Redstone1024/Redcraft
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
Files
6cc31f660c923ac3928f8bd91a63a0c1a523b632
Redcraft/Redcraft.Utility/Source/Public/Strings/Formatting.h

3184 lines
86 KiB
C++
Raw Blame History

#pragma once
#include "CoreTypes.h"
#include "TypeTraits/TypeTraits.h"
#include "Templates/Utility.h"
#include "Ranges/Utility.h"
#include "Ranges/View.h"
#include "Algorithms/Basic.h"
#include "Containers/StaticArray.h"
#include "Containers/Array.h"
#include "Numerics/Bit.h"
#include "Numerics/Math.h"
#include "Strings/Char.h"
#include "Miscellaneous/AssertionMacros.h"
#include <charconv>
#pragma warning(push)
#pragma warning(disable: 4244)
NAMESPACE_REDCRAFT_BEGIN
NAMESPACE_MODULE_BEGIN(Redcraft)
NAMESPACE_MODULE_BEGIN(Utility)
/** A concept specifies a type is a format description string context. */
template <typename T, typename CharType = char>
concept CFormatStringContext = CInputRange<T> && CCharType<CharType> && CSameAs<TRangeElement<T>, CharType>
&& requires(T& Context, TRangeIterator<T> Iter, size_t Index)
{
/** Set the iterator of the context. */
Context.AdvanceTo(MoveTemp(Iter));
/** @return The next automatic index. */
{ Context.GetNextIndex() } -> CSameAs<size_t>;
/** @return true if the manual index is valid. */
{ Context.CheckIndex(Index) } -> CBooleanTestable;
};
/** A concept specifies a type is a format output context. */
template <typename T, typename CharType = char>
concept CFormatObjectContext = COutputRange<T, CharType> && CCharType<CharType>
&& requires(T& Context, TRangeIterator<T> Iter, size_t Index)
{
/** Set the iterator of the context. */
Context.AdvanceTo(MoveTemp(Iter));
/** Visit the format argument by index, and the argument is always like const&. */
{ Context.Visit([](const auto&) { return 0; }, Index) } -> CIntegral;
/** Visit the format argument by index, and the argument is always like const&. */
{ Context.template Visit<int>([](const auto&) { return 0; }, Index) } -> CIntegral;
};
/**
* A template class that defines the formatting rules for a specific type.
*
* @tparam T - The type of object being formatted.
* @tparam CharType - The character type of the formatting target.
*/
template <typename T, CCharType CharType = char> requires (CSameAs<TRemoveCVRef<T>, T>)
class TFormatter
{
public:
static_assert(sizeof(T) == -1, "The type is not formattable.");
FORCEINLINE constexpr TFormatter() = delete;
FORCEINLINE constexpr TFormatter(const TFormatter&) = delete;
FORCEINLINE constexpr TFormatter(TFormatter&&) = delete;
FORCEINLINE constexpr TFormatter& operator=(const TFormatter&) = delete;
FORCEINLINE constexpr TFormatter& operator=(TFormatter&&) = delete;
/**
* Parses the format description string from the context.
* Assert that the format description string is valid.
*
* @return The iterator that points to the first unmatched character.
*/
template <CFormatStringContext<CharType> CTX>
FORCEINLINE constexpr TRangeIterator<CTX> Parse(CTX& Context);
/**
* Formats the object and writes the result to the context.
* Do not assert that the output range is always large enough, and return directly if it is insufficient.
* Specify, unlike visiting arguments from the context which is always like const&,
* the object argument is a forwarding reference.
*
* @return The iterator that points to the next position of the output.
*/
template <typename U, CFormatObjectContext<CharType> CTX> requires (CSameAs<TRemoveCVRef<T>, U>)
FORCEINLINE constexpr TRangeIterator<CTX> Format(U&& Object, CTX& Context) const;
};
NAMESPACE_PRIVATE_BEGIN
template <typename I, typename S, typename... Ts>
class TFormatStringContext
{
public:
FORCEINLINE constexpr TFormatStringContext(I InFirst, S InLast) : First(MoveTemp(InFirst)), Last(InLast), AutomaticIndex(0) { }
NODISCARD FORCEINLINE constexpr I Begin() requires (!CCopyable<I>) { return MoveTemp(First); }
NODISCARD FORCEINLINE constexpr I Begin() const requires ( CCopyable<I>) { return First; }
NODISCARD FORCEINLINE constexpr S End() const { return Last; }
NODISCARD FORCEINLINE constexpr size_t Num() const requires (CSizedSentinelFor<S, I>) { return Last - First; }
NODISCARD FORCEINLINE constexpr bool IsEmpty() const { return First == Last; }
FORCEINLINE constexpr void AdvanceTo(I Iter) { First = MoveTemp(Iter); }
NODISCARD FORCEINLINE constexpr size_t GetNextIndex()
{
bool bIsValid = AutomaticIndex < sizeof...(Ts) && AutomaticIndex != INDEX_NONE;
checkf(bIsValid, TEXT("Illegal automatic indexing. Already entered manual indexing mode."));
if (!bIsValid) return INDEX_NONE;
return AutomaticIndex++;
}
NODISCARD FORCEINLINE constexpr bool CheckIndex(size_t Index)
{
bool bIsValid = AutomaticIndex == 0 || AutomaticIndex == INDEX_NONE;
checkf(bIsValid, TEXT("Illegal manual indexing. Already entered automatic indexing mode."));
if (!bIsValid) return false;
AutomaticIndex = INDEX_NONE;
return Index < sizeof...(Ts);
}
private:
NO_UNIQUE_ADDRESS I First;
NO_UNIQUE_ADDRESS S Last;
size_t AutomaticIndex = 0;
};
static_assert(CFormatStringContext<TFormatStringContext<IInputIterator<char>, ISentinelFor<IInputIterator<char>>>>);
template <typename I, typename S, typename... Ts>
class TFormatObjectContext
{
public:
FORCEINLINE constexpr TFormatObjectContext(I InFirst, S InLast, Ts&... Args) : First(MoveTemp(InFirst)), Last(InLast), ArgsTuple(Args...) { }
NODISCARD FORCEINLINE constexpr I Begin() requires (!CCopyable<I>) { return MoveTemp(First); }
NODISCARD FORCEINLINE constexpr I Begin() const requires ( CCopyable<I>) { return First; }
NODISCARD FORCEINLINE constexpr S End() const { return Last; }
NODISCARD FORCEINLINE constexpr size_t Num() const requires (CSizedSentinelFor<S, I>) { return Last - First; }
NODISCARD FORCEINLINE constexpr bool IsEmpty() const { return First == Last; }
FORCEINLINE constexpr void AdvanceTo(I Iter) { First = MoveTemp(Iter); }
template <typename F> requires (((sizeof...(Ts) >= 1) && ... && CInvocable<F&&, Ts&>) && CCommonReference<TInvokeResult<F&&, const Ts&>...>)
FORCEINLINE constexpr decltype(auto) Visit(F&& Func, size_t Index) const { return ArgsTuple.Visit(Forward<F>(Func), Index); }
template <typename Ret, typename F> requires ((sizeof...(Ts) >= 1) && ... && CInvocableResult<Ret, F&&, const Ts&>)
FORCEINLINE constexpr Ret Visit(F&& Func, size_t Index) const { return ArgsTuple.template Visit<Ret>(Forward<F>(Func), Index); }
private:
NO_UNIQUE_ADDRESS I First;
NO_UNIQUE_ADDRESS S Last;
TTuple<Ts&...> ArgsTuple;
};
static_assert(CFormatObjectContext<TFormatObjectContext<IOutputIterator<char&>, ISentinelFor<IOutputIterator<char&>>, int>>);
NAMESPACE_PRIVATE_END
/** A concept specifies a type is formattable by the 'Algorithms::Format()'. */
template <typename T, typename CharType = char>
concept CFormattable = CCharType<CharType> && CSemiregular<TFormatter<TRemoveCVRef<T>, CharType>>
&& requires(TFormatter<TRemoveCVRef<T>, CharType>& Formatter, T& Object,
NAMESPACE_PRIVATE::TFormatStringContext< IInputIterator<CharType >, ISentinelFor< IInputIterator<CharType >>, T> FormatStringContext,
NAMESPACE_PRIVATE::TFormatObjectContext<IOutputIterator<CharType&>, ISentinelFor<IOutputIterator<CharType&>>, T> FormatObjectContext)
{
{ Formatter.Parse ( FormatStringContext) } -> CSameAs< IInputIterator<CharType >>;
{ Formatter.Format(Object, FormatObjectContext) } -> CSameAs<IOutputIterator<CharType&>>;
};
NAMESPACE_BEGIN(Algorithms)
/**
* Formats the objects and writes the result to the output range.
* Assert that the format description string is valid.
* If the output range is insufficient, return directly without asserting.
*
* @param Output - The output range to write the result.
* @param Fmt - The format description string.
* @param Args - The objects to format.
*
* @return The iterator that points to the next position of the output.
*/
template <CInputRange R1, COutputRange<TRangeElement<R1>> R2, CFormattable<TRangeElement<R1>>... Ts> requires (CBorrowedRange<R2>)
FORCEINLINE constexpr TRangeIterator<R2> Format(R2&& Output, R1&& Fmt, Ts&&... Args)
{
if constexpr (CSizedRange<R1&>)
{
checkf(Algorithms::Distance(Fmt) >= 0, TEXT("Illegal range. Please check Algorithms::Distance(Fmt)."));
}
if constexpr (CSizedRange<R2&>)
{
checkf(Algorithms::Distance(Output) >= 0, TEXT("Illegal range. Please check Algorithms::Distance(Output)."));
}
using FCharType = TRangeElement<R1>;
using FCharTraits = TChar<FCharType>;
using FFormatStringContext = NAMESPACE_PRIVATE::TFormatStringContext<TRangeIterator<R1>, TRangeSentinel<R1>, Ts...>;
using FFormatObjectContext = NAMESPACE_PRIVATE::TFormatObjectContext<TRangeIterator<R2>, TRangeSentinel<R2>, Ts...>;
FFormatStringContext FormatStringContext(Ranges::Begin(Fmt ), Ranges::End(Fmt ));
FFormatObjectContext FormatObjectContext(Ranges::Begin(Output), Ranges::End(Output), Args...);
auto FmtIter = Ranges::Begin(FormatStringContext);
auto FmtSent = Ranges::End (FormatStringContext);
auto OutIter = Ranges::Begin(FormatObjectContext);
auto OutSent = Ranges::End (FormatObjectContext);
// If the output range is insufficient.
