[ViewVC] Diff of: cvs/libecb/ecb.pod

Comparing libecb/ecb.pod (file contents):
Revision 1.81 by root, Mon Jan 20 21:01:29 2020 UTC vs.
Revision 1.105 by root, Fri Mar 25 15:22:17 2022 UTC

…		…
10		10
11	Its homepage can be found here:	11	Its homepage can be found here:
12		12
13	http://software.schmorp.de/pkg/libecb	13	http://software.schmorp.de/pkg/libecb
14		14
15	It mainly provides a number of wrappers around GCC built-ins, together	15	It mainly provides a number of wrappers around many compiler built-ins,
16	with replacement functions for other compilers. In addition to this,	16	together with replacement functions for other compilers. In addition
17	it provides a number of other lowlevel C utilities, such as endianness	17	to this, it provides a number of other low-level C utilities, such as
18	detection, byte swapping or bit rotations.	18	endianness detection, byte swapping or bit rotations.
19		19
20	Or in other words, things that should be built into any standard C system,	20	Or in other words, things that should be built into any standard C
21	but aren't, implemented as efficient as possible with GCC, and still	21	system, but aren't, implemented as efficient as possible with GCC (clang,
22	correct with other compilers.	22	MSVC...), and still correct with other compilers.
23		23
24	More might come.	24	More might come.
25		25
26	=head2 ABOUT THE HEADER	26	=head2 ABOUT THE HEADER
27		27
…		…
56	is usually implemented as a macro. Specifically, a "bool" in this manual	56	is usually implemented as a macro. Specifically, a "bool" in this manual
57	refers to any kind of boolean value, not a specific type.	57	refers to any kind of boolean value, not a specific type.
58		58
59	=head2 TYPES / TYPE SUPPORT	59	=head2 TYPES / TYPE SUPPORT
60		60
61	ecb.h makes sure that the following types are defined (in the expected way):	61	F<ecb.h> makes sure that the following types are defined (in the expected way):
62		62
63	int8_t uint8_	63	int8_t uint8_
64	int16_t uint16_t	64	int16_t uint16_t
65	int32_t uint32_	65	int32_t uint32_
66	int64_t uint64_t	66	int64_t uint64_t
…		…
80		80
81	All the following symbols expand to an expression that can be tested in	81	All the following symbols expand to an expression that can be tested in
82	preprocessor instructions as well as treated as a boolean (use C<!!> to	82	preprocessor instructions as well as treated as a boolean (use C<!!> to
83	ensure it's either C<0> or C<1> if you need that).	83	ensure it's either C<0> or C<1> if you need that).
84		84
85	=over 4	85	=over
86		86
87	=item ECB_C	87	=item ECB_C
88		88
89	True if the implementation defines the C<__STDC__> macro to a true value,	89	True if the implementation defines the C<__STDC__> macro to a true value,
90	while not claiming to be C++.	90	while not claiming to be C++, i..e C, but not C++.
91		91
92	=item ECB_C99	92	=item ECB_C99
93		93
94	True if the implementation claims to be compliant to C99 (ISO/IEC	94	True if the implementation claims to be compliant to C99 (ISO/IEC
95	9899:1999) or any later version, while not claiming to be C++.	95	9899:1999) or any later version, while not claiming to be C++.
…		…
109		109
110	=item ECB_CPP11, ECB_CPP14, ECB_CPP17	110	=item ECB_CPP11, ECB_CPP14, ECB_CPP17
111		111
112	True if the implementation claims to be compliant to C++11/C++14/C++17	112	True if the implementation claims to be compliant to C++11/C++14/C++17
113	(ISO/IEC 14882:2011, :2014, :2017) or any later version.	113	(ISO/IEC 14882:2011, :2014, :2017) or any later version.
		114
		115	Note that many C++20 features will likely have their own feature test
		116	macros (see e.g. L<http://eel.is/c++draft/cpp.predefined#1.8>).
114		117
115	=item ECB_OPTIMIZE_SIZE	118	=item ECB_OPTIMIZE_SIZE
116		119
117	Is C<1> when the compiler optimizes for size, C<0> otherwise. This symbol	120	Is C<1> when the compiler optimizes for size, C<0> otherwise. This symbol
118	can also be defined before including F<ecb.h>, in which case it will be	121	can also be defined before including F<ecb.h>, in which case it will be
119	unchanged.	122	unchanged.
120		123
121	=item ECB_GCC_VERSION (major, minor)	124	=item ECB_GCC_VERSION (major, minor)
122		125
123	Expands to a true value (suitable for testing in by the preprocessor)	126	Expands to a true value (suitable for testing by the preprocessor) if the
124	if the compiler used is GNU C and the version is the given version, or	127	compiler used is GNU C and the version is the given version, or higher.
125	higher.
126		128
127	This macro tries to return false on compilers that claim to be GCC	129	This macro tries to return false on compilers that claim to be GCC
128	compatible but aren't.	130	compatible but aren't.
129		131
130	=item ECB_EXTERN_C	132	=item ECB_EXTERN_C
…		…
149		151
150	ECB_EXTERN_C_END	152	ECB_EXTERN_C_END
151		153
152	=item ECB_STDFP	154	=item ECB_STDFP
153		155
154	If this evaluates to a true value (suitable for testing in by the	156	If this evaluates to a true value (suitable for testing by the
155	preprocessor), then C<float> and C<double> use IEEE 754 single/binary32	157	preprocessor), then C<float> and C<double> use IEEE 754 single/binary32
156	and double/binary64 representations internally I<and> the endianness of	158	and double/binary64 representations internally I<and> the endianness of
157	both types match the endianness of C<uint32_t> and C<uint64_t>.	159	both types match the endianness of C<uint32_t> and C<uint64_t>.
