mathematical 0.5.0 → 0.6.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/README.md +4 -4
- data/Rakefile +2 -2
- data/ext/mathematical/extconf.rb +2 -0
- data/ext/mathematical/mtex2MML/src/color_definitions.c +45 -0
- data/ext/mathematical/mtex2MML/src/deps/uthash/utarray.h +232 -0
- data/ext/mathematical/mtex2MML/src/deps/uthash/uthash.h +958 -0
- data/ext/mathematical/mtex2MML/src/deps/uthash/utlist.h +757 -0
- data/ext/mathematical/mtex2MML/src/deps/uthash/utstring.h +393 -0
- data/ext/mathematical/mtex2MML/src/lex.yy.c +4642 -3467
- data/ext/mathematical/mtex2MML/src/mtex2MML.h +31 -31
- data/ext/mathematical/mtex2MML/src/mtex2MML_ruby.c +2299 -0
- data/ext/mathematical/mtex2MML/src/parse_extras.c +223 -224
- data/ext/mathematical/mtex2MML/src/parse_extras.h +59 -58
- data/ext/mathematical/mtex2MML/src/string_extras.c +17 -194
- data/ext/mathematical/mtex2MML/src/string_extras.h +10 -28
- data/ext/mathematical/mtex2MML/src/y.tab.c +3622 -3063
- data/ext/mathematical/mtex2MML/src/y.tab.h +440 -336
- data/lib/mathematical/corrections.rb +0 -18
- data/lib/mathematical/version.rb +1 -1
- data/test/mathematical/corrections_test.rb +0 -40
- data/test/mathematical/fixtures/performance/big_file.text +8 -20
- data/test/mathematical/maliciousness_test.rb +12 -2
- data/test/mathematical/mathjax_test.rb +16 -0
- metadata +10 -4
- data/ext/mathematical/mtex2MML/src/stack.c +0 -80
- data/ext/mathematical/mtex2MML/src/stack.h +0 -101
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: a5c2b7f088d174aa989dd112d75604446c2a88ce
|
4
|
+
data.tar.gz: eba8f5b307969ca88aef92ce6def3c93deea3cff
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 6749ee0562805b3314d4f26e78e1ae28d4ee340bc332f84e986619379f867edfa3023660fd4243578003701fee34b77bdc46a2b014c4de78b0a098c1cf9a50ee
|
7
|
+
data.tar.gz: 1e215792250caab54bc1837ccef1bf766bdd7b7daf4692663f3f1e250dda3caaf3f6b8f9fb280633db03100e820fd3c39ece96cfca068563def4da53604e504d
|
data/README.md
CHANGED
@@ -30,7 +30,7 @@ require 'mathematical'
|
|
30
30
|
Mathematical::Render.new.render(string_with_math)
|
31
31
|
```
|
32
32
|
|
33
|
-
`string_with_math` should just be a string of
|
33
|
+
`string_with_math` should just be a string of math TeX inline (`$..$`) or display (`$$..$$`) style math.
|
34
34
|
|
35
35
|
The output will be a hash, with keys that depend on the format you want:
|
36
36
|
|
@@ -65,9 +65,9 @@ renderer.render('$a \ne b$')
|
|
65
65
|
|
66
66
|
### Supported commands and symbols
|
67
67
|
|
68
|
-
|
68
|
+
Check out [SUPPORTED.md on the mtex2MML website](https://github.com/gjtorikian/mtex2MML/blob/master/SUPPORTED.md).
|
69
69
|
|
70
|
-
**Note**: This library makes few assumptions about the strings that you pass in. It assumes that `$..$` is inline math, `$$..$$` is display math, and that slashes like `\\` are escaped (`\\\\`).
|
70
|
+
**Note**: This library makes a few assumptions about the strings that you pass in. It assumes that `$..$` is inline math, `$$..$$` is display math, and that your double-slashes (like `\\`) are escaped (`\\\\`).
|
71
71
|
|
72
72
|
## Dependencies
|
73
73
|
|
@@ -211,7 +211,7 @@ And thus a wrapper was born.
|
|
211
211
|
## More math stuff
|
212
212
|
|
213
213
|
Check out [math-to-itex](https://github.com/gjtorikian/math-to-itex/), which quickly
|
214
|
-
parses out
|
214
|
+
parses out TeX notation from strings.