if (OutIter == OutSent) UNLIKELY return OutIter;
// For each character in the format string.
for (FCharType Char; FmtIter != FmtSent; ++FmtIter)
{
Char = *FmtIter;
// If the character may be a replacement field.
if (Char == LITERAL(FCharType, '{'))
{
if (++FmtIter == FmtSent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Unmatched '{' in format string."));
break;
}
Char = *FmtIter;
// If the character just an escaped '{'.
if (Char == LITERAL(FCharType, '{'))
{
if (OutIter == OutSent) UNLIKELY return OutIter;
*OutIter++ = LITERAL(FCharType, '{');
continue;
}
// If available replacement fields.
if constexpr (sizeof...(Ts) >= 1)
{
size_t Index;
// If the replacement field has a manual index.
if (Char != LITERAL(FCharType, ':') && Char != LITERAL(FCharType, '}'))
{
Index = 0;
bool bIsValid = true;
do
{
const uint Digit = FCharTraits::ToDigit(Char);
if (Digit >= 10) bIsValid = false;
Index = Index * 10 + Digit;
if (++FmtIter == FmtSent) UNLIKELY break;
Char = *FmtIter;
}
while (Char != LITERAL(FCharType, ':') && Char != LITERAL(FCharType, '}'));
// If the index string contains illegal characters or the index is out of range.
if (!bIsValid || !FormatStringContext.CheckIndex(Index)) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the replacement field."));
break;
}
}
// If the replacement field need automatic indexing.
else
{
Index = FormatStringContext.GetNextIndex();
if (Index == INDEX_NONE)
{
checkf(false, TEXT("Illegal index. Please check the replacement field."));
break;
}
}
// Jump over the ':' character.
if (Char == LITERAL(FCharType, ':')) ++FmtIter;
if (FmtIter == FmtSent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
break;
}
Char = *FmtIter;
ForwardAsTuple(Forward<Ts>(Args)...).Visit([&]<typename T>(T&& Arg)
{
TFormatter<TRemoveCVRef<T>, FCharType> Formatter;
if (Char != LITERAL(FCharType, '}'))
{
FormatStringContext.AdvanceTo(MoveTemp(FmtIter));
// Parse the format description string.
FmtIter = Formatter.Parse(FormatStringContext);
if (FmtIter == FmtSent || *FmtIter != LITERAL(FCharType, '}')) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return;
}
}
FormatObjectContext.AdvanceTo(MoveTemp(OutIter));
// Format the object and write the result to the context.
OutIter = Formatter.Format(Forward<T>(Arg), FormatObjectContext);
}
, Index);
}
else
{
checkf(false, TEXT("Illegal index. Please check the replacement field."));
break;
}
}
// If the character just an escaped '}'.
else if (Char == LITERAL(FCharType, '}'))
{
// Confirm the character is an escaped '}'.
if (++FmtIter != FmtSent && *FmtIter == LITERAL(FCharType, '}'))
{
if (OutIter == OutSent) UNLIKELY return OutIter;
*OutIter++ = LITERAL(FCharType, '}');
continue;
}
checkf(false, TEXT("Illegal format string. Missing '{' in format string."));
break;
}
// If the output range is insufficient.
else if (OutIter == OutSent) UNLIKELY return OutIter;
// If the character is not a replacement field.
else *OutIter++ = Char;
}
return OutIter;
}
/**
* Formats the objects and writes the result to the output range.
* Assert that the format description string is valid.
* If the output range is insufficient, return directly without asserting.
*
* @param OutputFirst - The iterator of the output range to write the result.
* @param OutputLast - The sentinel of the output range to write the result.
* @param FmtFirst - The iterator of the format description string.
* @param FmtLast - The sentinel of the format description string.
* @param Args - The objects to format.
*
* @return The iterator that points to the next position of the output.
*/
template <CInputIterator I1, CSentinelFor<I1> S1, COutputIterator<TIteratorElement<I1>> I2, CSentinelFor<I2> S2, CFormattable<TIteratorElement<I1>>... Ts>
FORCEINLINE constexpr I2 Format(I2 OutputFirst, S2 OutputLast, I1 FmtFirst, S1 FmtLast, Ts&&... Args)
{
if constexpr (CSizedSentinelFor<S1, I1>)
{
checkf(FmtFirst - FmtLast <= 0, TEXT("Illegal range iterator. Please check HaystackFirst <= HaystackLast."));
}
if constexpr (CSizedSentinelFor<S2, I2>)
{
checkf(OutputFirst - OutputLast <= 0, TEXT("Illegal range iterator. Please check NeedleFirst <= NeedleLast."));
}
return Algorithms::Format(Ranges::View(MoveTemp(OutputFirst), OutputLast), Ranges::View(MoveTemp(FmtFirst), FmtLast), Forward<Ts>(Args)...);
}
NAMESPACE_END(Algorithms)
/**
* A formatter for null-terminated string.
*
* The syntax of format specifications is:
*
* [Fill And Align] [Width] [Precision] [Type] [!] [?]
*
* 1. The fill and align part:
*
* [Fill Character] <Align Option>
*
* i. Fill Character: The character is used to fill width of the object. It is optional and cannot be '{' or '}'.
* It should be representable as a single unicode otherwise it is undefined behavior.
*
* ii. Align Option: The character is used to indicate the direction of alignment.
*
* - '<': Align the formatted argument to the left of the available space
* by inserting n fill characters after the formatted argument.
* This is default option.
* - '^': Align the formatted argument to the center of the available space
* by inserting n fill characters around the formatted argument.
* If cannot absolute centering, offset to the left.
* - '>': Align the formatted argument ro the right of the available space
* by inserting n fill characters before the formatted argument.
*
* 2. The width part:
*
* - 'N': The number is used to specify the minimum field width of the object.
* N should be an unsigned non-zero decimal number.
* - '{N}': Dynamically determine the minimum field width of the object.
* N should be a valid index of the format integral argument.
* N is optional, and the default value is automatic indexing.
*
* 3. The precision part:
*
* - '.N': The number is used to specify the maximum field width of the object.
* N should be an unsigned non-zero decimal number.
* - '.{N}': Dynamically determine the maximum field width of the object.
* N should be a valid index of the format integral argument.
* N is optional, and the default value is automatic indexing.
*
* 4. The type indicator part:
*
* - none: Indicates the as-is formatting.
* - 'S': Indicates the as-is formatting.
* - 's': Indicates lowercase formatting.
*
* 5. The case indicators part:
*
* - '!': Indicates capitalize the entire string.
*
* 6. The escape indicators part:
*
* - '?': Indicates the escape formatting.
*
*/
template <CCharType T>
class TFormatter<T*, T>
{
private:
using FCharType = T;
using FCharTraits = TChar<FCharType>;
using FFillCharacter = TStaticArray<FCharType, FCharTraits::MaxCodeUnitLength>;
public:
template <CFormatStringContext<FCharType> CTX>
constexpr TRangeIterator<CTX> Parse(CTX& Context)
{
auto Iter = Ranges::Begin(Context);
auto Sent = Ranges::End (Context);
// Set the default values.
{
FillUnitLength = 1;
FillCharacter[0] = LITERAL(FCharType, ' ');
AlignOption = LITERAL(FCharType, '<');
MinFieldWidth = 0;
MaxFieldWidth = -1;
bDynamicMin = false;
bDynamicMax = false;
bLowercase = false;
bUppercase = false;
bEscape = false;
}
// If the format description string is empty.
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
FCharType Char = *Iter; ++Iter;
// Try to parse the fill and align part.
// This code assumes that the format string does not contain multi-unit characters, except for fill character.
// If the fill character is multi-unit.
if (!FCharTraits::IsValid(Char))
{
FillUnitLength = 1;
FillCharacter[0] = Char;
while (true)
{
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
// If the fill character ends.
if (FillUnitLength == FCharTraits::MaxCodeUnitLength || FCharTraits::IsValid(Char)) break;
FillCharacter[FillUnitLength++] = Char;
}
if (Char != LITERAL(FCharType, '<') && Char != LITERAL(FCharType, '^') && Char != LITERAL(FCharType, '>')) UNLIKELY
{
checkf(false, TEXT("Illegal format string. The fill character is not representable as a single unicode."));
return Iter;
}
AlignOption = Char;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
// If the fill character is single-unit.
else do
{
if (Iter == Sent) break;
// If the fill character is specified.
if (*Iter == LITERAL(FCharType, '<') || *Iter == LITERAL(FCharType, '^') || *Iter == LITERAL(FCharType, '>'))
{
FillUnitLength = 1;
FillCharacter[0] = Char;
Char = *Iter; ++Iter;
}
// If the fill character is not specified and the align option is not specified.
else if (Char != LITERAL(FCharType, '<') && Char != LITERAL(FCharType, '^') && Char != LITERAL(FCharType, '>')) break;
AlignOption = Char;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
while (false);
// Try to parse the width part.
{
if (Char == LITERAL(FCharType, '{'))
{
bDynamicMin = true;
MinFieldWidth = INDEX_NONE;
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
}
if ((bDynamicMin || Char != LITERAL(FCharType, '0')) && FCharTraits::IsDigit(Char))
{
MinFieldWidth = FCharTraits::ToDigit(Char);
while (true)
{
if (Iter == Sent)
{
checkf(!bDynamicMin, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
if (!bDynamicMin && *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
const uint Digit = FCharTraits::ToDigit(Char);
if (Digit >= 10) break;
MinFieldWidth = MinFieldWidth * 10 + Digit;
}
}
if (bDynamicMin)
{
if (Char != LITERAL(FCharType, '}')) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
do
{
// Try to automatic indexing.
if (MinFieldWidth == INDEX_NONE)
{
MinFieldWidth = Context.GetNextIndex();
if (MinFieldWidth == INDEX_NONE) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the field width."));
}
else break;
}
// Try to manual indexing.
else if (!Context.CheckIndex(MinFieldWidth)) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the field width."));
}
else break;
bDynamicMin = false;
MinFieldWidth = 0;
}
while (false);
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
}
// Try to parse the precision part.
if (Char == LITERAL(FCharType, '.'))