158		160
159	This means you can just copy the bits of a C<float> (or C<double>) to an	161	This means you can just copy the bits of a C<float> (or C<double>) to an
…		…
161	without having to think about format or endianness.	163	without having to think about format or endianness.
162		164
163	This is true for basically all modern platforms, although F<ecb.h> might	165	This is true for basically all modern platforms, although F<ecb.h> might
164	not be able to deduce this correctly everywhere and might err on the safe	166	not be able to deduce this correctly everywhere and might err on the safe
165	side.	167	side.
		168
		169	=item ECB_64BIT_NATIVE
		170
		171	Evaluates to a true value (suitable for both preprocessor and C code
		172	testing) if 64 bit integer types on this architecture are evaluated
		173	"natively", that is, with similar speeds as 32 bit integers. While 64 bit
		174	integer support is very common (and in fact required by libecb), 32 bit
		175	CPUs have to emulate operations on them, so you might want to avoid them.
166		176
167	=item ECB_AMD64, ECB_AMD64_X32	177	=item ECB_AMD64, ECB_AMD64_X32
168		178
169	These two macros are defined to C<1> on the x86_64/amd64 ABI and the X32	179	These two macros are defined to C<1> on the x86_64/amd64 ABI and the X32
170	ABI, respectively, and undefined elsewhere.	180	ABI, respectively, and undefined elsewhere.
…		…
177		187
178	=back	188	=back
179		189
180	=head2 MACRO TRICKERY	190	=head2 MACRO TRICKERY
181		191
182	=over 4	192	=over
183		193
184	=item ECB_CONCAT (a, b)	194	=item ECB_CONCAT (a, b)
185		195
186	Expands any macros in C<a> and C<b>, then concatenates the result to form	196	Expands any macros in C<a> and C<b>, then concatenates the result to form
187	a single token. This is mainly useful to form identifiers from components,	197	a single token. This is mainly useful to form identifiers from components,
…		…
228	declarations must be put before the whole declaration:	238	declarations must be put before the whole declaration:
229		239
230	ecb_const int mysqrt (int a);	240	ecb_const int mysqrt (int a);
231	ecb_unused int i;	241	ecb_unused int i;
232		242
233	=over 4	243	=over
234		244
235	=item ecb_unused	245	=item ecb_unused
236		246
237	Marks a function or a variable as "unused", which simply suppresses a	247	Marks a function or a variable as "unused", which simply suppresses a
238	warning by GCC when it detects it as unused. This is useful when you e.g.	248	warning by the compiler when it detects it as unused. This is useful when
239	declare a variable but do not always use it:	249	you e.g. declare a variable but do not always use it:
240		250
241	{	251	{
242	ecb_unused int var;	252	ecb_unused int var;
243		253
244	#ifdef SOMECONDITION	254	#ifdef SOMECONDITION
…		…
264		274
265	Expands either to (a compiler-specific equivalent of) C<static inline> or	275	Expands either to (a compiler-specific equivalent of) C<static inline> or
266	to just C<static>, if inline isn't supported. It should be used to declare	276	to just C<static>, if inline isn't supported. It should be used to declare
267	functions that should be inlined, for code size or speed reasons.	277	functions that should be inlined, for code size or speed reasons.
268		278
269	Example: inline this function, it surely will reduce codesize.	279	Example: inline this function, it surely will reduce code size.
270		280
271	ecb_inline int	281	ecb_inline int
272	negmul (int a, int b)	282	negmul (int a, int b)
273	{	283	{
274	return - (a * b);	284	return - (a * b);
…		…
374	speed-critical times, and keeping it in the cache might be a waste of said	384	speed-critical times, and keeping it in the cache might be a waste of said
375	cache.	385	cache.
376		386
377	In addition to placing cold functions together (or at least away from hot	387	In addition to placing cold functions together (or at least away from hot
378	functions), this knowledge can be used in other ways, for example, the	388	functions), this knowledge can be used in other ways, for example, the
379	function will be optimised for size, as opposed to speed, and codepaths	389	function will be optimised for size, as opposed to speed, and code paths
380	leading to calls to those functions can automatically be marked as if	390	leading to calls to those functions can automatically be marked as if
381	C<ecb_expect_false> had been used to reach them.	391	C<ecb_expect_false> had been used to reach them.
382		392
383	Good examples for such functions would be error reporting functions, or	393	Good examples for such functions would be error reporting functions, or
384	functions only called in exceptional or rare cases.	394	functions only called in exceptional or rare cases.
…		…
412		422
413	=back	423	=back
414		424
415	=head2 OPTIMISATION HINTS	425	=head2 OPTIMISATION HINTS
416		426
417	=over 4	427	=over
418		428
419	=item bool ecb_is_constant (expr)	429	=item bool ecb_is_constant (expr)
420		430
421	Returns true iff the expression can be deduced to be a compile-time	431	Returns true iff the expression can be deduced to be a compile-time
422	constant, and false otherwise.	432	constant, and false otherwise.
…		…
538	never be executed. Apart from suppressing a warning in some cases, this	548	never be executed. Apart from suppressing a warning in some cases, this
539	function can be used to implement C<ecb_assume> or similar functionality.	549	function can be used to implement C<ecb_assume> or similar functionality.