|
215
215
|
|
216
216
|
With it, you could do something fun like:
|
217
217
|
|
data/Rakefile
CHANGED
@@ -42,8 +42,8 @@ end
|
|
42
42
|
task :destroy_copies do
|
43
43
|
ext_dir = File.join(File.dirname(__FILE__), "ext", "mathematical")
|
44
44
|
Dir.glob("#{ext_dir}/*").select { |f| File.file?(f) }.each do |f|
|
45
|
-
next if f =~ /extconf.rb
|
45
|
+
next if f =~ /extconf.rb$/ || f =~ /mathematical.c$/
|
46
46
|
File.delete(f)
|
47
47
|
end
|
48
|
-
Dir.glob("#{ext_dir}/{lib,src,test,ext}").select { |d| FileUtils.rm_rf d }
|
48
|
+
Dir.glob("#{ext_dir}/{lib,src,test,ext,deps,uthash}").select { |d| FileUtils.rm_rf d }
|
49
49
|
end
|
data/ext/mathematical/extconf.rb
CHANGED
@@ -26,6 +26,8 @@ Dir.chdir(MTEX_DIR) do
|
|
26
26
|
end
|
27
27
|
|
28
28
|
FileUtils.cp_r(Dir.glob("#{MTEX_DIR}/*.{c,h,cc}"), File.dirname(__FILE__))
|
29
|
+
FileUtils.cp_r("#{MTEX_DIR}/deps/.", File.dirname(__FILE__))
|
30
|
+
FileUtils.cp_r(Dir.glob("#{MTEX_DIR}/deps/**/*"), File.dirname(__FILE__))
|
29
31
|
|
30
32
|
have_library("pangocairo-1.0")
|
31
33
|
|
@@ -0,0 +1,45 @@
|
|
1
|
+
#include "parse_extras.h"
|
2
|
+
#include "deps/uthash/uthash.h"
|
3
|
+
|
4
|
+
void create_css_colors(struct css_colors **colors)
|
5
|
+
{
|
6
|
+
const char **n, **h,
|
7
|
+
*names[] = { "\"Apricot\"","\"Aquamarine\"","\"Bittersweet\"",
|
8
|
+
"\"Black\"","\"Blue\"","\"BlueGreen\"","\"BlueViolet\"","\"BrickRed\"",
|
9
|
+
"\"Brown\"","\"BurntOrange\"","\"CadetBlue\"","\"CarnationPink\"","\"Cerulean\"",
|
10
|
+
"\"CornflowerBlue\"","\"Cyan\"","\"Dandelion\"","\"DarkOrchid\"","\"Emerald\"",
|
11
|
+
"\"ForestGreen\"","\"Fuchsia\"","\"Goldenrod\"","\"Gray\"","\"Green\"",
|
12
|
+
"\"GreenYellow\"","\"JungleGreen\"","\"Lavender\"","\"LimeGreen\"","\"Magenta\"",
|
13
|
+
"\"Mahogany\"","\"Maroon\"","\"Melon\"","\"MidnightBlue\"","\"Mulberry\"",
|
14
|
+
"\"NavyBlue\"","\"OliveGreen\"","\"Orange\"","\"OrangeRed\"","\"Orchid\"",
|
15
|
+
"\"Peach\"","\"Periwinkle\"","\"PineGreen\"","\"Plum\"","\"ProcessBlue\"",
|
16
|
+
"\"Purple\"","\"RawSienna\"","\"Red\"","\"RedOrange\"","\"RedViolet\"",
|
17
|
+
"\"Rhodamine\"","\"RoyalBlue\"","\"RoyalPurple\"","\"RubineRed\"","\"Salmon\"",
|
18
|
+
"\"SeaGreen\"","\"Sepia\"","\"SkyBlue\"","\"SpringGreen\"","\"Tan\"","\"TealBlue\"",
|
19
|
+
"\"Thistle\"","\"Turquoise\"","\"Violet\"","\"VioletRed\"","\"White\"",
|
20
|
+
"\"WildStrawberry\"","\"Yellow\"","\"YellowGreen\"", "\"YellowOrange\"", NULL
|
21
|
+
},
|
22
|
+
|
23
|
+
*hex_code[] = { "\"#FBB982\"","\"#00B5BE\"","\"#C04F17\"","\"#221E1F\"","\"#2D2F92\"",
|
24
|
+
"\"#00B3B8\"","\"#473992\"","\"#B6321C\"","\"#792500\"","\"#F7921D\"","\"#74729A\"",
|
25
|
+
"\"#F282B4\"","\"#00A2E3\"","\"#41B0E4\"","\"#00AEEF\"","\"#FDBC42\"","\"#A4538A\"",
|
26
|
+
"\"#00A99D\"","\"#009B55\"","\"#8C368C\"","\"#FFDF42\"","\"#949698\"","\"#00A64F\"",
|
27
|
+
"\"#DFE674\"","\"#00A99A\"","\"#F49EC4\"","\"#8DC73E\"","\"#EC008C\"","\"#A9341F\"",
|
28
|
+
"\"#AF3235\"","\"#F89E7B\"","\"#006795\"","\"#A93C93\"","\"#006EB8\"","\"#3C8031\"",