{
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing precision in format string."));
return Iter;
}
Char = *Iter; ++Iter;
if (Char == LITERAL(FCharType, '{'))
{
bDynamicMax = true;
MaxFieldWidth = INDEX_NONE;
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
}
if ((bDynamicMax || Char != LITERAL(FCharType, '0')) && FCharTraits::IsDigit(Char))
{
MaxFieldWidth = FCharTraits::ToDigit(Char);
while (true)
{
if (Iter == Sent)
{
checkf(!bDynamicMax, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
if (!bDynamicMax && *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
const uint Digit = FCharTraits::ToDigit(Char);
if (Digit >= 10) break;
MaxFieldWidth = MaxFieldWidth * 10 + Digit;
}
}
else if (!bDynamicMax)
{
checkf(false, TEXT("Illegal format string. Missing precision in format string."));
return Iter;
}
if (bDynamicMax)
{
if (Char != LITERAL(FCharType, '}')) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
do
{
// Try to automatic indexing.
if (MaxFieldWidth == INDEX_NONE)
{
MaxFieldWidth = Context.GetNextIndex();
if (MaxFieldWidth == INDEX_NONE) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the precision."));
}
else break;
}
// Try to manual indexing.
else if (!Context.CheckIndex(MaxFieldWidth)) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the precision."));
}
else break;
bDynamicMax = false;
MaxFieldWidth = -1;
}
while (false);
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
}
// Try to parse the type indicators part.
switch (Char)
{
case LITERAL(FCharType, 's'): bLowercase = true; break;
default: { }
}
switch (Char)
{
case LITERAL(FCharType, 'S'):
case LITERAL(FCharType, 's'): if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter; Char = *Iter; ++Iter; break;
default: { }
}
// Try to parse the case indicators part.
if (Char == LITERAL(FCharType, '!'))
{
bUppercase = true;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
// Try to parse the escape indicators part.
if (Char == LITERAL(FCharType, '?'))
{
bEscape = true;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
template <CFormatObjectContext<FCharType> CTX>
constexpr TRangeIterator<CTX> Format(const FCharType* Object, CTX& Context) const
{
auto Iter = Ranges::Begin(Context);
auto Sent = Ranges::End (Context);
size_t MinDynamicField = MinFieldWidth;
size_t MaxDynamicField = MaxFieldWidth;
// Visit the dynamic width argument.
if (bDynamicMin)
{
MinDynamicField = Context.Visit([]<typename U>(U&& Value) -> size_t
{
using FDecayU = TRemoveCVRef<U>;
if constexpr (CIntegral<FDecayU> && !CSameAs<FDecayU, bool>)
{
checkf(Value > 0, TEXT("Illegal format argument. The dynamic width argument must be a unsigned non-zero number."));
return Math::Max(Value, 1);
}
else
{
checkf(false, TEXT("Illegal format argument. The dynamic width argument must be an integral."));
return 0;
}
}
, MinFieldWidth);
}
// Visit the dynamic precision argument.
if (bDynamicMax)
{
MaxDynamicField = Context.Visit([]<typename U>(U&& Value) -> size_t
{
using FDecayU = TRemoveCVRef<U>;
if constexpr (CIntegral<FDecayU> && !CSameAs<FDecayU, bool>)
{
checkf(Value > 0, TEXT("Illegal format argument. The dynamic precision argument must be a unsigned non-zero number."));
return Math::Max(Value, 1);
}
else
{
checkf(false, TEXT("Illegal format argument. The dynamic precision argument must be an integral."));
return 0;
}
}
, MaxFieldWidth);
}
size_t LeftPadding = 0;
size_t RightPadding = 0;
// Estimate the field width.
if (MinDynamicField != 0)
{
// If escape formatting is enabled, add quotes characters.
size_t FieldWidth = bEscape ? 2 : 0;
for (const FCharType* Ptr = Object; *Ptr != LITERAL(FCharType, '\0'); ++Ptr)
{
if (bEscape)
{
switch (const FCharType Char = *Ptr)
{
case LITERAL(FCharType, '\"'):
case LITERAL(FCharType, '\\'):
case LITERAL(FCharType, '\a'):
case LITERAL(FCharType, '\b'):
case LITERAL(FCharType, '\f'):
case LITERAL(FCharType, '\n'):
case LITERAL(FCharType, '\r'):
case LITERAL(FCharType, '\t'):
case LITERAL(FCharType, '\v'): FieldWidth += 2; break;
default:
{
// Use '\x00' format for other non-printable characters.
if (!FCharTraits::IsASCII(Char) || !FCharTraits::IsPrint(Char))
{
FieldWidth += 2 + sizeof(FCharType) * 2;
}
else ++FieldWidth;
}
}
}
else ++FieldWidth;
}
const size_t PaddingWidth = MinDynamicField - Math::Min(FieldWidth, MinDynamicField, MaxDynamicField);
switch (AlignOption)
{
default:
case LITERAL(FCharType, '<'): RightPadding = PaddingWidth; break;
case LITERAL(FCharType, '>'): LeftPadding = PaddingWidth; break;
case LITERAL(FCharType, '^'):
LeftPadding = Math::DivAndFloor(PaddingWidth, 2);
RightPadding = PaddingWidth - LeftPadding;
}
}
// Write the left padding.
for (size_t Index = 0; Index != LeftPadding; ++Index)
{
for (size_t Jndex = 0; Jndex != FillUnitLength; ++Jndex)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = FillCharacter[Jndex];
}
}
// Write the left quote.
if (bEscape)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = LITERAL(FCharType, '\"');
}
const FCharType* Ptr = Object;
bool bComplete = false;
// Write the object, include escaped quotes in the counter.
for (size_t Index = bEscape ? 1 : 0; Index != MaxDynamicField; ++Index)
{
if (*Ptr == LITERAL(FCharType, '\0'))
{
bComplete = true;
break;
}
FCharType Char = *Ptr++;
if (Iter == Sent) UNLIKELY return Iter;
// Convert the character case.
if (bLowercase) Char = FCharTraits::ToLower(Char);
if (bUppercase) Char = FCharTraits::ToUpper(Char);
if (bEscape)
{
switch (Char)
{
case LITERAL(FCharType, '\"'): *Iter++ = LITERAL(FCharType, '\\'); *Iter++ = LITERAL(FCharType, '\"'); break;
case LITERAL(FCharType, '\\'): *Iter++ = LITERAL(FCharType, '\\'); *Iter++ = LITERAL(FCharType, '\\'); break;
case LITERAL(FCharType, '\a'): *Iter++ = LITERAL(FCharType, '\\'); *Iter++ = LITERAL(FCharType, 'a'); break;
case LITERAL(FCharType, '\b'): *Iter++ = LITERAL(FCharType, '\\'); *Iter++ = LITERAL(FCharType, 'b'); break;
case LITERAL(FCharType, '\f'): *Iter++ = LITERAL(FCharType, '\\'); *Iter++ = LITERAL(FCharType, 'f'); break;
case LITERAL(FCharType, '\n'): *Iter++ = LITERAL(FCharType, '\\'); *Iter++ = LITERAL(FCharType, 'n'); break;
case LITERAL(FCharType, '\r'): *Iter++ = LITERAL(FCharType, '\\'); *Iter++ = LITERAL(FCharType, 'r'); break;
case LITERAL(FCharType, '\t'): *Iter++ = LITERAL(FCharType, '\\'); *Iter++ = LITERAL(FCharType, 't'); break;
case LITERAL(FCharType, '\v'): *Iter++ = LITERAL(FCharType, '\\'); *Iter++ = LITERAL(FCharType, 'v'); break;
default:
{
// Use '\x00' format for other non-printable characters.
if (!FCharTraits::IsASCII(Char) || !FCharTraits::IsPrint(Char))
{
*Iter++ = LITERAL(FCharType, '\\');
*Iter++ = LITERAL(FCharType, 'x' );
using FUnsignedT = TMakeUnsigned<FCharType>;
constexpr size_t DigitNum = sizeof(FCharType) * 2;
FUnsignedT IntValue = static_cast<FUnsignedT>(Char);
TStaticArray<FCharType, DigitNum> Buffer;
for (size_t Jndex = 0; Jndex != DigitNum; ++Jndex)
{
Buffer[DigitNum - Jndex - 1] = FCharTraits::FromDigit(IntValue & 0xF);
IntValue >>= 4;
}
check(IntValue == 0);
for (size_t Jndex = 0; Jndex != DigitNum; ++Jndex)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = Buffer[Jndex];
}
}
else *Iter++ = Char;
}
}
}
else *Iter++ = Char;
}
// Write the right quote, if the field width is enough.
if (bEscape && bComplete)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = LITERAL(FCharType, '\"');
}
// Write the right padding.
for (size_t Index = 0; Index != RightPadding; ++Index)
{
for (size_t Jndex = 0; Jndex != FillUnitLength; ++Jndex)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = FillCharacter[Jndex];
}
}
return Iter;
}
private:
size_t FillUnitLength = 1;
FFillCharacter FillCharacter = { LITERAL(FCharType, ' ') };
FCharType AlignOption = LITERAL(FCharType, '<');
size_t MinFieldWidth = 0;
size_t MaxFieldWidth = -1;
bool bDynamicMin = false;
bool bDynamicMax = false;
bool bLowercase = false;
bool bUppercase = false;
bool bEscape = false;
};
template <CCharType T> class TFormatter<const T* , T> : public TFormatter<T*, T> { };
template <CCharType T, size_t N> class TFormatter< T[N], T> : public TFormatter<T*, T> { };
template <CCharType T, size_t N> class TFormatter<const T[N], T> : public TFormatter<T*, T> { };
static_assert(CFormattable< char*>);
static_assert(CFormattable<const char*>);
static_assert(CFormattable< char[256]>);
static_assert(CFormattable<const char[256]>);
/**
* A formatter for integral like types.
*
* The syntax of format specifications is:
*
* [Fill And Align] [Sign] [#] [0] [Width] [Base] [Type] [!] [?]
*
* 1. The fill and align part:
*
* [Fill Character] <Align Option>
*
* i. Fill Character: The character is used to fill width of the object. It is optional and cannot be '{' or '}'.
* It should be representable as a single unicode otherwise it is undefined behavior.
*
* ii. Align Option: The character is used to indicate the direction of alignment.
*
* - '<': Align the formatted argument to the left of the available space
* by inserting n fill characters after the formatted argument.
* - '^': Align the formatted argument to the center of the available space
* by inserting n fill characters around the formatted argument.
* If cannot absolute centering, offset to the left.
* - '>': Align the formatted argument ro the right of the available space
* by inserting n fill characters before the formatted argument.
* This is default option.
*
* 2. The sign part:
*
* This part is allowed only if the type indicator part is not 'C', 'c', 'S' or 's'.
*
* - '+': Always include a sign character before the number. Use '+' for positive.
* - '-': Only include a sign character before the number if the number is negative. This is default option.
* - ' ': Always include a sign character before the number. Use ' ' for positive.
*
* 3. The alternate form indicator part:
*
* This part is allowed only if the type indicator part is not 'C', 'c', 'S' or 's'.