540		550
541	=item ecb_prefetch (addr, rw, locality)	551	=item ecb_prefetch (addr, rw, locality)
542		552
543	Tells the compiler to try to prefetch memory at the given C<addr>ess	553	Tells the compiler to try to prefetch memory at the given I<addr>ess
544	for either reading (C<rw> = 0) or writing (C<rw> = 1). A C<locality> of	554	for either reading (I<rw> = 0) or writing (I<rw> = 1). A I<locality> of
545	C<0> means that there will only be one access later, C<3> means that	555	C<0> means that there will only be one access later, C<3> means that
546	the data will likely be accessed very often, and values in between mean	556	the data will likely be accessed very often, and values in between mean
547	something... in between. The memory pointed to by the address does not	557	something... in between. The memory pointed to by the address does not
548	need to be accessible (it could be a null pointer for example), but C<rw>	558	need to be accessible (it could be a null pointer for example), but C<rw>
549	and C<locality> must be compile-time constants.	559	and C<locality> must be compile-time constants.
…		…
579		589
580	=back	590	=back
581		591
582	=head2 BIT FIDDLING / BIT WIZARDRY	592	=head2 BIT FIDDLING / BIT WIZARDRY
583		593
584	=over 4	594	=over
585		595
586	=item bool ecb_big_endian ()	596	=item bool ecb_big_endian ()
587		597
588	=item bool ecb_little_endian ()	598	=item bool ecb_little_endian ()
589		599
…		…
610		620
611	For example:	621	For example:
612		622
613	ecb_ctz32 (3) = 0	623	ecb_ctz32 (3) = 0
614	ecb_ctz32 (6) = 1	624	ecb_ctz32 (6) = 1
		625
		626	=item int ecb_clz32 (uint32_t x)
		627
		628	=item int ecb_clz64 (uint64_t x)
		629
		630	=item int ecb_clz (T x) [C++]
		631
		632	Counts the number of leading zero bits in C<x>. If C<x> is 0 the result is
		633	undefined.
		634
		635	The overloaded C++ C<ecb_clz> function supports C<uint32_t> and
		636	C<uint64_t> types only.
		637
		638	It is often simpler to use one of the C<ecb_ld*> functions instead, whoise
		639	result only depends on the value and not the size of the type.
		640
		641	For example:
		642
		643	ecb_clz32 (3) = 30
		644	ecb_clz32 (6) = 29
615		645
616	=item bool ecb_is_pot32 (uint32_t x)	646	=item bool ecb_is_pot32 (uint32_t x)
617		647
618	=item bool ecb_is_pot64 (uint32_t x)	648	=item bool ecb_is_pot64 (uint32_t x)
619		649
…		…
721		751
722	=item uint64_t ecb_rotr64 (uint64_t x, unsigned int count)	752	=item uint64_t ecb_rotr64 (uint64_t x, unsigned int count)
723		753
724	These two families of functions return the value of C<x> after rotating	754	These two families of functions return the value of C<x> after rotating
725	all the bits by C<count> positions to the right (C<ecb_rotr>) or left	755	all the bits by C<count> positions to the right (C<ecb_rotr>) or left
726	(C<ecb_rotl>).	756	(C<ecb_rotl>). There are no restrictions on the value C<count>, i.e. both
		757	zero and values equal or larger than the word width work correctly. Also,
		758	notwithstanding C<count> being unsigned, negative numbers work and shift
		759	to the opposite direction.
727		760
728	Current GCC versions understand these functions and usually compile them	761	Current GCC/clang versions understand these functions and usually compile
729	to "optimal" code (e.g. a single C<rol> or a combination of C<shld> on	762	them to "optimal" code (e.g. a single C<rol> or a combination of C<shld>
730	x86).	763	on x86).
731		764
732	=item T ecb_rotl (T x, unsigned int count) [C++]	765	=item T ecb_rotl (T x, unsigned int count) [C++]
733		766
734	=item T ecb_rotr (T x, unsigned int count) [C++]	767	=item T ecb_rotr (T x, unsigned int count) [C++]
735		768
736	Overloaded C++ rotl/rotr functions.	769	Overloaded C++ rotl/rotr functions.
737		770
738	C<T> must be one of C<uint8_t>, C<uint16_t>, C<uint32_t> or C<uint64_t>.	771	C<T> must be one of C<uint8_t>, C<uint16_t>, C<uint32_t> or C<uint64_t>.
739		772
		773	=item uint_fast8_t ecb_gray8_encode (uint_fast8_t b)
		774
		775	=item uint_fast16_t ecb_gray16_encode (uint_fast16_t b)
		776
		777	=item uint_fast32_t ecb_gray32_encode (uint_fast32_t b)
		778
		779	=item uint_fast64_t ecb_gray64_encode (uint_fast64_t b)
		780
		781	Encode an unsigned into its corresponding (reflective) gray code - the
		782	kind of gray code meant when just talking about "gray code". These
		783	functions are very fast and all have identical implementation, so there is
		784	no need to use a smaller type, as long as your CPU can handle it natively.
