--- libecb/ecb.pod 2015/02/19 15:45:29 1.65 +++ libecb/ecb.pod 2021/11/22 17:15:50 1.101 @@ -12,14 +12,14 @@ http://software.schmorp.de/pkg/libecb -It mainly provides a number of wrappers around GCC built-ins, together -with replacement functions for other compilers. In addition to this, -it provides a number of other lowlevel C utilities, such as endianness -detection, byte swapping or bit rotations. - -Or in other words, things that should be built into any standard C system, -but aren't, implemented as efficient as possible with GCC, and still -correct with other compilers. +It mainly provides a number of wrappers around many compiler built-ins, +together with replacement functions for other compilers. In addition +to this, it provides a number of other lowlevel C utilities, such as +endianness detection, byte swapping or bit rotations. + +Or in other words, things that should be built into any standard C +system, but aren't, implemented as efficient as possible with GCC (clang, +msvc...), and still correct with other compilers. More might come. @@ -60,15 +60,21 @@ ecb.h makes sure that the following types are defined (in the expected way): - int8_t uint8_t int16_t uint16_t - int32_t uint32_t int64_t uint64_t - intptr_t uintptr_t + int8_t uint8_ + int16_t uint16_t + int32_t uint32_ + int64_t uint64_t + int_fast8_t uint_fast8_t + int_fast16_t uint_fast16_t + int_fast32_t uint_fast32_t + int_fast64_t uint_fast64_t + intptr_t uintptr_t The macro C is defined to the size of a pointer on this platform (currently C<4> or C<8>) and can be used in preprocessor expressions. -For C and C use C. +For C and C use C/C. =head2 LANGUAGE/ENVIRONMENT/COMPILER VERSIONS @@ -76,12 +82,12 @@ preprocessor instructions as well as treated as a boolean (use C to ensure it's either C<0> or C<1> if you need that). -=over 4 +=over =item ECB_C True if the implementation defines the C<__STDC__> macro to a true value, -while not claiming to be C++. +while not claiming to be C++, i..e C, but not C++. =item ECB_C99 @@ -91,26 +97,34 @@ Note that later versions (ECB_C11) remove core features again (for example, variable length arrays). -=item ECB_C11 +=item ECB_C11, ECB_C17 -True if the implementation claims to be compliant to C11 (ISO/IEC -9899:2011) or any later version, while not claiming to be C++. +True if the implementation claims to be compliant to C11/C17 (ISO/IEC +9899:2011, :20187) or any later version, while not claiming to be C++. =item ECB_CPP True if the implementation defines the C<__cplusplus__> macro to a true value, which is typically true for C++ compilers. -=item ECB_CPP11 +=item ECB_CPP11, ECB_CPP14, ECB_CPP17 -True if the implementation claims to be compliant to ISO/IEC 14882:2011 -(C++11) or any later version. +True if the implementation claims to be compliant to C++11/C++14/C++17 +(ISO/IEC 14882:2011, :2014, :2017) or any later version. + +Note that many C++20 features will likely have their own feature test +macros (see e.g. L). + +=item ECB_OPTIMIZE_SIZE + +Is C<1> when the compiler optimizes for size, C<0> otherwise. This symbol +can also be defined before including F, in which case it will be +unchanged. =item ECB_GCC_VERSION (major, minor) -Expands to a true value (suitable for testing in by the preprocessor) -if the compiler used is GNU C and the version is the given version, or -higher. +Expands to a true value (suitable for testing by the preprocessor) if the +compiler used is GNU C and the version is the given version, or higher. This macro tries to return false on compilers that claim to be GCC compatible but aren't. @@ -139,7 +153,7 @@ =item ECB_STDFP -If this evaluates to a true value (suitable for testing in by the +If this evaluates to a true value (suitable