*
* - '#': Insert the prefix '0x' for hexadecimal numbers, '0' for octal numbers, '0b' for binary numbers.
*
* 4. The zero padding part:
*
* This part is allowed only if the type indicator part is not 'C', 'c', 'S' or 's'.
*
* - '0': By adding the prefix '0' to satisfy the minimum field width of the object.
* if the object is normal number.
*
* 5. The width part:
*
* - 'N': The number is used to specify the minimum field width of the object.
* N should be an unsigned non-zero decimal number.
* - '{N}': Dynamically determine the minimum field width of the object.
* N should be a valid index of the format integral argument.
* N is optional, and the default value is automatic indexing.
*
* 5. The base part:
*
* This part is allowed only if the type indicator part is 'I' or 'i'.
*
* - '_N': The number is override the base of the number.
* N should be an unsigned non-zero decimal number.
* - '_{N}': Dynamically override the base of the number.
* N should be a valid index of the format integral argument.
* N is optional, and the default value is automatic indexing.
*
* 7. The type indicator part:
*
* - none: Same as 'D' if the object is integral type.
* Same as 'C' if the object is target character type.
* Same as 'S' if the object is boolean type.
* Others are asserted failure.
*
* - 'I': Indicates the uppercase integer formatting.
* - 'i': Indicates the lowercase integer formatting.
*
* - 'B': Indicates the binary formatting. Same as '_2I'.
* - 'b': Indicates the binary formatting. Same as '_2I'.
* - 'O': Indicates the octal formatting. Same as '_8I'.
* - 'o': Indicates the octal formatting. Same as '_8I'.
* - 'D': Indicates the decimal formatting. Same as '_10I'.
* - 'd': Indicates the decimal formatting. Same as '_10I'.
* - 'X': Indicates the uppercase hexadecimal formatting. Same as '_16I'.
* - 'x': Indicates the lowercase hexadecimal formatting. Same as '_16I'.
*
* If the object is not boolean type and is a valid character value.
*
* - 'C': Indicates the the as-is character formatting.
* - 'c': Indicates the lowercase character formatting.
*
* If the object is boolean type.
*
* - 'C': Indicates the uppercase character formatting, as 'T' or 'F'.
* - 'c': Indicates the lowercase character formatting, as 't' or 'f'.
*
* - 'S': Indicates the normal string formatting, as 'True' or 'False'.
* - 's': Indicates the lowercase string formatting, as 'true' or "false'.
*
* 8. The case indicators part:
*
* - '!': Indicates capitalize the entire string.
*
* 9. The escape indicators part:
*
* This part is allowed only if the type indicator part is 'C', 'c', 'S' or 's'.
*
* - '?': Indicates the escape formatting.
*
*/
template <CIntegral T, CCharType CharType> requires (CSameAs<TRemoveCVRef<T>, T>)
class TFormatter<T, CharType>
{
private:
using FCharType = CharType;
using FCharTraits = TChar<FCharType>;
using FFillCharacter = TStaticArray<FCharType, FCharTraits::MaxCodeUnitLength>;
public:
template <CFormatStringContext<FCharType> CTX>
constexpr TRangeIterator<CTX> Parse(CTX& Context)
{
auto Iter = Ranges::Begin(Context);
auto Sent = Ranges::End (Context);
// Set the default values.
{
FillUnitLength = 1;
FillCharacter[0] = LITERAL(FCharType, ' ');
AlignOption = CSameAs<T, FCharType> || CSameAs<T, bool> ? LITERAL(FCharType, '<') : LITERAL(FCharType, '>');
SignOption = LITERAL(FCharType, '-');
bAlternateForm = false;
bZeroPadding = false;
FieldWidth = 0;
IntegralBase = 10;
bDynamicWidth = false;
bDynamicBase = false;
bCharacter = CSameAs<T, FCharType>;
bString = CSameAs<T, bool>;
bLowercase = false;
bUppercase = false;
bEscape = false;
}
// If the format description string is empty.
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
FCharType Char = *Iter; ++Iter;
// Flag indicates that the part is specified.
bool bHasFillAndAlign = false;
bool bHasSignOption = false;
bool bHasIntegralBase = false;
// Try to parse the fill and align part.
// This code assumes that the format string does not contain multi-unit characters, except for fill character.
// If the fill character is multi-unit.
if (!FCharTraits::IsValid(Char))
{
FillUnitLength = 1;
FillCharacter[0] = Char;
while (true)
{
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
// If the fill character ends.
if (FillUnitLength == FCharTraits::MaxCodeUnitLength || FCharTraits::IsValid(Char)) break;
FillCharacter[FillUnitLength++] = Char;
}
if (Char != LITERAL(FCharType, '<') && Char != LITERAL(FCharType, '^') && Char != LITERAL(FCharType, '>')) UNLIKELY
{
checkf(false, TEXT("Illegal format string. The fill character is not representable as a single unicode."));
return Iter;
}
bHasFillAndAlign = true;
AlignOption = Char;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
// If the fill character is single-unit.
else do
{
if (Iter == Sent) break;
// If the fill character is specified.
if (*Iter == LITERAL(FCharType, '<') || *Iter == LITERAL(FCharType, '^') || *Iter == LITERAL(FCharType, '>'))
{
FillUnitLength = 1;
FillCharacter[0] = Char;
Char = *Iter; ++Iter;
}
// If the fill character is not specified and the align option is not specified.
else if (Char != LITERAL(FCharType, '<') && Char != LITERAL(FCharType, '^') && Char != LITERAL(FCharType, '>')) break;
bHasFillAndAlign = true;
AlignOption = Char;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
while (false);
// The flag indicates that the type indicator part defaults to 'D'.
// Use to check the sign part, alternate form indicator part or zero padding part is allowed
// when there is no type indicator part.
constexpr bool bIntegral = !CSameAs<T, FCharType> && !CSameAs<T, bool>;
// Try to parse the sign part.
switch (Char)
{
case LITERAL(FCharType, '+'):
case LITERAL(FCharType, '-'):
case LITERAL(FCharType, ' '):
bHasSignOption = true;
SignOption = Char;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}'))
{
checkf(bIntegral, TEXT("Illegal format string. The sign option is not allowed for 'C', 'c', 'S' or 's' type."));
return Iter;
}
Char = *Iter; ++Iter;
default: { }
}
// Try to parse the alternate form indicator part.
if (Char == LITERAL(FCharType, '#'))
{
bAlternateForm = true;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}'))
{
checkf(bIntegral, TEXT("Illegal format string. The alternate form is not allowed for 'C', 'c', 'S' or 's' type."));
return Iter;
}
Char = *Iter; ++Iter;
}
// Try to parse the zero padding part.
if (Char == LITERAL(FCharType, '0'))
{
bZeroPadding = true;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}'))
{
checkf(bIntegral, TEXT("Illegal format string. The zero padding is not allowed for 'C', 'c', 'S' or 's' type."));
return Iter;
}
Char = *Iter; ++Iter;
}
// Try to parse the width part.
{
if (Char == LITERAL(FCharType, '{'))
{
bDynamicWidth = true;
FieldWidth = INDEX_NONE;
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
}
if ((bDynamicWidth || Char != LITERAL(FCharType, '0')) && FCharTraits::IsDigit(Char))
{
FieldWidth = FCharTraits::ToDigit(Char);
while (true)
{
if (Iter == Sent)
{
checkf(!bDynamicWidth, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
if (!bDynamicWidth && *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
const uint Digit = FCharTraits::ToDigit(Char);
if (Digit >= 10) break;
FieldWidth = FieldWidth * 10 + Digit;
}
}
if (bDynamicWidth)
{
if (Char != LITERAL(FCharType, '}')) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
do
{
// Try to automatic indexing.
if (FieldWidth == INDEX_NONE)
{
FieldWidth = Context.GetNextIndex();
if (FieldWidth == INDEX_NONE) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the field width."));
}
else break;
}
// Try to manual indexing.
else if (!Context.CheckIndex(FieldWidth)) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the field width."));
}
else break;
bDynamicWidth = false;
FieldWidth = 0;
}
while (false);
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
}
// Try to parse the base part.
if (Char == LITERAL(FCharType, '_'))
{
bHasIntegralBase = true;
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing base in format string."));
return Iter;
}
Char = *Iter; ++Iter;
if (Char == LITERAL(FCharType, '{'))
{
bDynamicBase = true;
IntegralBase = INDEX_NONE;
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
}
if ((bDynamicBase || Char != LITERAL(FCharType, '0')) && FCharTraits::IsDigit(Char))
{
IntegralBase = FCharTraits::ToDigit(Char);
while (true)
{
if (Iter == Sent)
{
checkf(!bDynamicBase, TEXT("Illegal format string. Missing '}' in format string."));
checkf(false, TEXT("Illegal format string. Missing 'I' or 'i' in format string."));
return Iter;
}
if (!bDynamicBase && *Iter == LITERAL(FCharType, '}'))
{
checkf(false, TEXT("Illegal format string. Missing 'I' or 'i' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
const uint Digit = FCharTraits::ToDigit(Char);
if (Digit >= 10) break;
IntegralBase = IntegralBase * 10 + Digit;
}
}
else if (!bDynamicBase)
{
checkf(false, TEXT("Illegal format string. Missing base in format string."));
return Iter;
}
if (bDynamicBase)
{
if (Char != LITERAL(FCharType, '}')) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
do
{
// Try to automatic indexing.
if (IntegralBase == INDEX_NONE)
{
IntegralBase = Context.GetNextIndex();
if (IntegralBase == INDEX_NONE) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the base."));
}
else break;
}
// Try to manual indexing.
else if (!Context.CheckIndex(IntegralBase)) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the base."));
}
else break;
bDynamicBase = false;
IntegralBase = 0;
}
while (false);
if (Iter == Sent || *Iter == LITERAL(FCharType, '}'))
{
checkf(false, TEXT("Illegal format string. Missing 'I' or 'i' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
}
}
// Try to parse the type indicators part.
const bool bHasAlternateForm = bAlternateForm == true;
const bool bHasZeroPadding = bZeroPadding == true;
// If indicates this is lowercase.
switch (Char)
{
case LITERAL(FCharType, 'i'):
case LITERAL(FCharType, 'b'):
case LITERAL(FCharType, 'o'):
case LITERAL(FCharType, 'd'):
case LITERAL(FCharType, 'x'):
case LITERAL(FCharType, 'c'):
case LITERAL(FCharType, 's'): bLowercase = true; break;
default: { }
}
// If indicates this is variable base integer.