		785
		786	=item T ecb_gray_encode (T b) [C++]
		787
		788	Overloaded C++ version of the above, for C<uint{8,16,32,64}_t>.
		789
		790	=item uint_fast8_t ecb_gray8_decode (uint_fast8_t b)
		791
		792	=item uint_fast16_t ecb_gray16_decode (uint_fast16_t b)
		793
		794	=item uint_fast32_t ecb_gray32_decode (uint_fast32_t b)
		795
		796	=item uint_fast64_t ecb_gray64_decode (uint_fast64_t b)
		797
		798	Decode a gray code back into linear index form (the reverse of
		799	C<ecb_gray*_encode>. Unlike the encode functions, the decode functions
		800	have higher time complexity for larger types, so it can pay off to use a
		801	smaller type here.
		802
		803	=item T ecb_gray_decode (T b) [C++]
		804
		805	Overloaded C++ version of the above, for C<uint{8,16,32,64}_t>.
		806
		807	=back
		808
		809	=head2 HILBERT CURVES
		810
		811	These functions deal with (square, pseudo) Hilbert curves. The parameter
		812	I<order> indicates the size of the square and is specified in bits, that
		813	means for order C<8>, the coordinates range from C<0>..C<255>, and the
		814	curve index ranges from C<0>..C<65535>.
		815
		816	The 32 bit variants of these functions map a 32 bit index to two 16 bit
		817	coordinates, stored in a 32 bit variable, where the high order bits are
		818	the x-coordinate, and the low order bits are the y-coordinate, thus,
		819	these functions map 32 bit linear index on the curve to a 32 bit packed
		820	coordinate pair, and vice versa.
		821
		822	The 64 bit variants work similarly.
		823
		824	The I<order> can go from C<1> to C<16> for the 32 bit curve, and C<1> to
		825	C<32> for the 64 bit curve.
		826
		827	When going from one order to the next higher order, these functions
		828	replace the curve segments by smaller versions of the generating shape,
		829	while doubling the size (since they use integer coordinates), which is
		830	what you would expect mathematically. This means that the curve will be
		831	mirrored at the diagonal. If your goal is to simply cover more area while
		832	retaining existing point coordinates you should increase or decrease the
		833	I<order> by C<2> or, in the case of C<ecb_hilbert2d_index_to_coord>,
		834	simply specify the maximum I<order> of C<16> or C<32>, respectively, as
		835	these are constant-time.
		836
		837	=over
		838
		839	=item uint32_t ecb_hilbert2d_index_to_coord32 (int order, uint32_t index)
		840
		841	=item uint64_t ecb_hilbert2d_index_to_coord64 (int order, uint64_t index)
		842
		843	Map a point on a pseudo Hilbert curve from its linear distance from the
		844	origin on the curve to a x\|y coordinate pair. The result is a packed
		845	coordinate pair, to get the actual x and < coordinates, you could do
		846	something like this:
		847
		848	uint32_t xy = ecb_hilbert2d_index_to_coord32 (16, 255);
		849	uint16_t x = xy >> 16;
		850	uint16_t y = xy & 0xffffU;
		851
		852	uint64_t xy = ecb_hilbert2d_index_to_coord64 (32, 255);
		853	uint32_t x = xy >> 32;
		854	uint32_t y = xy & 0xffffffffU;
		855
		856	These functions work in constant time, so for many applications it is
		857	preferable to simply hard-code the order to the maximum (C<16> or C<32>).
		858
		859	This (production-ready, i.e. never run) example generates an SVG image of
		860	an order 8 pseudo Hilbert curve:
		861
		862	printf ("<svg xmlns='http://www.w3.org/2000/svg' width='%d' height='%d'>\n", 64 * 8, 64 * 8);
		863	printf ("<g transform='translate(4) scale(8)' stroke-width='0.25' stroke='black'>\n");
		864	for (uint32_t i = 0; i < 64*64 - 1; ++i)
		865	{
		866	uint32_t p1 = ecb_hilbert2d_index_to_coord32 (6, i );
		867	uint32_t p2 = ecb_hilbert2d_index_to_coord32 (6, i + 1);
		868	printf ("<line x1='%d' y1='%d' x2='%d' y2='%d'/>\n",
		869	p1 >> 16, p1 & 0xffff,
		870	p2 >> 16, p2 & 0xffff);
		871	}
		872	printf ("</g>\n");
		873	printf ("</svg>\n");
		874
		875	=item uint32_t ecb_hilbert2d_coord_to_index32 (int order, uint32_t xy)
		876
		877	=item uint64_t ecb_hilbert2d_coord_to_index64 (int order, uint64_t xy)
		878
		879	The reverse of C<ecb_hilbert2d_index_to_coord> - map a packed pair of
		880	coordinates to their linear index on the pseudo Hilbert curve of order
		881	I<order>.
		882
		883	They are an exact inverse of the C<ecb_hilbert2d_coord_to_index> functions
		884	for the same I<order>:
		885
		886	assert (
		887	u == ecb_hilbert2d_coord_to_index (32,
		888	ecb_hilbert2d_index_to_coord32 (32,
		889	u)));
		890
		891	Packing coordinates is done the same way, as well, from I<x> and I<y>:
		892
		893	uint32_t xy = ((uint32_t)x << 16) \| y; // for ecb_hilbert2d_coord_to_index32
		894	uint64_t xy = ((uint64_t)x << 32) \| y; // for ecb_hilbert2d_coord_to_index64
		895
		896	Unlike C<ecb_hilbert2d_coord_to_index>, these functions are O(I<order>),
		897	so it is preferable to use the lowest possible order.
		898
		899	=back
		900
		901	=head2 BIT MIXING, HASHING
		902
		903	Sometimes you have an integer and want to distribute its bits well, for
		904	example, to use it as a hash in a hash table. A common example is pointer
		905	values, which often only have a limited range (e.g. low and high bits are
		906	often zero).