for testing by the preprocessor), then C and C use IEEE 754 single/binary32 and double/binary64 representations internally I the endianness of both types match the endianness of C and C. @@ -152,6 +166,14 @@ not be able to deduce this correctly everywhere and might err on the safe side. +=item ECB_64BIT_NATIVE + +Evaluates to a true value (suitable for both preprocessor and C code +testing) if 64 bit integer types on this architecture are evaluated +"natively", that is, with similar speeds as 32 bit integers. While 64 bit +integer support is very common (and in fact required by libecb), 32 bit +cpus have to emulate operations on them, so you might want to avoid them. + =item ECB_AMD64, ECB_AMD64_X32 These two macros are defined to C<1> on the x86_64/amd64 ABI and the X32 @@ -167,7 +189,7 @@ =head2 MACRO TRICKERY -=over 4 +=over =item ECB_CONCAT (a, b) @@ -218,13 +240,13 @@ ecb_const int mysqrt (int a); ecb_unused int i; -=over 4 +=over =item ecb_unused Marks a function or a variable as "unused", which simply suppresses a -warning by GCC when it detects it as unused. This is useful when you e.g. -declare a variable but do not always use it: +warning by the compiler when it detects it as unused. This is useful when +you e.g. declare a variable but do not always use it: { ecb_unused int var; @@ -244,15 +266,15 @@ =item ecb_deprecated_message (message) -Same as C, but if possible, supply a diagnostic that is +Same as C, but if possible, the specified diagnostic is used instead of a generic depreciation message when the object is being used. =item ecb_inline -Expands either to C or to just C, if inline -isn't supported. It should be used to declare functions that should be -inlined, for code size or speed reasons. +Expands either to (a compiler-specific equivalent of) C or +to just C, if inline isn't supported. It should be used to declare +functions that should be inlined, for code size or speed reasons. Example: inline this function, it surely will reduce codesize. @@ -264,7 +286,7 @@ =item ecb_noinline -Prevent a function from being inlined - it might be optimised away, but +Prevents a function from being inlined - it might be optimised away, but not inlined into other functions. This is useful if you know your function is rarely called and large enough for inlining not to be helpful. @@ -402,7 +424,7 @@ =head2 OPTIMISATION HINTS -=over 4 +=over =item bool ecb_is_constant (expr) @@ -491,7 +513,7 @@ =item ecb_assume (cond) -Try to tell the compiler that some condition is true, even if it's not +Tries to tell the compiler that some condition is true, even if it's not obvious. This is not a function, but a statement: it cannot be used in another expression. @@ -569,7 +591,7 @@ =head2 BIT FIDDLING / BIT WIZARDRY -=over 4 +=over =item bool ecb_big_endian () @@ -585,12 +607,17 @@ =item int ecb_ctz64 (uint64_t x) +=item int ecb_ctz (T x) [C++] + Returns the index of the least significant bit set in C (or equivalently the number of bits set to 0 before the least significant bit set), starting from 0. If C is 0 the result is undefined. For smaller types than C you can safely use C. +The overloaded C++ C function supports C, C, +C and C types. + For example: ecb_ctz32 (3) = 0 @@ -600,14 +627,21 @@ =item bool ecb_is_pot64 (uint32_t x) -Return true iff C is a power of two or C. +=item bool ecb_is_pot (T x) [C++] -For smaller types then C you can safely use C. +Returns true iff C is a power of two or C. + +For smaller types than C you can safely use C. + +The overloaded C++ C function supports C, C, +C and C types. =item int ecb_ld32 (uint32_t x) =item int ecb_ld64 (uint64_t x) +=item int ecb_ld64 (T x) [C++] + Returns the index of the most significant bit set in C, or the