switch (Char)
{
case LITERAL(FCharType, 'I'): case LITERAL(FCharType, 'i'):
checkf( bHasIntegralBase, TEXT("Illegal format string. The base is required for 'I' or 'i' type."));
break;
default:
checkf(!bHasIntegralBase, TEXT("Illegal format string. The base is only allowed for 'I' or 'i' type."));
}
// If indicates this is integral.
switch (Char)
{
case LITERAL(FCharType, 'I'): case LITERAL(FCharType, 'i'): break;
case LITERAL(FCharType, 'B'): case LITERAL(FCharType, 'b'): IntegralBase = 2; bDynamicBase = false; break;
case LITERAL(FCharType, 'O'): case LITERAL(FCharType, 'o'): IntegralBase = 8; bDynamicBase = false; break;
case LITERAL(FCharType, 'D'): case LITERAL(FCharType, 'd'): IntegralBase = 10; bDynamicBase = false; break;
case LITERAL(FCharType, 'X'): case LITERAL(FCharType, 'x'): IntegralBase = 16; bDynamicBase = false; break;
default: if constexpr (bIntegral) break;
case LITERAL(FCharType, 'C'): case LITERAL(FCharType, 'c'):
case LITERAL(FCharType, 'S'): case LITERAL(FCharType, 's'):
checkf(!bHasSignOption, TEXT("Illegal format string. The sign option is not allowed for 'C', 'c', 'S' or 's' type."));
checkf(!bHasAlternateForm, TEXT("Illegal format string. The alternate form is not allowed for 'C', 'c', 'S' or 's' type."));
checkf(!bHasZeroPadding, TEXT("Illegal format string. The zero padding is not allowed for 'C', 'c', 'S' or 's' type."));
}
// If indicates this is character or string.
switch (Char)
{
case LITERAL(FCharType, 'I'): case LITERAL(FCharType, 'i'):
case LITERAL(FCharType, 'B'): case LITERAL(FCharType, 'b'):
case LITERAL(FCharType, 'O'): case LITERAL(FCharType, 'o'):
case LITERAL(FCharType, 'D'): case LITERAL(FCharType, 'd'):
case LITERAL(FCharType, 'X'): case LITERAL(FCharType, 'x'): bCharacter = false; bString = false; break;
case LITERAL(FCharType, 'C'): case LITERAL(FCharType, 'c'): bCharacter = true; bString = false; break;
case LITERAL(FCharType, 'S'): case LITERAL(FCharType, 's'): bCharacter = false; bString = true; break;
default: { }
}
if (!bHasFillAndAlign) AlignOption = bCharacter || bString ? LITERAL(FCharType, '<') : LITERAL(FCharType, '>');
checkf((!bString || CSameAs<T, bool>), TEXT("Illegal format string. The 'S' or 's' type is only allowed for boolean type."));
// If exists the type indicators part.
switch (Char)
{
case LITERAL(FCharType, 'I'): case LITERAL(FCharType, 'i'):
case LITERAL(FCharType, 'B'): case LITERAL(FCharType, 'b'):
case LITERAL(FCharType, 'O'): case LITERAL(FCharType, 'o'):
case LITERAL(FCharType, 'D'): case LITERAL(FCharType, 'd'):
case LITERAL(FCharType, 'X'): case LITERAL(FCharType, 'x'):
case LITERAL(FCharType, 'C'): case LITERAL(FCharType, 'c'):
case LITERAL(FCharType, 'S'): case LITERAL(FCharType, 's'): if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter; Char = *Iter; ++Iter; break;
default: { }
}
// Try to parse the case indicators part.
if (Char == LITERAL(FCharType, '!'))
{
bUppercase = true;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
// Try to parse the escape indicators part.
if (Char == LITERAL(FCharType, '?') && (bCharacter || bString))
{
bEscape = true;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
template <CFormatObjectContext<FCharType> CTX>
constexpr TRangeIterator<CTX> Format(T Object, CTX& Context) const
{
auto Iter = Ranges::Begin(Context);
auto Sent = Ranges::End (Context);
size_t TargetField = FieldWidth;
size_t TargetBase = IntegralBase;
// Visit the dynamic width argument.
if (bDynamicWidth)
{
TargetField = Context.Visit([]<typename U>(U&& Value) -> size_t
{
using FDecayU = TRemoveCVRef<U>;
if constexpr (CIntegral<FDecayU> && !CSameAs<FDecayU, bool>)
{
checkf(Value > 0, TEXT("Illegal format argument. The dynamic width argument must be a unsigned non-zero number."));
return Math::Max(Value, 1);
}
else
{
checkf(false, TEXT("Illegal format argument. The dynamic width argument must be an integral."));
return 0;
}
}
, FieldWidth);
}
// Visit the dynamic base argument.
if (bDynamicBase)
{
TargetBase = Context.Visit([]<typename U>(U&& Value) -> size_t
{
using FDecayU = TRemoveCVRef<U>;
if constexpr (CIntegral<FDecayU> && !CSameAs<FDecayU, bool>)
{
checkf(Math::IsWithinInclusive(Value, 2, 36), TEXT("Illegal format argument. The dynamic base argument must be in the range [2, 36]."));
return Math::Max(Value, 1);
}
else
{
checkf(false, TEXT("Illegal format argument. The dynamic base argument must be an integral."));
return 0;
}
}
, IntegralBase);
}
bool bNegative = false;
bool bNormal = false;
size_t TargetWidth;
const FCharType* Target = nullptr;
constexpr size_t BufferSize = sizeof(T) * 8;
TStaticArray<FCharType, BufferSize> Buffer;
do
{
// Handle the literal boolean type.
if constexpr (CSameAs<T, bool>) if (bCharacter || bString)
{
TargetWidth = bCharacter ? 1 : Object ? 4 : 5;
Target = Object ? LITERAL(FCharType, "True") : LITERAL(FCharType, "False");
// Convert the character case.
if (bLowercase) Target = Object ? LITERAL(FCharType, "true") : LITERAL(FCharType, "false");
if (bUppercase) Target = Object ? LITERAL(FCharType, "TRUE") : LITERAL(FCharType, "FALSE");
break;
}
// Handle the literal character type.
if constexpr (!CSameAs<T, bool>) if (bCharacter)
{
TargetWidth = 1;
FCharType Char = static_cast<FCharType>(Object);
checkf(Char == Object, TEXT("Illegal format argument. The integral value is not a valid character."));
// Convert the character case.
if (bLowercase) Char = FCharTraits::ToLower(Char);
if (bUppercase) Char = FCharTraits::ToUpper(Char);
Buffer[0] = Char;
Target = Buffer.GetData();
break;
}
bNormal = true;
// Handle the illegal base.
if (!Math::IsWithinInclusive(TargetBase, 2, 36))
{
checkf(false, TEXT("Illegal format argument. The base must be in the range [2, 36]."));
TargetBase = 10;
}
// Handle the integral boolean type.
if constexpr (CSameAs<T, bool>)
{
TargetWidth = 1;
Buffer[0] = Object ? LITERAL(FCharType, '1') : LITERAL(FCharType, '0');
Target = Buffer.GetData();
break;
}
// Handle the integral type.
else
{
using FUnsignedT = TMakeUnsigned<T>;
FUnsignedT Unsigned = static_cast<FUnsignedT>(Object);
if constexpr (CSigned<T>)
{
if (Object < 0)
{
bNegative = true;
Unsigned = static_cast<FUnsignedT>(-Object);
}
}
FCharType* DigitIter = Buffer.GetData() + BufferSize;
FCharType* DigitSent = Buffer.GetData() + BufferSize;
switch (TargetBase)
{
case 0x02: do { *--DigitIter = static_cast<FCharType>('0' + (Unsigned & 0b00001)); Unsigned >>= 1; } while (Unsigned != 0); break;
case 0x04: do { *--DigitIter = static_cast<FCharType>('0' + (Unsigned & 0b00011)); Unsigned >>= 2; } while (Unsigned != 0); break;
case 0x08: do { *--DigitIter = static_cast<FCharType>('0' + (Unsigned & 0b00111)); Unsigned >>= 3; } while (Unsigned != 0); break;
case 0x10: do { *--DigitIter = FCharTraits::FromDigit(Unsigned & 0b01111, bLowercase && !bUppercase); Unsigned >>= 4; } while (Unsigned != 0); break;
case 0X20: do { *--DigitIter = FCharTraits::FromDigit(Unsigned & 0b11111, bLowercase && !bUppercase); Unsigned >>= 5; } while (Unsigned != 0); break;
case 3:
case 5:
case 6:
case 7:
case 9:
case 10: do { *--DigitIter = static_cast<FCharType>('0' + Unsigned % TargetBase); Unsigned = static_cast<FUnsignedT>(Unsigned / TargetBase); } while (Unsigned != 0); break;
default: do { *--DigitIter = FCharTraits::FromDigit(Unsigned % TargetBase, bLowercase && !bUppercase); Unsigned = static_cast<FUnsignedT>(Unsigned / TargetBase); } while (Unsigned != 0); break;
}
TargetWidth = DigitSent - DigitIter;
Target = DigitIter;
break;
}
}
while (false);
size_t ZeroPadding = 0;
size_t LeftPadding = 0;
size_t RightPadding = 0;
// Estimate the field width.
if (TargetField != 0)
{
size_t LiteralWidth = TargetWidth;
// Handle the escape option.
if (bEscape) LiteralWidth += 2;
// Handle the sign option.
switch (SignOption)
{
case LITERAL(FCharType, '+'):
case LITERAL(FCharType, ' '): LiteralWidth += 1; break;
default: if (bNegative) LiteralWidth += 1;
}
// Handle the alternate form.
if (bAlternateForm) switch (TargetBase)
{
case 0x02: LiteralWidth += 2; break;
case 0x08: LiteralWidth += 1; break;
case 0x10: LiteralWidth += 2; break;
default: { }
}
const size_t PaddingWidth = TargetField - Math::Min(LiteralWidth, TargetField);
if (!bZeroPadding || !bNormal)
{
switch (AlignOption)
{
case LITERAL(FCharType, '<'): RightPadding = PaddingWidth; break;
case LITERAL(FCharType, '>'): LeftPadding = PaddingWidth; break;
case LITERAL(FCharType, '^'):
LeftPadding = Math::DivAndFloor(PaddingWidth, 2);
RightPadding = PaddingWidth - LeftPadding;
break;
default: check_no_entry();
}
}
else ZeroPadding = PaddingWidth;
}
// Write the left padding.
for (size_t Index = 0; Index != LeftPadding; ++Index)
{
for (size_t Jndex = 0; Jndex != FillUnitLength; ++Jndex)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = FillCharacter[Jndex];
}
}
// Write the left quote.
if (bEscape)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = bCharacter ? LITERAL(FCharType, '\'') : LITERAL(FCharType, '\"');
}
// Write the object.