		907
		908	The following functions try to mix the bits to get a good bias-free
		909	distribution. They were mainly made for pointers, but the underlying
		910	integer functions are exposed as well.
		911
		912	As an added benefit, the functions are reversible, so if you find it
		913	convenient to store only the hash value, you can recover the original
		914	pointer from the hash ("unmix"), as long as your pointers are 32 or 64 bit
		915	(if this isn't the case on your platform, drop us a note and we will add
		916	functions for other bit widths).
		917
		918	The unmix functions are very slightly slower than the mix functions, so
		919	it is equally very slightly preferable to store the original values wehen
		920	convenient.
		921
		922	The underlying algorithm if subject to change, so currently these
		923	functions are not suitable for persistent hash tables, as their result
		924	value can change between different versions of libecb.
		925
		926	=over
		927
		928	=item uintptr_t ecb_ptrmix (void *ptr)
		929
		930	Mixes the bits of a pointer so the result is suitable for hash table
		931	lookups. In other words, this hashes the pointer value.
		932
		933	=item uintptr_t ecb_ptrmix (T *ptr) [C++]
		934
		935	Overload the C<ecb_ptrmix> function to work for any pointer in C++.
		936
		937	=item void *ecb_ptrunmix (uintptr_t v)
		938
		939	Unmix the hash value into the original pointer. This only works as long
		940	as the hash value is not truncated, i.e. you used C<uintptr_t> (or
		941	equivalent) throughout to store it.
		942
		943	=item T *ecb_ptrunmix<T> (uintptr_t v) [C++]
		944
		945	The somewhat less useful template version of C<ecb_ptrunmix> for
		946	C++. Example:
		947
		948	sometype *myptr;
		949	uintptr_t hash = ecb_ptrmix (myptr);
		950	sometype *orig = ecb_ptrunmix<sometype> (hash);
		951
		952	=item uint32_t ecb_mix32 (uint32_t v)
		953
		954	=item uint64_t ecb_mix64 (uint64_t v)
		955
		956	Sometimes you don't have a pointer but an integer whose values are very
		957	badly distributed. In this case you can use these integer versions of the
		958	mixing function. No C++ template is provided currently.
		959
		960	=item uint32_t ecb_unmix32 (uint32_t v)
		961
		962	=item uint64_t ecb_unmix64 (uint64_t v)
		963
		964	The reverse of the C<ecb_mix> functions - they take a mixed/hashed value
		965	and recover the original value.
		966
740	=back	967	=back
741		968
742	=head2 HOST ENDIANNESS CONVERSION	969	=head2 HOST ENDIANNESS CONVERSION
743		970
744	=over 4	971	=over
745		972
746	=item uint_fast16_t ecb_be_u16_to_host (uint_fast16_t v)	973	=item uint_fast16_t ecb_be_u16_to_host (uint_fast16_t v)
747		974
748	=item uint_fast32_t ecb_be_u32_to_host (uint_fast32_t v)	975	=item uint_fast32_t ecb_be_u32_to_host (uint_fast32_t v)
749		976
…		…
777		1004
778	=back	1005	=back
779		1006
780	In C++ the following additional template functions are supported:	1007	In C++ the following additional template functions are supported:
781		1008
782	=over 4	1009	=over
783		1010
784	=item T ecb_be_to_host (T v)	1011	=item T ecb_be_to_host (T v)
785		1012
786	=item T ecb_le_to_host (T v)	1013	=item T ecb_le_to_host (T v)
787		1014
788	=item T ecb_host_to_be (T v)	1015	=item T ecb_host_to_be (T v)
789		1016
790	=item T ecb_host_to_le (T v)	1017	=item T ecb_host_to_le (T v)
		1018
		1019	=back
791		1020
792	These functions work like their C counterparts, above, but use templates,	1021	These functions work like their C counterparts, above, but use templates,
793	which make them useful in generic code.	1022	which make them useful in generic code.
794		1023
795	C<T> must be one of C<uint8_t>, C<uint16_t>, C<uint32_t> or C<uint64_t>	1024	C<T> must be one of C<uint8_t>, C<uint16_t>, C<uint32_t> or C<uint64_t>
…		…
798		1027
799	=head2 UNALIGNED LOAD/STORE	1028	=head2 UNALIGNED LOAD/STORE
800		1029
801	These function load or store unaligned multi-byte values.	1030	These function load or store unaligned multi-byte values.
802		1031
803	=over 4	1032	=over
804		1033
805	=item uint_fast16_t ecb_peek_u16_u (const void *ptr)	1034	=item uint_fast16_t ecb_peek_u16_u (const void *ptr)
806		1035
807	=item uint_fast32_t ecb_peek_u32_u (const void *ptr)	1036	=item uint_fast32_t ecb_peek_u32_u (const void *ptr)
808		1037
…		…
852		1081
853	=back	1082	=back
854		1083
855	In C++ the following additional template functions are supported:	1084	In C++ the following additional template functions are supported:
856		1085
857	=over 4	1086	=over
858		1087
859	=item T ecb_peek<T> (const void *ptr)	1088	=item T ecb_peek<T> (const void *ptr)
860		1089
861	=item T ecb_peek_be<T> (const void *ptr)	1090	=item T ecb_peek_be<T> (const void *ptr)
862		1091
…		…
893	=item ecb_poke_be_u (void *ptr, T v)	1122	=item ecb_poke_be_u (void *ptr, T v)
894		1123
895	=item ecb_poke_le_u (void *ptr, T v)	1124	=item ecb_poke_le_u (void *ptr, T v)
896		1125
897	Again, similarly to their C counterparts, these functions store an	1126	Again, similarly to their C counterparts, these functions store an
898	unsigned 8, 16, 32 or z64 bit value to memory, with optional conversion to	1127	unsigned 8, 16, 32 or 64 bit value to memory, with optional conversion to
899	big/little endian.	1128	big/little endian.