number of digits the number requires in binary (so that C<< 2**ld <= x < 2**(ld+1) >>). If C is 0 the result is undefined. A common use case is @@ -621,14 +655,22 @@ For smaller types than C you can safely use C. +The overloaded C++ C function supports C, C, +C and C types. + =item int ecb_popcount32 (uint32_t x) =item int ecb_popcount64 (uint64_t x) +=item int ecb_popcount (T x) [C++] + Returns the number of bits set to 1 in C. For smaller types than C you can safely use C. +The overloaded C++ C function supports C, C, +C and C types. + For example: ecb_popcount32 (7) = 3 @@ -640,24 +682,39 @@ =item uint32_t ecb_bitrev32 (uint32_t x) +=item T ecb_bitrev (T x) [C++] + Reverses the bits in x, i.e. the MSB becomes the LSB, MSB-1 becomes LSB+1 and so on. +The overloaded C++ C function supports C, C and C types. + Example: ecb_bitrev8 (0xa7) = 0xea ecb_bitrev32 (0xffcc4411) = 0x882233ff +=item T ecb_bitrev (T x) [C++] + +Overloaded C++ bitrev function. + +C must be one of C, C or C. + =item uint32_t ecb_bswap16 (uint32_t x) =item uint32_t ecb_bswap32 (uint32_t x) =item uint64_t ecb_bswap64 (uint64_t x) +=item T ecb_bswap (T x) + These functions return the value of the 16-bit (32-bit, 64-bit) value C after reversing the order of bytes (0x11223344 becomes 0x44332211 in C). +The overloaded C++ C function supports C, C, +C and C types. + =item uint8_t ecb_rotl8 (uint8_t x, unsigned int count) =item uint16_t ecb_rotl16 (uint16_t x, unsigned int count) @@ -676,38 +733,442 @@ These two families of functions return the value of C after rotating all the bits by C positions to the right (C) or left -(C). +(C). There are no restrictions on the value C, i.e. both +zero and values equal or larger than the word width work correctly. Also, +notwithstanding C being unsigned, negative numbers work and shift +to the opposite direction. + +Current GCC/clang versions understand these functions and usually compile +them to "optimal" code (e.g. a single C or a combination of C +on x86). + +=item T ecb_rotl (T x, unsigned int count) [C++] + +=item T ecb_rotr (T x, unsigned int count) [C++] + +Overloaded C++ rotl/rotr functions. + +C must be one of C, C, C or C. + +=back + +=head2 BIT MIXING, HASHING + +Sometimes you have an integer and want to distribute its bits well, for +example, to use it as a hash in a hashtable. A common example is pointer +values, which often only have a limited range (e.g. low and high bits are +often zero). + +The following functions try to mix the bits to get a good bias-free +distribution. They were mainly made for pointers, but the underlying +integer functions are exposed as well. + +As an added benefit, the functions are reversible, so if you find it +convenient to store only the hash value, you can recover the original +pointer from the hash ("unmix"), as long as your pinters are 32 or 64 bit +(if this isn't the case on your platform, drop us a note and we will add +functions for other bit widths). + +The unmix functions are very slightly slower than the mix functions, so +it is equally very slightly preferable to store the original values wehen +convenient. + +The underlying algorithm if subject to change, so currently these +functions are not suitable for persistent hash tables, as their result +value can change between diferent versions of libecb. + +=over + +=item uintptr_t ecb_ptrmix (void *ptr) + +Mixes the bits of a pointer so the result is suitable for hash table +lookups. In other words, this hashes the pointer value. + +=item uintptr_t ecb_ptrmix (T *ptr) [C++] + +Overload the C function to work for any pointer in C++. + +=item void *ecb_ptrunmix (uintptr_t v) + +Unmix the hash value into the original pointer. This only works as long +as the hash