|
29
|
+
"\"#F58137\"","\"#ED135A\"","\"#AF72B0\"","\"#F7965A\"","\"#7977B8\"","\"#008B72\"",
|
30
|
+
"\"#92268F\"","\"#00B0F0\"","\"#99479B\"","\"#974006\"","\"#ED1B23\"","\"#F26035\"",
|
31
|
+
"\"#A1246B\"","\"#EF559F\"","\"#0071BC\"","\"#613F99\"","\"#ED017D\"","\"#F69289\"",
|
32
|
+
"\"#3FBC9D\"","\"#671800\"","\"#46C5DD\"","\"#C6DC67\"","\"#DA9D76\"","\"#00AEB3\"",
|
33
|
+
"\"#D883B7\"","\"#00B4CE\"","\"#58429B\"","\"#EF58A0\"","\"#FFFFFF\"","\"#EE2967\"",
|
34
|
+
"\"#FFF200\"","\"#98CC70\"","\"#FAA21A\"", NULL
|
35
|
+
};
|
36
|
+
|
37
|
+
struct css_colors *c = NULL;
|
38
|
+
|
39
|
+
for (n = names, h = hex_code; *n != NULL; n++, h++) {
|
40
|
+
c = malloc(sizeof(struct css_colors));
|
41
|
+
strncpy(c->name, *n, 22);
|
42
|
+
strncpy(c->color, *h, 10);
|
43
|
+
HASH_ADD_STR(*colors, name, c );
|
44
|
+
}
|
45
|
+
}
|
@@ -0,0 +1,232 @@
|
|
1
|
+
/*
|
2
|
+
Copyright (c) 2008-2014, Troy D. Hanson http://troydhanson.github.com/uthash/
|
3
|
+
All rights reserved.
|
4
|
+
|
5
|
+
Redistribution and use in source and binary forms, with or without
|
6
|
+
modification, are permitted provided that the following conditions are met:
|
7
|
+
|
8
|
+
* Redistributions of source code must retain the above copyright
|
9
|
+
notice, this list of conditions and the following disclaimer.
|
10
|
+
|
11
|
+
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS
|
12
|
+
IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
|
13
|
+
TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
|
14
|
+
PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
|
15
|
+
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
|
16
|
+
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
|
17
|
+
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
|
18
|
+
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
|
19
|
+
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
|
20
|
+
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
|
21
|
+
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
22
|
+
*/
|
23
|
+
|
24
|
+
/* a dynamic array implementation using macros
|
25
|
+
*/
|
26
|
+
#ifndef UTARRAY_H
|
27
|
+
#define UTARRAY_H
|
28
|
+
|
29
|
+
#define UTARRAY_VERSION 1.9.9
|
30
|
+
|
31
|
+
#ifdef __GNUC__
|
32
|
+
#define _UNUSED_ __attribute__ ((__unused__))
|
33
|
+
#else
|
34
|
+
#define _UNUSED_
|
35
|
+
#endif
|
36
|
+
|
37
|
+
#include <stddef.h> /* size_t */
|
38
|
+
#include <string.h> /* memset, etc */
|
39
|
+
#include <stdlib.h> /* exit */
|
40
|
+
|
41
|
+
#define oom() exit(-1)
|
42
|
+
|
43
|
+
typedef void (ctor_f)(void *dst, const void *src);
|
44
|
+
typedef void (dtor_f)(void *elt);
|
45
|
+
typedef void (init_f)(void *elt);
|
46
|
+
typedef struct {
|
47
|
+
size_t sz;
|
48
|
+
init_f *init;
|
49
|
+
ctor_f *copy;
|
50
|
+
dtor_f *dtor;
|
51
|
+
} UT_icd;
|
52
|
+
|
53
|
+
typedef struct {
|
54
|
+
unsigned i,n;/* i: index of next available slot, n: num slots */
|
55
|
+
UT_icd icd; /* initializer, copy and destructor functions */
|