{
if (Iter == Sent) UNLIKELY return Iter;
// Handle the sign option.
switch (SignOption)
{
case LITERAL(FCharType, '+'): *Iter++ = bNegative ? LITERAL(FCharType, '-') : LITERAL(FCharType, '+'); break;
case LITERAL(FCharType, ' '): *Iter++ = bNegative ? LITERAL(FCharType, '-') : LITERAL(FCharType, ' '); break;
default: if (bNegative) *Iter++ = LITERAL(FCharType, '-');
}
// Handle the alternate form.
if (bAlternateForm)
{
if (Iter == Sent) UNLIKELY return Iter;
switch (TargetBase)
{
case 0x02:
case 0x08:
case 0x10: *Iter++ = LITERAL(FCharType, '0'); break;
default: { }
}
if (Iter == Sent) UNLIKELY return Iter;
switch (TargetBase)
{
case 0x02: *Iter++ = bUppercase ? LITERAL(FCharType, 'B') : LITERAL(FCharType, 'b'); break;
case 0x10: *Iter++ = bUppercase ? LITERAL(FCharType, 'X') : LITERAL(FCharType, 'x'); break;
default: { }
}
}
// Handle the zero padding.
for (size_t Index = 0; Index != ZeroPadding; ++Index)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = LITERAL(FCharType, '0');
}
// Write the target object.
for (size_t Index = 0; Index != TargetWidth; ++Index)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = Target[Index];
}
}
// Write the right quote, if the field width is enough.
if (bEscape)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = bCharacter ? LITERAL(FCharType, '\'') : LITERAL(FCharType, '\"');
}
// Write the right padding.
for (size_t Index = 0; Index != RightPadding; ++Index)
{
for (size_t Jndex = 0; Jndex != FillUnitLength; ++Jndex)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = FillCharacter[Jndex];
}
}
return Iter;
}
private:
size_t FillUnitLength = 1;
FFillCharacter FillCharacter = { LITERAL(FCharType, ' ') };
FCharType AlignOption = CSameAs<T, FCharType> || CSameAs<T, bool> ? LITERAL(FCharType, '<') : LITERAL(FCharType, '>');
FCharType SignOption = LITERAL(FCharType, '-');
bool bAlternateForm = false;
bool bZeroPadding = false;
size_t FieldWidth = 0;
size_t IntegralBase = 10;
bool bDynamicWidth = false;
bool bDynamicBase = false;
bool bCharacter = CSameAs<T, FCharType>;
bool bString = CSameAs<T, bool>;
bool bLowercase = false;
bool bUppercase = false;
bool bEscape = false;
};
static_assert(CFormattable<int >);
static_assert(CFormattable<char>);
static_assert(CFormattable<bool>);
/**
* A formatter for the floating-point types.
*
* The syntax of format specifications is:
*
* [Fill And Align] [Sign] [#] [0] [Width] [Precision] [Type] [!]
*
* 1. The fill and align part:
*
* [Fill Character] <Align Option>
*
* i. Fill Character: The character is used to fill width of the object. It is optional and cannot be '{' or '}'.
* It should be representable as a single unicode otherwise it is undefined behavior.
*
* ii. Align Option: The character is used to indicate the direction of alignment.
*
* - '<': Align the formatted argument to the left of the available space
* by inserting n fill characters after the formatted argument.
* - '^': Align the formatted argument to the center of the available space
* by inserting n fill characters around the formatted argument.
* If cannot absolute centering, offset to the left.
* - '>': Align the formatted argument ro the right of the available space
* by inserting n fill characters before the formatted argument.
* This is default option.
*
* 2. The sign part:
*
* - '+': Always include a sign character before the number. Use '+' for positive.
* - '-': Only include a sign character before the number if the number is negative. This is default option.
* - ' ': Always include a sign character before the number. Use ' ' for positive.
*
* 3. The alternate form indicator part:
*
* - '#': Insert the decimal point character unconditionally,
* and do not remove trailing zeros if the type indicator part is 'G' or 'g'.
*
* 4. The zero padding part:
*
* - '0': By adding the prefix '0' to satisfy the minimum field width of the object.
* if the object is normal number.
*
* 5. The width part:
*
* - 'N': The number is used to specify the minimum field width of the object.
* N should be an unsigned non-zero decimal number.
* - '{N}': Dynamically determine the minimum field width of the object.
* N should be a valid index of the format integral argument.
* N is optional, and the default value is automatic indexing.
*
* 6. The precision part:
*
* - '.N': The number is used to specify the precision of the floating-point number.
* N should be an unsigned non-zero decimal number.
* - '.{N}': Dynamically determine the precision of the floating-point number.
* N should be a valid index of the format integral argument.
* N is optional, and the default value is automatic indexing.
*
* 7. The type indicator part:
*
* - none: Indicates the normal formatting.
* - 'G': Indicates the general formatting.
* - 'g': Indicates the general formatting.
* - 'F': Indicates the fixed-point formatting.
* - 'f': Indicates the fixed-point formatting.
* - 'E': Indicates the scientific formatting.
* - 'e': Indicates the scientific formatting.
* - 'A': Indicates the uppercase hexadecimal formatting.
* - 'a': Indicates the lowercase hexadecimal formatting.
*
* 8. The case indicators part:
*
* - '!': Indicates capitalize the entire string.
*
*/
template <CFloatingPoint T, CCharType CharType> requires (CSameAs<TRemoveCVRef<T>, T>)
class TFormatter<T, CharType>
{
private:
using FCharType = CharType;
using FCharTraits = TChar<FCharType>;
using FFillCharacter = TStaticArray<FCharType, FCharTraits::MaxCodeUnitLength>;
public:
template <CFormatStringContext<FCharType> CTX>
constexpr TRangeIterator<CTX> Parse(CTX& Context)
{
auto Iter = Ranges::Begin(Context);
auto Sent = Ranges::End (Context);
// Set the default values.
{
FillUnitLength = 1;
FillCharacter = { LITERAL(FCharType, ' ') };
AlignOption = LITERAL(FCharType, '>');
SignOption = LITERAL(FCharType, '-');
bAlternateForm = false;
bZeroPadding = false;
bHasPrecision = false;
FieldWidth = 0;
Precision = 0;
bDynamicWidth = false;
bDynamicPrecision = false;
bGeneral = false;
bFixedPoint = false;
bScientific = false;
bHexadecimal = false;
bLowercase = false;
bUppercase = false;
}
// If the format description string is empty.
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
FCharType Char = *Iter; ++Iter;
// Flag indicates that the part is specified.
bool bHasFillAndAlign = false;
// Try to parse the fill and align part.
// This code assumes that the format string does not contain multi-unit characters, except for fill character.
// If the fill character is multi-unit.
if (!FCharTraits::IsValid(Char))
{
FillUnitLength = 1;
FillCharacter[0] = Char;
while (true)
{
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
// If the fill character ends.
if (FillUnitLength == FCharTraits::MaxCodeUnitLength || FCharTraits::IsValid(Char)) break;
FillCharacter[FillUnitLength++] = Char;
}
if (Char != LITERAL(FCharType, '<') && Char != LITERAL(FCharType, '^') && Char != LITERAL(FCharType, '>')) UNLIKELY
{
checkf(false, TEXT("Illegal format string. The fill character is not representable as a single unicode."));
return Iter;
}
bHasFillAndAlign = true;
AlignOption = Char;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
// If the fill character is single-unit.
else do
{
if (Iter == Sent) break;
// If the fill character is specified.
if (*Iter == LITERAL(FCharType, '<') || *Iter == LITERAL(FCharType, '^') || *Iter == LITERAL(FCharType, '>'))
{
FillUnitLength = 1;
FillCharacter[0] = Char;
Char = *Iter; ++Iter;
}
// If the fill character is not specified and the align option is not specified.
else if (Char != LITERAL(FCharType, '<') && Char != LITERAL(FCharType, '^') && Char != LITERAL(FCharType, '>')) break;
bHasFillAndAlign = true;
AlignOption = Char;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
while (false);
// Try to parse the sign part.
switch (Char)
{
case LITERAL(FCharType, '+'):
case LITERAL(FCharType, '-'):
case LITERAL(FCharType, ' '):
SignOption = Char;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
default: { }
}
// Try to parse the alternate form indicator part.
if (Char == LITERAL(FCharType, '#'))
{
bAlternateForm = true;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
// Try to parse the zero padding part.
if (Char == LITERAL(FCharType, '0'))
{
bZeroPadding = true;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
// Try to parse the width part.
{
if (Char == LITERAL(FCharType, '{'))
{
bDynamicWidth = true;
FieldWidth = INDEX_NONE;
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
}
if ((bDynamicWidth || Char != LITERAL(FCharType, '0')) && FCharTraits::IsDigit(Char))
{
FieldWidth = FCharTraits::ToDigit(Char);
while (true)
{
if (Iter == Sent)
{
checkf(!bDynamicWidth, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
if (!bDynamicWidth && *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
const uint Digit = FCharTraits::ToDigit(Char);
if (Digit >= 10) break;
FieldWidth = FieldWidth * 10 + Digit;
}
}
if (bDynamicWidth)
{
if (Char != LITERAL(FCharType, '}')) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
do
{
// Try to automatic indexing.
if (FieldWidth == INDEX_NONE)
{
FieldWidth = Context.GetNextIndex();
if (FieldWidth == INDEX_NONE) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the field width."));
}
else break;
}
// Try to manual indexing.
else if (!Context.CheckIndex(FieldWidth)) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the field width."));
}
else break;
bDynamicWidth = false;
FieldWidth = 0;
}
while (false);
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
}
// Try to parse the precision part.
if (Char == LITERAL(FCharType, '.'))
{
bHasPrecision = true;
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing precision in format string."));
return Iter;
}
Char = *Iter; ++Iter;
if (Char == LITERAL(FCharType, '{'))
{
bDynamicPrecision = true;
Precision = INDEX_NONE;
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
}
if ((bDynamicPrecision || Char != LITERAL(FCharType, '0')) && FCharTraits::IsDigit(Char))
{
Precision = FCharTraits::ToDigit(Char);
while (true)
{
if (Iter == Sent)
{
checkf(!bDynamicPrecision, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
if (!bDynamicPrecision && *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
const uint Digit = FCharTraits::ToDigit(Char);
if (Digit >= 10) break;
Precision = Precision * 10 + Digit;
}
}
else if (!bDynamicPrecision)
{
checkf(false, TEXT("Illegal format string. Missing precision in format string."));
return Iter;
}
if (bDynamicPrecision)
{
if (Char != LITERAL(FCharType, '}')) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
do
{
// Try to automatic indexing.
if (Precision == INDEX_NONE)
{
Precision = Context.GetNextIndex();
if (Precision == INDEX_NONE) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the precision."));
}
else break;
}
// Try to manual indexing.
else if (!Context.CheckIndex(Precision)) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the precision."));
}
else break;
bDynamicPrecision = false;
Precision = 0;
}
while (false);
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
}
// Try to parse the type indicators part.