value is not truncated, i.e. you used C (or +equivalent) throughout to store it. + +=item T *ecb_ptrunmix (uintptr_t v) [C++] + +The somewhat less useful template version of C for +C++. Example: + + sometype *myptr; + uintptr_t hash = ecb_ptrmix (myptr); + sometype *orig = ecb_ptrunmix (hash); + +=item uint32_t ecb_mix32 (uint32_t v) + +=item uint64_t ecb_mix64 (uint64_t v) + +Sometimes you don't have a pointer but an integer whose values are very +badly distributed. In this case you cna sue these integer versions of the +mixing function. No C++ template is provided currently. + +=item uint32_t ecb_unmix32 (uint32_t v) + +=item uint64_t ecb_unmix64 (uint64_t v) + +The reverse of the C functions - they take a mixed/hashed value +and recover the original value. -Current GCC versions understand these functions and usually compile them -to "optimal" code (e.g. a single C or a combination of C on -x86). +=back + +=head2 HOST ENDIANNESS CONVERSION + +=over + +=item uint_fast16_t ecb_be_u16_to_host (uint_fast16_t v) + +=item uint_fast32_t ecb_be_u32_to_host (uint_fast32_t v) + +=item uint_fast64_t ecb_be_u64_to_host (uint_fast64_t v) + +=item uint_fast16_t ecb_le_u16_to_host (uint_fast16_t v) + +=item uint_fast32_t ecb_le_u32_to_host (uint_fast32_t v) + +=item uint_fast64_t ecb_le_u64_to_host (uint_fast64_t v) + +Convert an unsigned 16, 32 or 64 bit value from big or little endian to host byte order. + +The naming convention is C(C|C)C<_u>C<16|32|64>C<_to_host>, +where C and C stand for big endian and little endian, respectively. + +=item uint_fast16_t ecb_host_to_be_u16 (uint_fast16_t v) + +=item uint_fast32_t ecb_host_to_be_u32 (uint_fast32_t v) + +=item uint_fast64_t ecb_host_to_be_u64 (uint_fast64_t v) + +=item uint_fast16_t ecb_host_to_le_u16 (uint_fast16_t v) + +=item uint_fast32_t ecb_host_to_le_u32 (uint_fast32_t v) + +=item uint_fast64_t ecb_host_to_le_u64 (uint_fast64_t v) + +Like above, but converts I host byte order to the specified +endianness. + +=back + +In C++ the following additional template functions are supported: + +=over + +=item T ecb_be_to_host (T v) + +=item T ecb_le_to_host (T v) + +=item T ecb_host_to_be (T v) + +=item T ecb_host_to_le (T v) + +=back + +These functions work like their C counterparts, above, but use templates, +which make them useful in generic code. + +C must be one of C, C, C or C +(so unlike their C counterparts, there is a version for C, which +again can be useful in generic code). + +=head2 UNALIGNED LOAD/STORE + +These function load or store unaligned multi-byte values. + +=over + +=item uint_fast16_t ecb_peek_u16_u (const void *ptr) + +=item uint_fast32_t ecb_peek_u32_u (const void *ptr) + +=item uint_fast64_t ecb_peek_u64_u (const void *ptr) + +These functions load an unaligned, unsigned 16, 32 or 64 bit value from +memory. + +=item uint_fast16_t ecb_peek_be_u16_u (const void *ptr) + +=item uint_fast32_t ecb_peek_be_u32_u (const void *ptr) + +=item uint_fast64_t ecb_peek_be_u64_u (const void *ptr) + +=item uint_fast16_t ecb_peek_le_u16_u (const void *ptr) + +=item uint_fast32_t ecb_peek_le_u32_u (const void *ptr) + +=item uint_fast64_t ecb_peek_le_u64_u (const void *ptr) + +Like above, but additionally convert from big endian (C) or little +endian (C) byte order to host byte order while doing so. + +=item ecb_poke_u16_u (void *ptr, uint16_t v) + +=item ecb_poke_u32_u (void *ptr, uint32_t v) + +=item ecb_poke_u64_u (void *ptr, uint64_t v) + +These functions store an unaligned, unsigned 16, 32 or 64 bit value to +memory. + +=item ecb_poke_be_u16_u (void *ptr, uint_fast16_t v) + +=item ecb_poke_be_u32_u (void *ptr, uint_fast32_t v) + +=item ecb_poke_be_u64_u (void *ptr, uint_fast64_t v) + +=item ecb_poke_le_u16_u (void *ptr, uint_fast16_t v) + +=item ecb_poke_le_u32_u (void *ptr, uint_fast32_t v) + +=item ecb_poke_le_u64_u (void *ptr, uint_fast64_t v) + +Like above, but additionally convert from host byte order to big endian +(C) or little endian (C) byte order while doing so. + +=back + +In C++ the following additional template functions are supported: + +=over + +=item T ecb_peek (const void *ptr) + +=item T ecb_peek_be (const void *ptr) + +=item T ecb_peek_le (const void *ptr) + +=item T ecb_peek_u (const void *ptr) + +=item T ecb_peek_be_u (const void *ptr) + +=item T ecb_peek_le_u (const void *ptr) + +Similarly to their C counterparts, these functions load an unsigned 8, 16, +32 or 64 bit value from memory, with optional conversion from big/little +endian. + +Since the type cannot be deduced, it has to be specified explicitly, e.g. + + uint_fast16_t v = ecb_peek (ptr); + +C must be one of C, C, C or C. + +Unlike their C counterparts, these functions support 8 bit quantities +(C) and also have an aligned version (without the C<_u> prefix), +all of which hopefully makes them more useful in generic code. + +=item ecb_poke (void *ptr, T v) + +=item ecb_poke_be (void *ptr, T v) + +=item ecb_poke_le (void *ptr, T v) + +=item ecb_poke_u (void *ptr, T v) + +=item ecb_poke_be_u (void *ptr, T v) + +=item ecb_poke_le_u (void *ptr, T v) + +Again, similarly to their C counterparts, these functions store an +unsigned 8, 16, 32 or z64 bit value to memory, with optional conversion to +big/little endian. + +C must be one of C, C, C or C. + +Unlike their C counterparts, these functions support 8 bit quantities +(C) and also have an aligned version (without the C<_u> prefix), +all of which hopefully makes them more useful in generic code. + +=back + +=head2 FAST INTEGER TO STRING + +Libecb defines a set of very fast integer to decimal string (or integer +to ascii, short C) functions. These work by converting the integer +to a fixed point representation and then successively multiplying out +the topmost digits. Unlike some other, also very fast, libraries, ecb's +algorithm should be completely branchless per digit, and does not rely on +the presence of special cpu functions (such as clz). + +There is a high level API that takes an C, C, +C or C as argument, and a low-level API, which is +harder to use but supports slightly more formatting options. + +=head3 HIGH LEVEL API + +The high level API consists of four functions, one each for C, +C, C and C: + +Example: + + char buf[ECB_I2A_MAX_DIGITS + 1]; + char *end = ecb_i2a_i32 (buf, 17262); + *end = 0; + // buf now contains "17262" + +=over + +=item ECB_I2A_I32_DIGITS (=11) + +=item char *ecb_i2a_u32 (char *ptr, uint32_t value) + +Takes an C I and formats it as a decimal number starting +at I, using at most C characters. Returns a +pointer to just after the generated string, where you would normally put +the terminating C<0> character. This function outputs the minimum number +of digits. + +=item ECB_I2A_U32_DIGITS (=10) + +=item char *ecb_i2a_i32 (char *ptr, int32_t value) + +Same as C, but formats a C value, including a minus +sign if needed. + +=item ECB_I2A_I64_DIGITS (=20) + +=item char *ecb_i2a_u64 (char *ptr, uint64_t value) + +=item ECB_I2A_U64_DIGITS (=21) + +=item char *ecb_i2a_i64 (char *ptr, int64_t value) + +Similar to their 32 bit counterparts, these take a 64 bit argument. + +=item ECB_I2A_MAX_DIGITS (=21) + +Instead of using a type specific length macro, you can just use +C, which is good enough for any C function. + +=back + +=head3 LOW-LEVEL API + +The functions above use a number of low-level APIs which have some strict +limitations, but can be used as building blocks (studying C +and related functions is recommended). + +There