56
|
+
char *d; /* n slots of size icd->sz*/
|
57
|
+
} UT_array;
|
58
|
+
|
59
|
+
#define utarray_init(a,_icd) do { \
|
60
|
+
memset(a,0,sizeof(UT_array)); \
|
61
|
+
(a)->icd=*_icd; \
|
62
|
+
} while(0)
|
63
|
+
|
64
|
+
#define utarray_done(a) do { \
|
65
|
+
if ((a)->n) { \
|
66
|
+
if ((a)->icd.dtor) { \
|
67
|
+
size_t _ut_i; \
|
68
|
+
for(_ut_i=0; _ut_i < (a)->i; _ut_i++) { \
|
69
|
+
(a)->icd.dtor(utarray_eltptr(a,_ut_i)); \
|
70
|
+
} \
|
71
|
+
} \
|
72
|
+
free((a)->d); \
|
73
|
+
} \
|
74
|
+
(a)->n=0; \
|
75
|
+
} while(0)
|
76
|
+
|
77
|
+
#define utarray_new(a,_icd) do { \
|
78
|
+
a=(UT_array*)malloc(sizeof(UT_array)); \
|
79
|
+
utarray_init(a,_icd); \
|
80
|
+
} while(0)
|
81
|
+
|
82
|
+
#define utarray_free(a) do { \
|
83
|
+
utarray_done(a); \
|
84
|
+
free(a); \
|
85
|
+
} while(0)
|
86
|
+
|
87
|
+
#define utarray_reserve(a,by) do { \
|
88
|
+
if (((a)->i+by) > ((a)->n)) { \
|
89
|
+
while(((a)->i+by) > ((a)->n)) { (a)->n = ((a)->n ? (2*(a)->n) : 8); } \
|
90
|
+
if ( ((a)->d=(char*)realloc((a)->d, (a)->n*(a)->icd.sz)) == NULL) oom(); \
|
91
|
+
} \
|
92
|
+
} while(0)
|
93
|
+
|
94
|
+
#define utarray_push_back(a,p) do { \
|
95
|
+
utarray_reserve(a,1); \
|
96
|
+
if ((a)->icd.copy) { (a)->icd.copy( _utarray_eltptr(a,(a)->i++), p); } \
|
97
|
+
else { memcpy(_utarray_eltptr(a,(a)->i++), p, (a)->icd.sz); }; \
|
98
|
+
} while(0)
|
99
|
+
|
100
|
+
#define utarray_pop_back(a) do { \
|
101
|
+
if ((a)->icd.dtor) { (a)->icd.dtor( _utarray_eltptr(a,--((a)->i))); } \
|
102
|
+
else { (a)->i--; } \
|
103
|
+
} while(0)
|
104
|
+
|
105
|
+
#define utarray_extend_back(a) do { \
|
106
|
+
utarray_reserve(a,1); \
|
107
|
+
if ((a)->icd.init) { (a)->icd.init(_utarray_eltptr(a,(a)->i)); } \
|
108
|
+
else { memset(_utarray_eltptr(a,(a)->i),0,(a)->icd.sz); } \
|
109
|
+
(a)->i++; \
|
110
|
+
} while(0)
|
111
|
+
|
112
|
+
#define utarray_len(a) ((a)->i)
|
113
|
+
|
114
|
+
#define utarray_eltptr(a,j) (((j) < (a)->i) ? _utarray_eltptr(a,j) : NULL)
|
115
|
+
#define _utarray_eltptr(a,j) ((char*)((a)->d + ((a)->icd.sz*(j) )))
|
116
|
+
|
117
|
+
#define utarray_insert(a,p,j) do { \
|
118
|
+
if (j > (a)->i) utarray_resize(a,j); \
|
119
|
+
utarray_reserve(a,1); \
|
120
|
+
if ((j) < (a)->i) { \
|
121
|
+
memmove( _utarray_eltptr(a,(j)+1), _utarray_eltptr(a,j), \
|
122
|
+
((a)->i - (j))*((a)->icd.sz)); \
|
123
|
+
} \
|
124
|
+
if ((a)->icd.copy) { (a)->icd.copy( _utarray_eltptr(a,j), p); } \
|
125
|
+
else { memcpy(_utarray_eltptr(a,j), p, (a)->icd.sz); }; \
|
126
|
+
(a)->i++; \
|
127
|
+
} while(0)
|
128
|
+
|
129
|
+
#define utarray_inserta(a,w,j) do { \
|
130
|
+
if (utarray_len(w) == 0) break; \
|
131
|
+
if (j > (a)->i) utarray_resize(a,j); \
|
132
|
+
utarray_reserve(a,utarray_len(w)); \
|
133
|
+
if ((j) < (a)->i) { \
|
134
|
+
memmove(_utarray_eltptr(a,(j)+utarray_len(w)), \
|
135
|
+
_utarray_eltptr(a,j), \
|
136
|
+
((a)->i - (j))*((a)->icd.sz)); \
|
137
|
+
} \
|
138
|
+
if ((a)->icd.copy) { \
|
139
|
+
size_t _ut_i; \
|
140
|
+
