implicitly converted to the enum class.
The remaining "hole" is that direct references to constants (Enum::Constant) are
still implicitly convertible to integral types, because they have naked
pre-C++11 enum type.
These return values of an optional type better_enums::optional<T>. This type is
defined in the spirit of boost::optional<T> and std::optional<T>, but is easy to
manipulate at compile time. Two additional macros BETTER_ENUMS_USE_OPTIONAL and
BETTER_ENUMS_EXTRA_INCLUDE are honored, whose intent is for the user to be able
to inject an alternative option type. However, there are currently no viable
alternatives. boost::optional<T> does not play well with constexpr, and I failed
to make the code compile with std::optional<T>. I did not try very hard. I
intend to support std::optional<T> in the future. Perhaps it will be the
default, when available.
A Better Enum is normally implicitly convertible to its internal enum type,
which makes it then implicitly convertible to an integer as well. The former
conversion is necessary for Better Enums to be usable in switch statements.
This change makes it possible to define BETTER_ENUMS_SAFER_SWITCH, which makes
Better Enums convert to an enum class, preventing the implicit conversion to
integers. The drawback is that switch cases have to be written as
case Enum::_Case::A:
instead of
case Enum::A:
The interface is now uniformly constexpr, including to_string and the _names
iterable. Without the weak symbol, the remaining code is also entirely standard
C++.
The compile-time string trimming code in this commit has a negative impact on
performance. The performance test is now twice as slow as including <iostream>,
whereas before it was faster. That test declares an excessive number of enums,
though, so perhaps in typical usage, and with some future optimizations, the
impact will not be so significant.
There may be other ways to solve this, such as providing a version of the macro
that does not trim strings at compile time, but only checks if they need
trimming. If some string does need trimming, that macro would fail a
static_assert and ask the user to use the slow macro.