900		1129
901	C<T> must be one of C<uint8_t>, C<uint16_t>, C<uint32_t> or C<uint64_t>.	1130	C<T> must be one of C<uint8_t>, C<uint16_t>, C<uint32_t> or C<uint64_t>.
902		1131
903	Unlike their C counterparts, these functions support 8 bit quantities	1132	Unlike their C counterparts, these functions support 8 bit quantities
904	(C<uint8_t>) and also have an aligned version (without the C<_u> prefix),	1133	(C<uint8_t>) and also have an aligned version (without the C<_u> prefix),
905	all of which hopefully makes them more useful in generic code.	1134	all of which hopefully makes them more useful in generic code.
906		1135
907	=back	1136	=back
908		1137
		1138	=head2 FAST INTEGER TO STRING
		1139
		1140	Libecb defines a set of very fast integer to decimal string (or integer
		1141	to ASCII, short C<i2a>) functions. These work by converting the integer
		1142	to a fixed point representation and then successively multiplying out
		1143	the topmost digits. Unlike some other, also very fast, libraries, ecb's
		1144	algorithm should be completely branchless per digit, and does not rely on
		1145	the presence of special CPU functions (such as C<clz>).
		1146
		1147	There is a high level API that takes an C<int32_t>, C<uint32_t>,
		1148	C<int64_t> or C<uint64_t> as argument, and a low-level API, which is
		1149	harder to use but supports slightly more formatting options.
		1150
		1151	=head3 HIGH LEVEL API
		1152
		1153	The high level API consists of four functions, one each for C<int32_t>,
		1154	C<uint32_t>, C<int64_t> and C<uint64_t>:
		1155
		1156	Example:
		1157
		1158	char buf[ECB_I2A_MAX_DIGITS + 1];
		1159	char *end = ecb_i2a_i32 (buf, 17262);
		1160	*end = 0;
		1161	// buf now contains "17262"
		1162
		1163	=over
		1164
		1165	=item ECB_I2A_I32_DIGITS (=11)
		1166
		1167	=item char ecb_i2a_u32 (char ptr, uint32_t value)
		1168
		1169	Takes an C<uint32_t> I<value> and formats it as a decimal number starting
		1170	at I<ptr>, using at most C<ECB_I2A_I32_DIGITS> characters. Returns a
		1171	pointer to just after the generated string, where you would normally put
		1172	the terminating C<0> character. This function outputs the minimum number
		1173	of digits.
		1174
		1175	=item ECB_I2A_U32_DIGITS (=10)
		1176
		1177	=item char ecb_i2a_i32 (char ptr, int32_t value)
		1178
		1179	Same as C<ecb_i2a_u32>, but formats a C<int32_t> value, including a minus
		1180	sign if needed.
		1181
		1182	=item ECB_I2A_I64_DIGITS (=20)
		1183
		1184	=item char ecb_i2a_u64 (char ptr, uint64_t value)
		1185
		1186	=item ECB_I2A_U64_DIGITS (=21)
		1187
		1188	=item char ecb_i2a_i64 (char ptr, int64_t value)
		1189
		1190	Similar to their 32 bit counterparts, these take a 64 bit argument.
		1191
		1192	=item ECB_I2A_MAX_DIGITS (=21)
		1193
		1194	Instead of using a type specific length macro, you can just use
		1195	C<ECB_I2A_MAX_DIGITS>, which is good enough for any C<ecb_i2a> function.
		1196
		1197	=back
		1198
		1199	=head3 LOW-LEVEL API
		1200
		1201	The functions above use a number of low-level APIs which have some strict
		1202	limitations, but can be used as building blocks (studying C<ecb_i2a_i32>
		1203	and related functions is recommended).
		1204
		1205	There are three families of functions: functions that convert a number
		1206	to a fixed number of digits with leading zeroes (C<ecb_i2a_0N>, C<0>
		1207	for "leading zeroes"), functions that generate up to N digits, skipping
		1208	leading zeroes (C<_N>), and functions that can generate more digits, but
		1209	the leading digit has limited range (C<_xN>).