for(_ut_i=0;_ut_i<(w)->i;_ut_i++) { \
|
141
|
+
(a)->icd.copy(_utarray_eltptr(a,j+_ut_i), _utarray_eltptr(w,_ut_i)); \
|
142
|
+
} \
|
143
|
+
} else { \
|
144
|
+
memcpy(_utarray_eltptr(a,j), _utarray_eltptr(w,0), \
|
145
|
+
utarray_len(w)*((a)->icd.sz)); \
|
146
|
+
} \
|
147
|
+
(a)->i += utarray_len(w); \
|
148
|
+
} while(0)
|
149
|
+
|
150
|
+
#define utarray_resize(dst,num) do { \
|
151
|
+
size_t _ut_i; \
|
152
|
+
if (dst->i > (size_t)(num)) { \
|
153
|
+
if ((dst)->icd.dtor) { \
|
154
|
+
for(_ut_i=num; _ut_i < dst->i; _ut_i++) { \
|
155
|
+
(dst)->icd.dtor(utarray_eltptr(dst,_ut_i)); \
|
156
|
+
} \
|
157
|
+
} \
|
158
|
+
} else if (dst->i < (size_t)(num)) { \
|
159
|
+
utarray_reserve(dst,num-dst->i); \
|
160
|
+
if ((dst)->icd.init) { \
|
161
|
+
for(_ut_i=dst->i; _ut_i < num; _ut_i++) { \
|
162
|
+
(dst)->icd.init(utarray_eltptr(dst,_ut_i)); \
|
163
|
+
} \
|
164
|
+
} else { \
|
165
|
+
memset(_utarray_eltptr(dst,dst->i),0,(dst)->icd.sz*(num-dst->i)); \
|
166
|
+
} \
|
167
|
+
} \
|
168
|
+
dst->i = num; \
|
169
|
+
} while(0)
|
170
|
+
|
171
|
+
#define utarray_concat(dst,src) do { \
|
172
|
+
utarray_inserta((dst),(src),utarray_len(dst)); \
|
173
|
+
} while(0)
|
174
|
+
|
175
|
+
#define utarray_erase(a,pos,len) do { \
|
176
|
+
if ((a)->icd.dtor) { \
|
177
|
+
size_t _ut_i; \
|
178
|
+
for(_ut_i=0; _ut_i < len; _ut_i++) { \
|
179
|
+
(a)->icd.dtor(utarray_eltptr((a),pos+_ut_i)); \
|
180
|
+
} \
|
181
|
+
} \
|
182
|
+
if ((a)->i > (pos+len)) { \
|
183
|
+
memmove( _utarray_eltptr((a),pos), _utarray_eltptr((a),pos+len), \
|
184
|
+
(((a)->i)-(pos+len))*((a)->icd.sz)); \
|
185
|
+
} \
|
186
|
+
(a)->i -= (len); \
|
187
|
+
} while(0)
|
188
|
+
|
189
|
+
#define utarray_renew(a,u) do { \
|
190
|
+
if (a) utarray_clear(a); \
|
191
|
+
else utarray_new((a),(u)); \
|
192
|
+
} while(0)
|
193
|
+
|
194
|
+
#define utarray_clear(a) do { \
|
195
|
+
if ((a)->i > 0) { \
|
196
|
+
if ((a)->icd.dtor) { \
|
197
|
+
size_t _ut_i; \
|
198
|
+
for(_ut_i=0; _ut_i < (a)->i; _ut_i++) { \
|
199
|
+
(a)->icd.dtor(utarray_eltptr(a,_ut_i)); \
|
200
|
+
} \
|
201
|
+
} \
|
202
|
+
(a)->i = 0; \
|
203
|
+
} \
|
204
|
+
} while(0)
|
205
|
+
|
206
|
+
#define utarray_sort(a,cmp) do { \
|
207
|
+
qsort((a)->d, (a)->i, (a)->icd.sz, cmp); \
|
208
|
+
} while(0)
|
209
|
+
|
210
|
+
#define utarray_find(a,v,cmp) bsearch((v),(a)->d,(a)->i,(a)->icd.sz,cmp)
|
211
|
+
|
212
|
+
#define utarray_front(a) (((a)->i) ? (_utarray_eltptr(a,0)) : NULL)
|
213
|
+
#define utarray_next(a,e) (((e)==NULL) ? utarray_front(a) : ((((a)->i) > (utarray_eltidx(a,e)+1)) ? _utarray_eltptr(a,utarray_eltidx(a,e)+1) : NULL))
|
214
|
+
#define utarray_prev(a,e) (((e)==NULL) ? utarray_back(a) : ((utarray_eltidx(a,e) > 0) ? _utarray_eltptr(a,utarray_eltidx(a,e)-1) : NULL))
|
215
|
+
#define utarray_back(a) (((a)->i) ? (_utarray_eltptr(a,(a)->i-1)) : NULL)
|
216
|
+
#define utarray_eltidx(a,e) (((char*)(e) >= (char*)((a)->d)) ? (((char*)(e) - (char*)((a)->d))/(size_t)(a)->icd.sz) : -1)
|
217
|
+
|
218
|
+