are three families of functions: functions that convert a number +to a fixed number of digits with leading zeroes (C, C<0> +for "leading zeroes"), functions that generate up to N digits, skipping +leading zeroes (C<_N>), and functions that can generate more digits, but +the leading digit has limited range (C<_xN>). + +None of the functions deal with negative numbers. + +Example: convert an IP address in an u32 into dotted-quad: + + uint32_t ip = 0x0a000164; // 10.0.1.100 + char ips[3 * 4 + 3 + 1]; + char *ptr = ips; + ptr = ecb_i2a_3 (ptr, ip >> 24 ); *ptr++ = '.'; + ptr = ecb_i2a_3 (ptr, (ip >> 16) & 0xff); *ptr++ = '.'; + ptr = ecb_i2a_3 (ptr, (ip >> 8) & 0xff); *ptr++ = '.'; + ptr = ecb_i2a_3 (ptr, ip & 0xff); *ptr++ = 0; + printf ("ip: %s\n", ips); // prints "ip: 10.0.1.100" + +=over + +=item char *ecb_i2a_02 (char *ptr, uint32_t value) // 32 bit + +=item char *ecb_i2a_03 (char *ptr, uint32_t value) // 32 bit + +=item char *ecb_i2a_04 (char *ptr, uint32_t value) // 32 bit + +=item char *ecb_i2a_05 (char *ptr, uint32_t value) // 64 bit + +=item char *ecb_i2a_06 (char *ptr, uint32_t value) // 64 bit + +=item char *ecb_i2a_07 (char *ptr, uint32_t value) // 64 bit + +=item char *ecb_i2a_08 (char *ptr, uint32_t value) // 64 bit + +=item char *ecb_i2a_09 (char *ptr, uint32_t value) // 64 bit + +The C<< ecb_i2a_0I >> functions take an unsigned I and convert +them to exactly I digits, returning a pointer to the first character +after the digits. The I must be in range. The functions marked with +I<32 bit> do their calculations internally in 32 bit, the ones marked with +I<64 bit> internally use 64 bit integers, which might be slow on 32 bit +architectures (the high level API decides on 32 vs. 64 bit versions using +C). + +=item char *ecb_i2a_2 (char *ptr, uint32_t value) // 32 bit + +=item char *ecb_i2a_3 (char *ptr, uint32_t value) // 32 bit + +=item char *ecb_i2a_4 (char *ptr, uint32_t value) // 32 bit + +=item char *ecb_i2a_5 (char *ptr, uint32_t value) // 64 bit + +=item char *ecb_i2a_6 (char *ptr, uint32_t value) // 64 bit + +=item char *ecb_i2a_7 (char *ptr, uint32_t value) // 64 bit + +=item char *ecb_i2a_8 (char *ptr, uint32_t value) // 64 bit + +=item char *ecb_i2a_9 (char *ptr, uint32_t value) // 64 bit + +Similarly, the C<< ecb_i2a_I >> functions take an unsigned I +and convert them to at most I digits, suppressing leading zeroes, and +returning a pointer to the first character after the digits. + +=item ECB_I2A_MAX_X5 (=59074) + +=item char *ecb_i2a_x5 (char *ptr, uint32_t value) // 32 bit + +=item ECB_I2A_MAX_X10 (=2932500665) + +=item char *ecb_i2a_x10 (char *ptr, uint32_t value) // 64 bit + +The C<< ecb_i2a_xI >> functions are similar to the C<< ecb_i2a_I >> +functions, but they can generate one digit more, as long as the number +is within range, which is given by the symbols C (almost +16 bit range) and C (a bit more than 31 bit range), +respectively. + +For example, the digit part of a 32 bit signed integer just fits into the +C range, so while C cannot convert a 10 +digit number, it can convert all 32 bit signed numbers. Sadly, it's not +good enough for 32 bit unsigned numbers. =back =head2 FLOATING POINT FIDDLING -=over 4 +=over -=item ECB_INFINITY +=item ECB_INFINITY [-UECB_NO_LIBM] Evaluates to positive infinity if supported by the platform, otherwise to a truly huge number. -=item ECB_NAN +=item ECB_NAN [-UECB_NO_LIBM] Evaluates to a quiet NAN if supported by the platform, otherwise to C. -=item float ecb_ldexpf (float x, int exp) +=item float ecb_ldexpf (float x, int exp) [-UECB_NO_LIBM] Same as C, but always available. +=item uint32_t ecb_float_to_binary16 (float x) [-UECB_NO_LIBM] + =item uint32_t ecb_float_to_binary32 (float