		1210
		1211	None of the functions deal with negative numbers.
		1212
		1213	Example: convert an IP address in an C<uint32_t> into dotted-quad:
		1214
		1215	uint32_t ip = 0x0a000164; // 10.0.1.100
		1216	char ips[3 * 4 + 3 + 1];
		1217	char *ptr = ips;
		1218	ptr = ecb_i2a_3 (ptr, ip >> 24 ); *ptr++ = '.';
		1219	ptr = ecb_i2a_3 (ptr, (ip >> 16) & 0xff); *ptr++ = '.';
		1220	ptr = ecb_i2a_3 (ptr, (ip >> 8) & 0xff); *ptr++ = '.';
		1221	ptr = ecb_i2a_3 (ptr, ip & 0xff); *ptr++ = 0;
		1222	printf ("ip: %s\n", ips); // prints "ip: 10.0.1.100"
		1223
		1224	=over
		1225
		1226	=item char ecb_i2a_02 (char ptr, uint32_t value) // 32 bit
		1227
		1228	=item char ecb_i2a_03 (char ptr, uint32_t value) // 32 bit
		1229
		1230	=item char ecb_i2a_04 (char ptr, uint32_t value) // 32 bit
		1231
		1232	=item char ecb_i2a_05 (char ptr, uint32_t value) // 64 bit
		1233
		1234	=item char ecb_i2a_06 (char ptr, uint32_t value) // 64 bit
		1235
		1236	=item char ecb_i2a_07 (char ptr, uint32_t value) // 64 bit
		1237
		1238	=item char ecb_i2a_08 (char ptr, uint32_t value) // 64 bit
		1239
		1240	=item char ecb_i2a_09 (char ptr, uint32_t value) // 64 bit
		1241
		1242	The C<< ecb_i2a_0I<N> >> functions take an unsigned I<value> and convert
		1243	them to exactly I<N> digits, returning a pointer to the first character
		1244	after the digits. The I<value> must be in range. The functions marked with
		1245	I<32 bit> do their calculations internally in 32 bit, the ones marked with
		1246	I<64 bit> internally use 64 bit integers, which might be slow on 32 bit
		1247	architectures (the high level API decides on 32 vs. 64 bit versions using
		1248	C<ECB_64BIT_NATIVE>).
		1249
		1250	=item char ecb_i2a_2 (char ptr, uint32_t value) // 32 bit
		1251
		1252	=item char ecb_i2a_3 (char ptr, uint32_t value) // 32 bit
		1253
		1254	=item char ecb_i2a_4 (char ptr, uint32_t value) // 32 bit
		1255
		1256	=item char ecb_i2a_5 (char ptr, uint32_t value) // 64 bit
		1257
		1258	=item char ecb_i2a_6 (char ptr, uint32_t value) // 64 bit
		1259
		1260	=item char ecb_i2a_7 (char ptr, uint32_t value) // 64 bit
		1261
		1262	=item char ecb_i2a_8 (char ptr, uint32_t value) // 64 bit
		1263
		1264	=item char ecb_i2a_9 (char ptr, uint32_t value) // 64 bit
		1265
		1266	Similarly, the C<< ecb_i2a_I<N> >> functions take an unsigned I<value>
		1267	and convert them to at most I<N> digits, suppressing leading zeroes, and
		1268	returning a pointer to the first character after the digits.
		1269
		1270	=item ECB_I2A_MAX_X5 (=59074)
		1271
		1272	=item char ecb_i2a_x5 (char ptr, uint32_t value) // 32 bit
		1273
		1274	=item ECB_I2A_MAX_X10 (=2932500665)
		1275
		1276	=item char ecb_i2a_x10 (char ptr, uint32_t value) // 64 bit
		1277
		1278	The C<< ecb_i2a_xI<N> >> functions are similar to the C<< ecb_i2a_I<N> >>
		1279	functions, but they can generate one digit more, as long as the number
		1280	is within range, which is given by the symbols C<ECB_I2A_MAX_X5> (almost
		1281	16 bit range) and C<ECB_I2A_MAX_X10> (a bit more than 31 bit range),
		1282	respectively.
		1283
		1284	For example, the digit part of a 32 bit signed integer just fits into the
		1285	C<ECB_I2A_MAX_X10> range, so while C<ecb_i2a_x10> cannot convert a 10
		1286	digit number, it can convert all 32 bit signed numbers. Sadly, it's not
		1287	good enough for 32 bit unsigned numbers.
		1288
		1289	=back
		1290
909	=head2 FLOATING POINT FIDDLING	1291	=head2 FLOATING POINT FIDDLING
910		1292
911	=over 4	1293	=over
912		1294
913	=item ECB_INFINITY [-UECB_NO_LIBM]	1295	=item ECB_INFINITY [-UECB_NO_LIBM]
914		1296
915	Evaluates to positive infinity if supported by the platform, otherwise to	1297	Evaluates to positive infinity if supported by the platform, otherwise to
916	a truly huge number.	1298	a truly huge number.
…		…
941	IEEE compliant, of course at a speed and code size penalty, and of course	1323	IEEE compliant, of course at a speed and code size penalty, and of course
942	also within reasonable limits (it tries to convert NaNs, infinities and	1324	also within reasonable limits (it tries to convert NaNs, infinities and
943	denormals, but will likely convert negative zero to positive zero).	1325	denormals, but will likely convert negative zero to positive zero).
944		1326
945	On all modern platforms (where C<ECB_STDFP> is true), the compiler should	1327	On all modern platforms (where C<ECB_STDFP> is true), the compiler should
946	be able to optimise away this function completely.	1328	be able to completely optimise away the 32 and 64 bit functions.
947		1329
948	These functions can be helpful when serialising floats to the network - you	1330	These functions can be helpful when serialising floats to the network - you
949	can serialise the return value like a normal uint16_t/uint32_t/uint64_t.	1331	can serialise the return value like a normal uint16_t/uint32_t/uint64_t.