// If indicates this is lowercase.
switch (Char)
{
case LITERAL(FCharType, 'g'):
case LITERAL(FCharType, 'f'):
case LITERAL(FCharType, 'e'):
case LITERAL(FCharType, 'a'): bLowercase = true; break;
default: { }
}
// If indicates this is not normal formatting.
switch (Char)
{
case LITERAL(FCharType, 'G'): case LITERAL(FCharType, 'g'): bGeneral = true; break;
case LITERAL(FCharType, 'F'): case LITERAL(FCharType, 'f'): bFixedPoint = true; break;
case LITERAL(FCharType, 'E'): case LITERAL(FCharType, 'e'): bScientific = true; break;
case LITERAL(FCharType, 'A'): case LITERAL(FCharType, 'a'): bHexadecimal = true; break;
default: { }
}
// If exists the type indicators part.
switch (Char)
{
case LITERAL(FCharType, 'G'): case LITERAL(FCharType, 'g'):
case LITERAL(FCharType, 'F'): case LITERAL(FCharType, 'f'):
case LITERAL(FCharType, 'E'): case LITERAL(FCharType, 'e'):
case LITERAL(FCharType, 'A'): case LITERAL(FCharType, 'a'): if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter; Char = *Iter; ++Iter; break;
default: { }
}
// Try to parse the case indicators part.
if (Char == LITERAL(FCharType, '!'))
{
bUppercase = true;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
template <CFormatObjectContext<FCharType> CTX>
constexpr TRangeIterator<CTX> Format(T Object, CTX& Context) const
{
auto Iter = Ranges::Begin(Context);
auto Sent = Ranges::End (Context);
size_t TargetField = FieldWidth;
size_t TargetPrecision = Precision;
// Visit the dynamic width argument.
if (bDynamicWidth)
{
TargetField = Context.Visit([]<typename U>(U&& Value) -> size_t
{
using FDecayU = TRemoveCVRef<U>;
if constexpr (CIntegral<FDecayU> && !CSameAs<FDecayU, bool>)
{
checkf(Value > 0, TEXT("Illegal format argument. The dynamic width argument must be a unsigned non-zero number."));
return Math::Max(Value, 1);
}
else
{
checkf(false, TEXT("Illegal format argument. The dynamic width argument must be an integral."));
return 0;
}
}
, FieldWidth);
}
// Visit the dynamic precision argument.
if (bDynamicPrecision)
{
TargetPrecision = Context.Visit([]<typename U>(U&& Value) -> size_t
{
using FDecayU = TRemoveCVRef<U>;
if constexpr (CIntegral<FDecayU> && !CSameAs<FDecayU, bool>)
{
checkf(Value >= 0, TEXT("Illegal format argument. The dynamic precision argument must be a unsigned number."));
return Math::Max(Value, 1);
}
else
{
checkf(false, TEXT("Illegal format argument. The dynamic precision argument must be an integral."));
return 0;
}
}
, Precision);
}
const bool bNegative = Math::IsNegative(Object);
bool bNormal = false;
size_t TargetWidth;
const char* Target = nullptr;
constexpr size_t StartingBufferSize = 64;
TArray<char, TInlineAllocator<StartingBufferSize>> Buffer(StartingBufferSize);
// Handle the infinite value.
if (Math::IsInfinity(Object))
{
TargetWidth = 8;
Target = TEXT("Infinity");
// Convert the character case.
if (bLowercase) Target = TEXT("infinity");
if (bUppercase) Target = TEXT("INFINITY");
}
// Handle the NaN value.
else if (Math::IsNaN(Object))
{
TargetWidth = 3;
Target = TEXT("NaN");
// Convert the character case.
if (bLowercase) Target = TEXT("nan");
if (bUppercase) Target = TEXT("NAN");
}
// Handle the normal value.
else
{
bNormal = true;
NAMESPACE_STD::to_chars_result ConvertResult;
while (true)
{
if (bHasPrecision)
{
check(static_cast<int>(TargetPrecision) >= 0);
if (bGeneral ) ConvertResult = NAMESPACE_STD::to_chars(Buffer.GetData(), Buffer.GetData() + Buffer.Num(), Object, NAMESPACE_STD::chars_format::general , static_cast<int>(TargetPrecision));
else if (bFixedPoint ) ConvertResult = NAMESPACE_STD::to_chars(Buffer.GetData(), Buffer.GetData() + Buffer.Num(), Object, NAMESPACE_STD::chars_format::fixed , static_cast<int>(TargetPrecision));
else if (bScientific ) ConvertResult = NAMESPACE_STD::to_chars(Buffer.GetData(), Buffer.GetData() + Buffer.Num(), Object, NAMESPACE_STD::chars_format::scientific, static_cast<int>(TargetPrecision));
else if (bHexadecimal) ConvertResult = NAMESPACE_STD::to_chars(Buffer.GetData(), Buffer.GetData() + Buffer.Num(), Object, NAMESPACE_STD::chars_format::hex , static_cast<int>(TargetPrecision));
else ConvertResult = NAMESPACE_STD::to_chars(Buffer.GetData(), Buffer.GetData() + Buffer.Num(), Object, NAMESPACE_STD::chars_format::general , static_cast<int>(TargetPrecision));
}
else
{
if (bGeneral ) ConvertResult = NAMESPACE_STD::to_chars(Buffer.GetData(), Buffer.GetData() + Buffer.Num(), Object, NAMESPACE_STD::chars_format::general );
else if (bFixedPoint ) ConvertResult = NAMESPACE_STD::to_chars(Buffer.GetData(), Buffer.GetData() + Buffer.Num(), Object, NAMESPACE_STD::chars_format::fixed );
else if (bScientific ) ConvertResult = NAMESPACE_STD::to_chars(Buffer.GetData(), Buffer.GetData() + Buffer.Num(), Object, NAMESPACE_STD::chars_format::scientific);
else if (bHexadecimal) ConvertResult = NAMESPACE_STD::to_chars(Buffer.GetData(), Buffer.GetData() + Buffer.Num(), Object, NAMESPACE_STD::chars_format::hex );
else ConvertResult = NAMESPACE_STD::to_chars(Buffer.GetData(), Buffer.GetData() + Buffer.Num(), Object);
}
if (ConvertResult.ec != NAMESPACE_STD::errc::value_too_large) break;
Buffer.SetNum(Buffer.Num() * 2);
}
Buffer.SetNum(ConvertResult.ptr - Buffer.GetData());
// Remove the negative sign.
if (Buffer.Front() == TEXT('-')) Buffer.StableErase(Buffer.Begin());
// Handle the alternate form.
if (bAlternateForm)
{
const char ExponentChar = bHexadecimal ? TEXT('p') : TEXT('e');
auto BufferIter = Buffer.Begin();
// Insert the decimal point character.
while (true)
{
if (BufferIter == Buffer.End())
{
Buffer.PushBack(TEXT('.'));
BufferIter = Algorithms::Prev(Buffer.End());
break;
}
if (*BufferIter == ExponentChar)
{
BufferIter = Buffer.Insert(BufferIter, TEXT('.'));
break;
}
if (*BufferIter == TEXT('.')) break;
++BufferIter;
}
// Restore trailing zeros.
if (bGeneral)
{
if (!bHasPrecision) TargetPrecision = 6;
size_t DigitNum = BufferIter - Buffer.Begin();
++BufferIter;
while (true)
{
if (DigitNum >= TargetPrecision) break;
if (BufferIter == Buffer.End())
{
Buffer.SetNum(Buffer.Num() + TargetPrecision - DigitNum, TEXT('0'));
break;
}
if (*BufferIter == ExponentChar)
{
Buffer.Insert(BufferIter, TargetPrecision - DigitNum, TEXT('0'));
break;
}
++BufferIter;
++DigitNum;
}
}
}
// Convert the character case.
if (!bLowercase || bUppercase) for (char& Char : Buffer)
{
// Convert the exponent character.
if ( bHexadecimal && Char == TEXT('p')) Char = bUppercase ? TEXT('P') : TEXT('p');
else if (!bHexadecimal && Char == TEXT('e')) Char = bUppercase ? TEXT('E') : TEXT('e');
// Convert the digit character.
else if (!bLowercase) Char = FChar::ToUpper(Char);
}
TargetWidth = Buffer.Num();
Target = Buffer.GetData();
}
size_t ZeroPadding = 0;
size_t LeftPadding = 0;
size_t RightPadding = 0;
// Estimate the field width.
if (TargetField != 0)
{
size_t LiteralWidth = TargetWidth;
// Handle the sign option.
switch (SignOption)
{
case LITERAL(FCharType, '+'):
case LITERAL(FCharType, ' '): LiteralWidth += 1; break;
default: if (bNegative) LiteralWidth += 1;
}
const size_t PaddingWidth = TargetField - Math::Min(LiteralWidth, TargetField);
if (!bZeroPadding || !bNormal)
{
switch (AlignOption)
{
case LITERAL(FCharType, '<'): RightPadding = PaddingWidth; break;
case LITERAL(FCharType, '>'): LeftPadding = PaddingWidth; break;
case LITERAL(FCharType, '^'):
LeftPadding = Math::DivAndFloor(PaddingWidth, 2);
RightPadding = PaddingWidth - LeftPadding;
break;
default: check_no_entry();
}
}
else ZeroPadding = PaddingWidth;
}
// Write the left padding.
for (size_t Index = 0; Index != LeftPadding; ++Index)
{
for (size_t Jndex = 0; Jndex != FillUnitLength; ++Jndex)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = FillCharacter[Jndex];
}
}
// Write the object.
