| 1 |
TODO: naming of cutnions, index_to_coord, xy_to_s, xy_to_linear/seq, 32/64 |
| 2 |
TODO: returning packed coordinates, really |
| 3 |
TODO: support order=0? |
| 4 |
|
| 5 |
- chatgpt monday approved. |
| 6 |
- ecb_i2a_u32 etc. did not advance the pointer always on !64bit. |
| 7 |
|
| 8 |
0x0001000c |
| 9 |
- added ecb_hilbert2d function family. |
| 10 |
- use __builtin_popcountl on 64 bit clang and gcc. |
| 11 |
- use 64 bit builtins, not 128 bit builtins, for ctz/clz on 64 bit cpus. |
| 12 |
- add ecb_clz* functions, |
| 13 |
- renamed ecb_gray* functions to have trailing bit width. |
| 14 |
- use ecb_function_ consistently for newly added functions. |
| 15 |
- say goodbye to gnu-style function definitions - they were not |
| 16 |
used consistently anyway, and imho make the code harder to read, |
| 17 |
and didn't even work for indented definitions. |
| 18 |
- windows/ms versions of clz returned the wrongt value (reported by b_honas). |
| 19 |
|
| 20 |
0x0001000b |
| 21 |
- added ecb_gray*_{de,en}code functions. |
| 22 |
|
| 23 |
0x0001000a |
| 24 |
- added ecb_ptrmix. |
| 25 |
|
| 26 |
0x00010009 |
| 27 |
- added ecb_i2a family of functions. |
| 28 |
- added ECB_64BIT_NATIVE. |
| 29 |
|
| 30 |
TODO: |
| 31 |
08:30:06 <b_jonas> I think it could be worth to add a macro that works like alignof or _Alignof on sane |
| 32 |
compilers, and like __alignof on MS compilers that support it, see |
| 33 |
http://msdn.microsoft.com/en-us/library/45t0s5f4.aspx |
| 34 |
08:30:24 <b_jonas> even if you can't support it on all the old compilers |
| 35 |
08:31:17 <b_jonas> I'd also like a macro for alignas, but sadly, that seems impossible in general, because |
| 36 |
the MS compiler only has some half-attempt to do something similar but with different and |
| 37 |
more broken semantics, see http://msdn.microsoft.com/en-us/library/83ythb65.aspx |
| 38 |
08:31:35 <b_jonas> but I wonder if some special case could still be worth to support |
| 39 |
08:32:23 <b_jonas> probably not, because it would just account to making a union with a highly aligned type, |
| 40 |
which is something I can do on any compiler portably |
| 41 |
|
| 42 |
TODO: #define ECB_IS_INTEGRAL(x) !((1 ? 1 : (x)) / 2) |
| 43 |
#define ECB_IS_INTEGRAL(x) (sizeof ((x) + 1.0f) != sizeof((x) + 1ULL)) |
| 44 |
|
| 45 |
TODO: ecb_minpot, either using bit tricks or ecb_ldXX |
| 46 |
|
| 47 |
TODO: __builtin_popcountll exists... |
| 48 |
|
| 49 |
TODO: __builtin_powi |
| 50 |
|
| 51 |
TODO: https://gustedt.wordpress.com/2010/06/08/detect-empty-macro-arguments/ |
| 52 |
|
| 53 |
implement is_constant for c11: https://gustedt.wordpress.com/2013/08/22/testing-compile-time-constness-and-null-pointers-with-c11s-_generic/ |
| 54 |
|
| 55 |
#ifdef _MSC_VER |
| 56 |
|
| 57 |
#include <stdlib.h> |
| 58 |
#define bswap_32(x) _byteswap_ulong(x) |
| 59 |
#define bswap_64(x) _byteswap_uint64(x) |
| 60 |
|
| 61 |
#elif defined(__APPLE__) |
| 62 |
|
| 63 |
// Mac OS X / Darwin features |
| 64 |
#include <libkern/OSByteOrder.h> |
| 65 |
#define bswap_32(x) OSSwapInt32(x) |
| 66 |
#define bswap_64(x) OSSwapInt64(x) |
| 67 |
|
| 68 |
#elif defined(__sun) || defined(sun) |
| 69 |
|
| 70 |
#include <sys/byteorder.h> |
| 71 |
#define bswap_32(x) BSWAP_32(x) |
| 72 |
#define bswap_64(x) BSWAP_64(x) |
| 73 |
|
| 74 |
#elif defined(__FreeBSD__) |
| 75 |
|
| 76 |
#include <sys/endian.h> |
| 77 |
#define bswap_32(x) bswap32(x) |
| 78 |
#define bswap_64(x) bswap64(x) |
| 79 |
|
| 80 |
#elif defined(__OpenBSD__) |
| 81 |
|
| 82 |
#include <sys/types.h> |
| 83 |
#define bswap_32(x) swap32(x) |
| 84 |