/* last we pre-define a few icd for common utarrays of ints and strings */
|
219
|
+
static void utarray_str_cpy(void *dst, const void *src) {
|
220
|
+
char **_src = (char**)src, **_dst = (char**)dst;
|
221
|
+
*_dst = (*_src == NULL) ? NULL : strdup(*_src);
|
222
|
+
}
|
223
|
+
static void utarray_str_dtor(void *elt) {
|
224
|
+
char **eltc = (char**)elt;
|
225
|
+
if (*eltc) free(*eltc);
|
226
|
+
}
|
227
|
+
static const UT_icd ut_str_icd _UNUSED_ = {sizeof(char*),NULL,utarray_str_cpy,utarray_str_dtor};
|
228
|
+
static const UT_icd ut_int_icd _UNUSED_ = {sizeof(int),NULL,NULL,NULL};
|
229
|
+
static const UT_icd ut_ptr_icd _UNUSED_ = {sizeof(void*),NULL,NULL,NULL};
|
230
|
+
|
231
|
+
|
232
|
+
#endif /* UTARRAY_H */
|
@@ -0,0 +1,958 @@
|
|
1
|
+
/*
|
2
|
+
Copyright (c) 2003-2014, Troy D. Hanson http://troydhanson.github.com/uthash/
|
3
|
+
All rights reserved.
|
4
|
+
|
5
|
+
Redistribution and use in source and binary forms, with or without
|
6
|
+
modification, are permitted provided that the following conditions are met:
|
7
|
+
|
8
|
+
* Redistributions of source code must retain the above copyright
|
9
|
+
notice, this list of conditions and the following disclaimer.
|
10
|
+
|
11
|
+
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS
|
12
|
+
IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
|
13
|
+
TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
|
14
|
+
PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
|
15
|
+
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
|
16
|
+
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
|
17
|
+
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
|
18
|
+
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
|
19
|
+
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
|
20
|
+
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
|
21
|
+
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
22
|
+
*/
|
23
|
+
|
24
|
+
#ifndef UTHASH_H
|
25
|
+
#define UTHASH_H
|
26
|
+
|
27
|
+
#include <string.h> /* memcmp,strlen */
|
28
|
+
#include <stddef.h> /* ptrdiff_t */
|
29
|
+
#include <stdlib.h> /* exit() */
|
30
|
+
|
31
|
+
/* These macros use decltype or the earlier __typeof GNU extension.
|
32
|
+
As decltype is only available in newer compilers (VS2010 or gcc 4.3+
|
33
|
+
when compiling c++ source) this code uses whatever method is needed
|
34
|
+
or, for VS2008 where neither is available, uses casting workarounds. */
|
35
|
+
#if defined(_MSC_VER) /* MS compiler */
|
36
|
+
#if _MSC_VER >= 1600 && defined(__cplusplus) /* VS2010 or newer in C++ mode */
|
37
|
+
#define DECLTYPE(x) (decltype(x))
|
38
|
+
#else /* VS2008 or older (or VS2010 in C mode) */
|
39
|
+
#define NO_DECLTYPE
|
40
|
+
#define DECLTYPE(x)
|
41
|
+
#endif
|
42
|
+
#elif defined(__BORLANDC__) || defined(__LCC__) || defined(__WATCOMC__)
|
43
|
+
#define NO_DECLTYPE
|
44
|
+
#define DECLTYPE(x)
|
45
|
+
#else /* GNU, Sun and other compilers */
|
46
|
+
#define DECLTYPE(x) (__typeof(x))
|
47
|
+
#endif
|
48
|
