ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libecb/Changes
Revision: 1.42
Committed: Fri Mar 25 15:22:17 2022 UTC (2 years, 3 months ago) by root
Branch: MAIN
Changes since 1.41: +1 -0 lines
Log Message:
*** empty log message ***

File Contents

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