#define bswap_64(x) swap64(x) |
| 85 |
|
| 86 |
#elif defined(__NetBSD__) |
| 87 |
|
| 88 |
#include <sys/types.h> |
| 89 |
#include <machine/bswap.h> |
| 90 |
#if defined(__BSWAP_RENAME) && !defined(__bswap_32) |
| 91 |
#define bswap_32(x) bswap32(x) |
| 92 |
#define bswap_64(x) bswap64(x) |
| 93 |
#endif |
| 94 |
|
| 95 |
#else |
| 96 |
|
| 97 |
#include <byteswap.h> |
| 98 |
|
| 99 |
#endif |
| 100 |
|
| 101 |
TODO: generic poprcount etc., also fast_t for them? |
| 102 |
|
| 103 |
- allow any rotate count in rcb_rot*, at the expense |
| 104 |
of bad optimisation results on some platforms or with |
| 105 |
some word sizes. |
| 106 |
- try to use inttypes.h on mingw. |
| 107 |
|
| 108 |
0x00010008 |
| 109 |
- aligned/unaligned load/store, bswap, host order |
| 110 |
conversion. |
| 111 |
- generic C++ ecb_rot[lr], ecb_popcount, ecb_ctz, |
| 112 |
ecb_is_pot32 and ecb_bitrev functions. |
| 113 |
|
| 114 |
0x00010007 |
| 115 |
- new ECB_OPTIMIZE_SIZE symbol. |
| 116 |
|
| 117 |
0x00010006 |
| 118 |
- new ECB_MEMORY_FENCE_RELAXED memory fence. |
| 119 |
- use acquire/eelease memory barriers on sun workshop pro, not read/write. |
| 120 |
- rely on c++ compiler barriers to do the right thing in gcc/clang. |
| 121 |
- change release memory fence to memory barrier on ia32/ia64. |
| 122 |
- apply ctz/ld patch for msc by Zsbán Ambrus. |
| 123 |
- ECB_PTRSIZE erroneously was 8 on most 32bit systems ( |
| 124 |
found by Zsbán Ambrus). |
| 125 |
- improved compiletime detection of endianness, also, allow |
| 126 |
runtime detection to indicate other-than-big/little endianness. |
| 127 |
- no memory barrier neded on arm < 6. |
| 128 |
|
| 129 |
0x00010005 |
| 130 |
- improve ecb_binary16_to_float. |
| 131 |
- add ecb_float_to_binary16. |
| 132 |
- add ecb_binary16_to_binary32 and ecb_binary32_to_binary16 pair. |
| 133 |
|
| 134 |
0x00010001 |
| 135 |
- add ecb_is_pot32/64. |
| 136 |
- add intptr_t/uintptr_t. |
| 137 |
- add ECB_PTRSIZE. |
| 138 |
- more macros for C/C++ version checks. |
| 139 |
- support C11 atomics for memory fences. |
| 140 |
- support gcc-4.7 atomics for memory fences. |
| 141 |
- support m68k, m88k and sh (patch by Miod Vallat). |
| 142 |
- add ecb_binary16_to_float. |
| 143 |
|
| 144 |
TODO: ecb_restrict_array etc. http://ue.tst.eu/5093eafd713ec5fda776d8065070aa4c.txt |
| 145 |
TODO: ffs |
| 146 |
64 bit variants of everything |
| 147 |
TODO: examples from X for clz/ctz |
| 148 |
TODO: arithmetic right shift |
| 149 |
TODO: template/generic functions for x32/x64 and so on |
| 150 |
TODO: #define ecb_integer_multiples_of(n,d) ((char (*)[d])(n) - (char (*)[d])0) |
| 151 |
TODO: generalised shift |
| 152 |
TODO: #define ECB_FAST_UNALIGNED_ACCESS |
| 153 |
unsigned long gensh(unsigned long v, int x) { |
| 154 |
int a, b; |
| 155 |
a = (v << x) & -(((unsigned int)x) < 32); |
| 156 |
x = -x; |
| 157 |
b = (v >> x) & -(((unsigned int)x) < 32); |
| 158 |
return a|b; |
| 159 |
} |
| 160 |
|
| 161 |
TODO: export(=dllexport) & hidden |
| 162 |
TODO: flatten |
| 163 |
TODO: warning(msg) |
| 164 |
TODO: error(msg) |
| 165 |
TODO: leaf (uh), noclone (hmmm) |
| 166 |
TODO: nonnull, returns_nonnull |
| 167 |
TODO: nothrow |
| 168 |
TODO: used |
| 169 |
TODO: trap |
| 170 |
TODO: http://llvm.org/docs/doxygen/html/Compiler_8h_source.html |
| 171 |
|
| 172 |
TODO: read/write unaligned macros |
| 173 |
TODO: htonl and friends |
| 174 |
TODO: macro to convert from unsigned to signed "the natural way" |
| 175 |
TODO: ecb_static_assert, with message (just like boost), or somesuch, using array-declaration |
| 176 |
TODO: alignof |
| 177 |
|
| 178 |
|