+
|
49
|
+
#ifdef NO_DECLTYPE
|
50
|
+
#define DECLTYPE_ASSIGN(dst,src) \
|
51
|
+
do { \
|
52
|
+
char **_da_dst = (char**)(&(dst)); \
|
53
|
+
*_da_dst = (char*)(src); \
|
54
|
+
} while(0)
|
55
|
+
#else
|
56
|
+
#define DECLTYPE_ASSIGN(dst,src) \
|
57
|
+
do { \
|
58
|
+
(dst) = DECLTYPE(dst)(src); \
|
59
|
+
} while(0)
|
60
|
+
#endif
|
61
|
+
|
62
|
+
/* a number of the hash function use uint32_t which isn't defined on Pre VS2010 */
|
63
|
+
#if defined (_WIN32)
|
64
|
+
#if defined(_MSC_VER) && _MSC_VER >= 1600
|
65
|
+
#include <stdint.h>
|
66
|
+
#elif defined(__WATCOMC__)
|
67
|
+
#include <stdint.h>
|
68
|
+
#else
|
69
|
+
typedef unsigned int uint32_t;
|
70
|
+
typedef unsigned char uint8_t;
|
71
|
+
#endif
|
72
|
+
#else
|
73
|
+
#include <stdint.h>
|
74
|
+
#endif
|
75
|
+
|
76
|
+
#define UTHASH_VERSION 1.9.9
|
77
|
+
|
78
|
+
#ifndef uthash_fatal
|
79
|
+
#define uthash_fatal(msg) exit(-1) /* fatal error (out of memory,etc) */
|
80
|
+
#endif
|
81
|
+
#ifndef uthash_malloc
|
82
|
+
#define uthash_malloc(sz) malloc(sz) /* malloc fcn */
|
83
|
+
#endif
|
84
|
+
#ifndef uthash_free
|
85
|
+
#define uthash_free(ptr,sz) free(ptr) /* free fcn */
|
86
|
+
#endif
|
87
|
+
|
88
|
+
#ifndef uthash_noexpand_fyi
|
89
|
+
#define uthash_noexpand_fyi(tbl) /* can be defined to log noexpand */
|
90
|
+
#endif
|
91
|
+
#ifndef uthash_expand_fyi
|
92
|
+
#define uthash_expand_fyi(tbl) /* can be defined to log expands */
|
93
|
+
#endif
|
94
|
+
|
95
|
+
/* initial number of buckets */
|
96
|
+
#define HASH_INITIAL_NUM_BUCKETS 32 /* initial number of buckets */
|
97
|
+
#define HASH_INITIAL_NUM_BUCKETS_LOG2 5 /* lg2 of initial number of buckets */
|
98
|
+
#define HASH_BKT_CAPACITY_THRESH 10 /* expand when bucket count reaches */
|
99
|
+
|
100
|
+
/* calculate the element whose hash handle address is hhe */
|
101
|
+
#define ELMT_FROM_HH(tbl,hhp) ((void*)(((char*)(hhp)) - ((tbl)->hho)))
|
102
|
+
|
103
|
+
#define HASH_FIND(hh,head,keyptr,keylen,out) \
|
104
|
+
do { \
|
105
|
+
unsigned _hf_bkt,_hf_hashv; \
|
106
|
+
out=NULL; \
|
107
|
+
if (head) { \
|
108
|
+
HASH_FCN(keyptr,keylen, (head)->hh.tbl->num_buckets, _hf_hashv, _hf_bkt); \
|
109
|
+
if (HASH_BLOOM_TEST((head)->hh.tbl, _hf_hashv)) { \
|
110
|
+
HASH_FIND_IN_BKT((head)->hh.tbl, hh, (head)->hh.tbl->buckets[ _hf_bkt ], \
|
111
|
+
keyptr,keylen,out); \
|
112
|
+
} \
|
113
|
+
} \
|
114
|
+
} while (0)
|
115
|
+
|
116
|
+
#ifdef HASH_BLOOM
|
117
|
+
#define HASH_BLOOM_BITLEN (1ULL << HASH_BLOOM)
|
118
|
+
#define HASH_BLOOM_BYTELEN (HASH_BLOOM_BITLEN/8) + ((HASH_BLOOM_BITLEN%8) ? 1:0)
|
119
|
+
#define HASH_BLOOM_MAKE(tbl) \
|
120
|
+
do { \
|
121
|
+
(tbl)->bloom_nbits = HASH_BLOOM; \
|
122
|
+
(tbl)->bloom_bv = (uint8_t*)uthash_malloc(HASH_BLOOM_BYTELEN); \
|
123
|
+
if (!((tbl)->bloom_bv)) { uthash_fatal( "out of memory"); } \
|
124
|
+
memset((tbl)->bloom_bv, 0, HASH_BLOOM_BYTELEN); \
|
125
|
+
(tbl)->bloom_sig = HASH_BLOOM_SIGNATURE; \
|
126
|
+
} while (0)
|
127
|
+
|
128
|
+
#define HASH_BLOOM_FREE(tbl) \
