ViewVC Help
View File | Revision Log | Show Annotations | Download File
/cvs/libecb/Changes
Revision: 1.45
Committed: Fri Mar 25 15:34:12 2022 UTC (2 years, 2 months ago) by root
Branch: MAIN
Changes since 1.44: +2 -0 lines
Log Message:
*** empty log message ***

File Contents

# User Rev Content
1 root 1.39 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 root 1.40 - use __builtin_popcountl on 64 bit clang and gcc.
7 root 1.41 - use 64 bit builtins, not 128 bit builtins, for ctz/clz on 64 bit cpus.
8 root 1.42 - add ecb_clz* functions,
9 root 1.43 - renamed ecb_gray* functions to have trailing bit width.
10 root 1.44 - use ecb_function_ consistently for newly added functions.
11 root 1.45 - say goodbye to gnu-style function definitions - they were not
12     used consistently anyway, and imho make the code harder to read.
13 root 1.39
14 root 1.38 0x0001000b
15     - added ecb_gray*_{de,en}code functions.
16    
17 root 1.37 0x0001000a
18     - added ecb_ptrmix.
19    
20 root 1.34 0x00010009
21 root 1.35 - added ecb_i2a family of functions.
22 root 1.34 - added ECB_64BIT_NATIVE.
23    
24 root 1.15 TODO:
25     08:30:06 <b_jonas> I think it could be worth to add a macro that works like alignof or _Alignof on sane
26     compilers, and like __alignof on MS compilers that support it, see
27     http://msdn.microsoft.com/en-us/library/45t0s5f4.aspx
28     08:30:24 <b_jonas> even if you can't support it on all the old compilers
29     08:31:17 <b_jonas> I'd also like a macro for alignas, but sadly, that seems impossible in general, because
30     the MS compiler only has some half-attempt to do something similar but with different and
31     more broken semantics, see http://msdn.microsoft.com/en-us/library/83ythb65.aspx
32     08:31:35 <b_jonas> but I wonder if some special case could still be worth to support
33     08:32:23 <b_jonas> probably not, because it would just account to making a union with a highly aligned type,
34     which is something I can do on any compiler portably
35    
36     TODO: #define ECB_IS_INTEGRAL(x) !((1 ? 1 : (x)) / 2)
37     #define ECB_IS_INTEGRAL(x) (sizeof ((x) + 1.0f) != sizeof((x) + 1ULL))
38    
39 root 1.28 TODO: ecb_minpot, either using bit tricks or ecb_ldXX
40    
41     TODO: __builtin_popcountll exists...
42    
43 root 1.26 TODO: __builtin_powi
44    
45 root 1.15 TODO: https://gustedt.wordpress.com/2010/06/08/detect-empty-macro-arguments/
46    
47 root 1.13 implement is_constant for c11: https://gustedt.wordpress.com/2013/08/22/testing-compile-time-constness-and-null-pointers-with-c11s-_generic/
48 root 1.12
49     #ifdef _MSC_VER
50    
51     #include <stdlib.h>
52     #define bswap_32(x) _byteswap_ulong(x)
53     #define bswap_64(x) _byteswap_uint64(x)
54    
55     #elif defined(__APPLE__)
56    
57     // Mac OS X / Darwin features
58     #include <libkern/OSByteOrder.h>
59     #define bswap_32(x) OSSwapInt32(x)
60     #define bswap_64(x) OSSwapInt64(x)
61    
62     #elif defined(__sun) || defined(sun)
63    
64     #include <sys/byteorder.h>
65     #define bswap_32(x) BSWAP_32(x)
66     #define bswap_64(x) BSWAP_64(x)
67    
68     #elif defined(__FreeBSD__)
69    
70     #include <sys/endian.h>
71     #define bswap_32(x) bswap32(x)
72     #define bswap_64(x) bswap64(x)
73    
74     #elif defined(__OpenBSD__)
75    
76     #include <sys/types.h>
77     #define bswap_32(x) swap32(x)
78     #define bswap_64(x) swap64(x)
79    
80     #elif defined(__NetBSD__)
81    
82     #include <sys/types.h>
83     #include <machine/bswap.h>
84     #if defined(__BSWAP_RENAME) && !defined(__bswap_32)
85     #define bswap_32(x) bswap32(x)
86     #define bswap_64(x) bswap64(x)
87     #endif
88    
89     #else
90    
91     #include <byteswap.h>
92    
93     #endif
94    
95 root 1.32 TODO: generic poprcount etc., also fast_t for them?