x) [-UECB_NO_LIBM] =item uint64_t ecb_double_to_binary64 (double x) [-UECB_NO_LIBM] These functions each take an argument in the native C or C -type and return the IEEE 754 bit representation of it. +type and return the IEEE 754 bit representation of it (binary16/half, +binary32/single or binary64/double precision). The bit representation is just as IEEE 754 defines it, i.e. the sign bit will be the most significant bit, followed by exponent and mantissa. @@ -718,10 +1179,10 @@ denormals, but will likely convert negative zero to positive zero). On all modern platforms (where C is true), the compiler should -be able to optimise away this function completely. +be able to completely optimise away the 32 and 64 bit functions. These functions can be helpful when serialising floats to the network - you -can serialise the return value like a normal uint32_t/uint64_t. +can serialise the return value like a normal uint16_t/uint32_t/uint64_t. Another use for these functions is to manipulate floating point values directly. @@ -738,11 +1199,12 @@ =item float ecb_binary32_to_float (uint32_t x) [-UECB_NO_LIBM] -=item double ecb_binary32_to_double (uint64_t x) [-UECB_NO_LIBM] +=item double ecb_binary64_to_double (uint64_t x) [-UECB_NO_LIBM] The reverse operation of the previous function - takes the bit -representation of an IEEE binary16, binary32 or binary64 number and -converts it to the native C or C format. +representation of an IEEE binary16, binary32 or binary64 number (half, +single or double precision) and converts it to the native C or +C format. This function should work even when the native floating point format isn't IEEE compliant, of course at a speed and code size penalty, and of course @@ -753,11 +1215,24 @@ On all modern platforms (where C is true), the compiler should be able to optimise away this function completely. +=item uint16_t ecb_binary32_to_binary16 (uint32_t x) + +=item uint32_t ecb_binary16_to_binary32 (uint16_t x) + +Convert a IEEE binary32/single precision to binary16/half format, and vice +versa, handling all details (round-to-nearest-even, subnormals, infinity +and NaNs) correctly. + +These are functions are available under C<-DECB_NO_LIBM>, since +they do not rely on the platform floating point format. The +C and C functions are +usually what you want. + =back =head2 ARITHMETIC -=over 4 +=over =item x = ecb_mod (m, n) @@ -774,8 +1249,8 @@ type (this typically excludes the minimum signed integer value, the same limitation as for C and C<%> in C). -Current GCC versions compile this into an efficient branchless sequence on -almost all CPUs. +Current GCC/clang versions compile this into an efficient branchless +sequence on almost all CPUs. For example, when you want to rotate forward through the members of an array for increasing C (which might be negative), then you should use @@ -798,7 +1273,7 @@ =head2 UTILITY -=over 4 +=over =item element_count = ecb_array_length (name) @@ -816,7 +1291,7 @@ These symbols need to be defined before including F the first time. -=over 4 +=over =item ECB_NO_THREADS @@ -842,4 +1317,21 @@ =back +=head1 UNDOCUMENTED FUNCTIONALITY + +F is full of undocumented functionality as well, some of which is +intended to be internal-use only, some of which we forgot to document, and +some of which we hide because we are not sure we will keep the interface +stable. + +While you are welcome to rummage around and use whatever you find useful +(we don't want to stop you), keep in mind that we will change undocumented +functionality in incompatible ways without thinking twice, while we are +considerably more conservative with documented things. + +=head1 AUTHORS + +C is designed and maintained by: + Emanuele Giaquinta + Marc Alexander Lehmann