{
static_assert(FChar::IsASCII() && FCharTraits::IsASCII());
if (Iter == Sent) UNLIKELY return Iter;
// Handle the sign option.
switch (SignOption)
{
case LITERAL(FCharType, '+'): *Iter++ = bNegative ? LITERAL(FCharType, '-') : LITERAL(FCharType, '+'); break;
case LITERAL(FCharType, ' '): *Iter++ = bNegative ? LITERAL(FCharType, '-') : LITERAL(FCharType, ' '); break;
default: if (bNegative) *Iter++ = LITERAL(FCharType, '-');
}
// Handle the zero padding.
for (size_t Index = 0; Index != ZeroPadding; ++Index)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = LITERAL(FCharType, '0');
}
// Write the target object.
for (size_t Index = 0; Index != TargetWidth; ++Index)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = static_cast<FCharType>(Target[Index]);
}
}
// Write the right padding.
for (size_t Index = 0; Index != RightPadding; ++Index)
{
for (size_t Jndex = 0; Jndex != FillUnitLength; ++Jndex)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = FillCharacter[Jndex];
}
}
return Iter;
}
private:
size_t FillUnitLength = 1;
FFillCharacter FillCharacter = { LITERAL(FCharType, ' ') };
FCharType AlignOption = LITERAL(FCharType, '>');
FCharType SignOption = LITERAL(FCharType, '-');
bool bAlternateForm = false;
bool bZeroPadding = false;
bool bHasPrecision = false;
size_t FieldWidth = 0;
size_t Precision = 0;
bool bDynamicWidth = false;
bool bDynamicPrecision = false;
bool bGeneral = false;
bool bFixedPoint = false;
bool bScientific = false;
bool bHexadecimal = false;
bool bLowercase = false;
bool bUppercase = false;
};
static_assert(CFormattable<float>);
/**
* A formatter for pointer or member pointer types.
*
* The syntax of format specifications is:
*
* [Fill And Align] [Width] [Type] [!]
*
* 1. The fill and align part:
*
* [Fill Character] <Align Option>
*
* i. Fill Character: The character is used to fill width of the object. It is optional and cannot be '{' or '}'.
* It should be representable as a single unicode otherwise it is undefined behavior.
*
* ii. Align Option: The character is used to indicate the direction of alignment.
*
* - '<': Align the formatted argument to the left of the available space
* by inserting n fill characters after the formatted argument.
* This is default option.
* - '^': Align the formatted argument to the center of the available space
* by inserting n fill characters around the formatted argument.
* If cannot absolute centering, offset to the left.
* - '>': Align the formatted argument ro the right of the available space
* by inserting n fill characters before the formatted argument.
*
* 2. The width part:
*
* - 'N': The number is used to specify the minimum field width of the object.
* N should be an unsigned non-zero decimal number.
* - '{N}': Dynamically determine the minimum field width of the object.
* N should be a valid index of the format integral argument.
* N is optional, and the default value is automatic indexing.
*
* 3. The type indicator part:
*
* - none: Indicates the normal formatting.
* - 'P': Indicates the normal formatting.
* - 'p': Indicates lowercase formatting.
*
* 4. The case indicators part:
*
* - '!': Indicates capitalize the entire string.
*
*/
template <typename T, CCharType CharType> requires ((CNullPointer<T> || CPointer<T> || CMemberPointer<T>) && CSameAs<TRemoveCVRef<T>, T>)
class TFormatter<T, CharType>
{
private:
using FCharType = CharType;
using FCharTraits = TChar<FCharType>;
using FFillCharacter = TStaticArray<FCharType, FCharTraits::MaxCodeUnitLength>;
public:
template <CFormatStringContext<FCharType> CTX>
constexpr TRangeIterator<CTX> Parse(CTX& Context)
{
auto Iter = Ranges::Begin(Context);
auto Sent = Ranges::End (Context);
// Set the default values.
{
FillUnitLength = 1;
FillCharacter[0] = LITERAL(FCharType, ' ');
AlignOption = LITERAL(FCharType, '>');
FieldWidth = 0;
bDynamicWidth = false;
bLowercase = false;
bUppercase = false;
}
// If the format description string is empty.
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
FCharType Char = *Iter; ++Iter;
// Try to parse the fill and align part.
// This code assumes that the format string does not contain multi-unit characters, except for fill character.
// If the fill character is multi-unit.
if (!FCharTraits::IsValid(Char))
{
FillUnitLength = 1;
FillCharacter[0] = Char;
while (true)
{
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
// If the fill character ends.
if (FillUnitLength == FCharTraits::MaxCodeUnitLength || FCharTraits::IsValid(Char)) break;
FillCharacter[FillUnitLength++] = Char;
}
if (Char != LITERAL(FCharType, '<') && Char != LITERAL(FCharType, '^') && Char != LITERAL(FCharType, '>')) UNLIKELY
{
checkf(false, TEXT("Illegal format string. The fill character is not representable as a single unicode."));
return Iter;
}
AlignOption = Char;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
// If the fill character is single-unit.
else do
{
if (Iter == Sent) break;
// If the fill character is specified.
if (*Iter == LITERAL(FCharType, '<') || *Iter == LITERAL(FCharType, '^') || *Iter == LITERAL(FCharType, '>'))
{
FillUnitLength = 1;
FillCharacter[0] = Char;
Char = *Iter; ++Iter;
}
// If the fill character is not specified and the align option is not specified.
else if (Char != LITERAL(FCharType, '<') && Char != LITERAL(FCharType, '^') && Char != LITERAL(FCharType, '>')) break;
AlignOption = Char;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
while (false);
// Try to parse the width part.
{
if (Char == LITERAL(FCharType, '{'))
{
bDynamicWidth = true;
FieldWidth = INDEX_NONE;
if (Iter == Sent) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
Char = *Iter; ++Iter;
}
if ((bDynamicWidth || Char != LITERAL(FCharType, '0')) && FCharTraits::IsDigit(Char))
{
FieldWidth = FCharTraits::ToDigit(Char);
while (true)
{
if (Iter == Sent)
{
checkf(!bDynamicWidth, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
if (!bDynamicWidth && *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
const uint Digit = FCharTraits::ToDigit(Char);
if (Digit >= 10) break;
FieldWidth = FieldWidth * 10 + Digit;
}
}
if (bDynamicWidth)
{
if (Char != LITERAL(FCharType, '}')) UNLIKELY
{
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
do
{
// Try to automatic indexing.
if (FieldWidth == INDEX_NONE)
{
FieldWidth = Context.GetNextIndex();
if (FieldWidth == INDEX_NONE) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the field width."));
}
else break;
}
// Try to manual indexing.
else if (!Context.CheckIndex(FieldWidth)) UNLIKELY
{
checkf(false, TEXT("Illegal index. Please check the field width."));
}
else break;
bDynamicWidth = false;
FieldWidth = 0;
}
while (false);
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
}
// Try to parse the type indicators part.
switch (Char)
{
case LITERAL(FCharType, 'p'): bLowercase = true; break;
default: { }
}
switch (Char)
{
case LITERAL(FCharType, 'P'):
case LITERAL(FCharType, 'p'): if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter; Char = *Iter; ++Iter; break;
default: { }
}
// Try to parse the case indicators part.
if (Char == LITERAL(FCharType, '!'))
{
bUppercase = true;
if (Iter == Sent || *Iter == LITERAL(FCharType, '}')) return Iter;
Char = *Iter; ++Iter;
}
checkf(false, TEXT("Illegal format string. Missing '}' in format string."));
return Iter;
}
template <CFormatObjectContext<FCharType> CTX>
constexpr TRangeIterator<CTX> Format(T Object, CTX& Context) const
{
auto Iter = Ranges::Begin(Context);
auto Sent = Ranges::End (Context);
size_t TargetField = FieldWidth;
// Visit the dynamic width argument.
if (bDynamicWidth)
{
TargetField = Context.Visit([]<typename U>(U&& Value) -> size_t
{
using FDecayU = TRemoveCVRef<U>;
if constexpr (CIntegral<FDecayU> && !CSameAs<FDecayU, bool>)
{
checkf(Value > 0, TEXT("Illegal format argument. The dynamic width argument must be a unsigned non-zero number."));
return Math::Max(Value, 1);
}
else
{
checkf(false, TEXT("Illegal format argument. The dynamic width argument must be an integral."));
return 0;
}
}
, FieldWidth);
}
size_t LeftPadding = 0;
size_t RightPadding = 0;
// Estimate the field width.
if (TargetField != 0)
{
const size_t LiteralWidth = 2 * sizeof(T) + 2;
const size_t PaddingWidth = TargetField - Math::Min(LiteralWidth, TargetField);
switch (AlignOption)
{
case LITERAL(FCharType, '<'): RightPadding = PaddingWidth; break;
case LITERAL(FCharType, '>'): LeftPadding = PaddingWidth; break;
case LITERAL(FCharType, '^'):
LeftPadding = Math::DivAndFloor(PaddingWidth, 2);
RightPadding = PaddingWidth - LeftPadding;
break;
default: check_no_entry();
}
}
// Write the left padding.
for (size_t Index = 0; Index != LeftPadding; ++Index)
{
for (size_t Jndex = 0; Jndex != FillUnitLength; ++Jndex)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = FillCharacter[Jndex];
}
}
// Write the object, include escaped quotes in the counter.
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = LITERAL(FCharType, '0');
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = bUppercase ? LITERAL(FCharType, 'X') : LITERAL(FCharType, 'x');
const uint8* Ptr = reinterpret_cast<const uint8*>(&Object);
if constexpr (Math::EEndian::Native != Math::EEndian::Little)
{
for (size_t Index = 0; Index != sizeof(T); ++Index)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = FCharTraits::FromDigit(Ptr[Index] >> 4, bLowercase);
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = FCharTraits::FromDigit(Ptr[Index] & 0x0F, bLowercase);
}
}
else
{
for (size_t Index = 0; Index != sizeof(T); ++Index)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = FCharTraits::FromDigit(Ptr[sizeof(T) - Index - 1] >> 4, bLowercase);
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = FCharTraits::FromDigit(Ptr[sizeof(T) - Index - 1] & 0x0F, bLowercase);
}
}
}
// Write the right padding.
for (size_t Index = 0; Index != RightPadding; ++Index)
{
for (size_t Jndex = 0; Jndex != FillUnitLength; ++Jndex)
{
if (Iter == Sent) UNLIKELY return Iter;
*Iter++ = FillCharacter[Jndex];
}
}
return Iter;
}
private:
size_t FillUnitLength = 1;
FFillCharacter FillCharacter = { LITERAL(FCharType, ' ') };
FCharType AlignOption = LITERAL(FCharType, '>');
size_t FieldWidth = 0;
bool bDynamicWidth = false;
bool bLowercase = false;
bool bUppercase = false;
};
NAMESPACE_MODULE_END(Utility)
NAMESPACE_MODULE_END(Redcraft)
NAMESPACE_REDCRAFT_END
#pragma warning(pop)
Reference in New Issue View Git Blame Copy Permalink