950		1332
951	Another use for these functions is to manipulate floating point values	1333	Another use for these functions is to manipulate floating point values
…		…
994		1376
995	=back	1377	=back
996		1378
997	=head2 ARITHMETIC	1379	=head2 ARITHMETIC
998		1380
999	=over 4	1381	=over
1000		1382
1001	=item x = ecb_mod (m, n)	1383	=item x = ecb_mod (m, n)
1002		1384
1003	Returns C<m> modulo C<n>, which is the same as the positive remainder	1385	Returns C<m> modulo C<n>, which is the same as the positive remainder
1004	of the division operation between C<m> and C<n>, using floored	1386	of the division operation between C<m> and C<n>, using floored
…		…
1011	C<n> must be strictly positive (i.e. C<< >= 1 >>), while C<m> must be	1393	C<n> must be strictly positive (i.e. C<< >= 1 >>), while C<m> must be
1012	negatable, that is, both C<m> and C<-m> must be representable in its	1394	negatable, that is, both C<m> and C<-m> must be representable in its
1013	type (this typically excludes the minimum signed integer value, the same	1395	type (this typically excludes the minimum signed integer value, the same
1014	limitation as for C</> and C<%> in C).	1396	limitation as for C</> and C<%> in C).
1015		1397
1016	Current GCC versions compile this into an efficient branchless sequence on	1398	Current GCC/clang versions compile this into an efficient branchless
1017	almost all CPUs.	1399	sequence on almost all CPUs.
1018		1400
1019	For example, when you want to rotate forward through the members of an	1401	For example, when you want to rotate forward through the members of an
1020	array for increasing C<m> (which might be negative), then you should use	1402	array for increasing C<m> (which might be negative), then you should use
1021	C<ecb_mod>, as the C<%> operator might give either negative results, or	1403	C<ecb_mod>, as the C<%> operator might give either negative results, or
1022	change direction for negative values:	1404	change direction for negative values:
…		…
1035		1417
1036	=back	1418	=back
1037		1419
1038	=head2 UTILITY	1420	=head2 UTILITY
1039		1421
1040	=over 4	1422	=over
1041		1423
1042	=item element_count = ecb_array_length (name)	1424	=item element_count = ecb_array_length (name)
1043		1425
1044	Returns the number of elements in the array C<name>. For example:	1426	Returns the number of elements in the array C<name>. For example:
1045		1427
…		…
1053		1435
1054	=head2 SYMBOLS GOVERNING COMPILATION OF ECB.H ITSELF	1436	=head2 SYMBOLS GOVERNING COMPILATION OF ECB.H ITSELF
1055		1437
1056	These symbols need to be defined before including F<ecb.h> the first time.	1438	These symbols need to be defined before including F<ecb.h> the first time.
1057		1439
1058	=over 4	1440	=over
1059		1441
1060	=item ECB_NO_THREADS	1442	=item ECB_NO_THREADS
1061		1443
1062	If F<ecb.h> is never used from multiple threads, then this symbol can	1444	If F<ecb.h> is never used from multiple threads, then this symbol can
1063	be defined, in which case memory fences (and similar constructs) are	1445	be defined, in which case memory fences (and similar constructs) are
…		…
1067		1449
1068	=item ECB_NO_SMP	1450	=item ECB_NO_SMP
1069		1451
1070	The weaker version of C<ECB_NO_THREADS> - if F<ecb.h> is used from	1452	The weaker version of C<ECB_NO_THREADS> - if F<ecb.h> is used from
1071	multiple threads, but never concurrently (e.g. if the system the program	1453	multiple threads, but never concurrently (e.g. if the system the program
1072	runs on has only a single CPU with a single core, no hyperthreading and so	1454	runs on has only a single CPU with a single core, no hyper-threading and so
1073	on), then this symbol can be defined, leading to more efficient code and	1455	on), then this symbol can be defined, leading to more efficient code and
1074	fewer dependencies.	1456	fewer dependencies.
1075		1457
1076	=item ECB_NO_LIBM	1458	=item ECB_NO_LIBM
1077		1459
…		…
1087	intended to be internal-use only, some of which we forgot to document, and	1469	intended to be internal-use only, some of which we forgot to document, and
1088	some of which we hide because we are not sure we will keep the interface	1470	some of which we hide because we are not sure we will keep the interface
1089	stable.	1471	stable.
1090		1472
1091	While you are welcome to rummage around and use whatever you find useful	1473	While you are welcome to rummage around and use whatever you find useful
1092	(we can't stop you), keep in mind that we will change undocumented	1474	(we don't want to stop you), keep in mind that we will change undocumented
1093	functionality in incompatible ways without thinking twice, while we are	1475	functionality in incompatible ways without thinking twice, while we are
1094	considerably more conservative with documented things.	1476	considerably more conservative with documented things.
1095		1477
1096	=head1 AUTHORS	1478	=head1 AUTHORS
1097		1479
1098	C<libecb> is designed and maintained by:	1480	C<libecb> is designed and maintained by:
1099		1481
1100	Emanuele Giaquinta <e.giaquinta@glauco.it>	1482	Emanuele Giaquinta <e.giaquinta@glauco.it>
1101	Marc Alexander Lehmann <schmorp@schmorp.de>	1483	Marc Alexander Lehmann <schmorp@schmorp.de>
1102
1103

Diff Legend

-–
+Removed lines
-+
+Added lines
-<
+Changed lines
->
+Changed lines

Comparing libecb/ecb.pod (file contents): Revision 1.81 by root, Mon Jan 20 21:01:29 2020 UTC vs. Revision 1.105 by root, Fri Mar 25 15:22:17 2022 UTC

Diff Legend

Comparing libecb/ecb.pod (file contents):
Revision 1.81 by root, Mon Jan 20 21:01:29 2020 UTC vs.
Revision 1.105 by root, Fri Mar 25 15:22:17 2022 UTC