96 root 1.33
97 root 1.36 - allow any rotate count in rcb_rot*, at the expense
98     of bad optimisation results on some platforms or with
99     some word sizes.
100 root 1.33 - try to use inttypes.h on mingw.
101    
102 root 1.30 0x00010008
103     - aligned/unaligned load/store, bswap, host order
104     conversion.
105 root 1.32 - generic C++ ecb_rot[lr], ecb_popcount, ecb_ctz,
106     ecb_is_pot32 and ecb_bitrev functions.
107 root 1.30
108 root 1.29 0x00010007
109     - new ECB_OPTIMIZE_SIZE symbol.
110    
111 root 1.28 0x00010006
112 root 1.29 - new ECB_MEMORY_FENCE_RELAXED memory fence.
113 root 1.28 - use acquire/eelease memory barriers on sun workshop pro, not read/write.
114     - rely on c++ compiler barriers to do the right thing in gcc/clang.
115 root 1.27 - change release memory fence to memory barrier on ia32/ia64.
116 root 1.23 - apply ctz/ld patch for msc by Zsbán Ambrus.
117 root 1.24 - ECB_PTRSIZE erroneously was 8 on most 32bit systems (
118     found by Zsbán Ambrus).
119 root 1.25 - improved compiletime detection of endianness, also, allow
120     runtime detection to indicate other-than-big/little endianness.
121 root 1.26 - no memory barrier neded on arm < 6.
122 root 1.23
123 root 1.22 0x00010005
124     - improve ecb_binary16_to_float.
125     - add ecb_float_to_binary16.
126     - add ecb_binary16_to_binary32 and ecb_binary32_to_binary16 pair.
127    
128 root 1.6 0x00010001
129     - add ecb_is_pot32/64.
130     - add intptr_t/uintptr_t.
131     - add ECB_PTRSIZE.
132 root 1.7 - more macros for C/C++ version checks.
133     - support C11 atomics for memory fences.
134     - support gcc-4.7 atomics for memory fences.
135 root 1.10 - support m68k, m88k and sh (patch by Miod Vallat).
136 root 1.11 - add ecb_binary16_to_float.
137 root 1.6
138 root 1.9 TODO: ecb_restrict_array etc. http://ue.tst.eu/5093eafd713ec5fda776d8065070aa4c.txt
139 sf-exg 1.20 TODO: ffs
140 root 1.1 64 bit variants of everything
141 root 1.2 TODO: examples from X for clz/ctz
142 root 1.3 TODO: arithmetic right shift
143     TODO: template/generic functions for x32/x64 and so on
144 root 1.4 TODO: #define ecb_integer_multiples_of(n,d) ((char (*)[d])(n) - (char (*)[d])0)
145 root 1.3 TODO: generalised shift
146 root 1.8 TODO: #define ECB_FAST_UNALIGNED_ACCESS
147 root 1.3 unsigned long gensh(unsigned long v, int x) {
148     int a, b;
149     a = (v << x) & -(((unsigned int)x) < 32);
150     x = -x;
151     b = (v >> x) & -(((unsigned int)x) < 32);
152     return a|b;
153     }
154 root 1.14
155 root 1.16 TODO: export(=dllexport) & hidden
156     TODO: flatten
157     TODO: warning(msg)
158     TODO: error(msg)
159 root 1.17 TODO: leaf (uh), noclone (hmmm)
160     TODO: nonnull, returns_nonnull
161     TODO: nothrow
162     TODO: used
163     TODO: trap
164     TODO: http://llvm.org/docs/doxygen/html/Compiler_8h_source.html
165 root 1.16
166 root 1.14 TODO: read/write unaligned macros
167     TODO: htonl and friends
168 sf-exg 1.20 TODO: macro to convert from unsigned to signed "the natural way"
169 root 1.14 TODO: ecb_static_assert, with message (just like boost), or somesuch, using array-declaration
170     TODO: alignof
171    
172