|
129
|
+
do { \
|
130
|
+
uthash_free((tbl)->bloom_bv, HASH_BLOOM_BYTELEN); \
|
131
|
+
} while (0)
|
132
|
+
|
133
|
+
#define HASH_BLOOM_BITSET(bv,idx) (bv[(idx)/8] |= (1U << ((idx)%8)))
|
134
|
+
#define HASH_BLOOM_BITTEST(bv,idx) (bv[(idx)/8] & (1U << ((idx)%8)))
|
135
|
+
|
136
|
+
#define HASH_BLOOM_ADD(tbl,hashv) \
|
137
|
+
HASH_BLOOM_BITSET((tbl)->bloom_bv, (hashv & (uint32_t)((1ULL << (tbl)->bloom_nbits) - 1)))
|
138
|
+
|
139
|
+
#define HASH_BLOOM_TEST(tbl,hashv) \
|
140
|
+
HASH_BLOOM_BITTEST((tbl)->bloom_bv, (hashv & (uint32_t)((1ULL << (tbl)->bloom_nbits) - 1)))
|
141
|
+
|
142
|
+
#else
|
143
|
+
#define HASH_BLOOM_MAKE(tbl)
|
144
|
+
#define HASH_BLOOM_FREE(tbl)
|
145
|
+
#define HASH_BLOOM_ADD(tbl,hashv)
|
146
|
+
#define HASH_BLOOM_TEST(tbl,hashv) (1)
|
147
|
+
#define HASH_BLOOM_BYTELEN 0
|
148
|
+
#endif
|
149
|
+
|
150
|
+
#define HASH_MAKE_TABLE(hh,head) \
|
151
|
+
do { \
|
152
|
+
(head)->hh.tbl = (UT_hash_table*)uthash_malloc( \
|
153
|
+
sizeof(UT_hash_table)); \
|
154
|
+
if (!((head)->hh.tbl)) { uthash_fatal( "out of memory"); } \
|
155
|
+
memset((head)->hh.tbl, 0, sizeof(UT_hash_table)); \
|
156
|
+
(head)->hh.tbl->tail = &((head)->hh); \
|
157
|
+
(head)->hh.tbl->num_buckets = HASH_INITIAL_NUM_BUCKETS; \
|
158
|
+
(head)->hh.tbl->log2_num_buckets = HASH_INITIAL_NUM_BUCKETS_LOG2; \
|
159
|
+
(head)->hh.tbl->hho = (char*)(&(head)->hh) - (char*)(head); \
|
160
|
+
(head)->hh.tbl->buckets = (UT_hash_bucket*)uthash_malloc( \
|
161
|
+
HASH_INITIAL_NUM_BUCKETS*sizeof(struct UT_hash_bucket)); \
|
162
|
+
if (! (head)->hh.tbl->buckets) { uthash_fatal( "out of memory"); } \
|
163
|
+
memset((head)->hh.tbl->buckets, 0, \
|
164
|
+
HASH_INITIAL_NUM_BUCKETS*sizeof(struct UT_hash_bucket)); \
|
165
|
+
HASH_BLOOM_MAKE((head)->hh.tbl); \
|
166
|
+
(head)->hh.tbl->signature = HASH_SIGNATURE; \
|
167
|
+
} while(0)
|
168
|
+
|
169
|
+
#define HASH_ADD(hh,head,fieldname,keylen_in,add) \
|
170
|
+
HASH_ADD_KEYPTR(hh,head,&((add)->fieldname),keylen_in,add)
|
171
|
+
|
172
|
+
#define HASH_REPLACE(hh,head,fieldname,keylen_in,add,replaced) \
|
173
|
+
do { \
|
174
|
+
replaced=NULL; \
|
175
|
+
HASH_FIND(hh,head,&((add)->fieldname),keylen_in,replaced); \
|
176
|
+
if (replaced!=NULL) { \
|
177
|
+
HASH_DELETE(hh,head,replaced); \
|
178
|
+
}; \
|
179
|
+
HASH_ADD(hh,head,fieldname,keylen_in,add); \
|
180
|
+
} while(0)
|
181
|
+
|
182
|
+
#define HASH_ADD_KEYPTR(hh,head,keyptr,keylen_in,add) \
|
183
|
+
do { \
|
184
|
+
unsigned _ha_bkt; \
|
185
|
+
(add)->hh.next = NULL; \
|
186
|
+
(add)->hh.key = (char*)(keyptr); \
|
187
|
+
(add)->hh.keylen = (unsigned)(keylen_in); \
|
188
|
+
if (!(head)) { \
|
189
|
+
head = (add); \
|
190
|
+
(head)->hh.prev = NULL; \
|
191
|
+
HASH_MAKE_TABLE(hh,head); \
|
192
|
+
} else { \
|
193
|
+
(head)->hh.tbl->tail->next = (add); \
|
194
|
+
(add)->hh.prev = ELMT_FROM_HH((head)->hh.tbl, (head)->hh.tbl->tail); \
|
195
|
+
(head)->hh.tbl->tail = &((add)->hh); \
|
196
|
+
} \
|
197
|
+
(head)->hh.tbl->num_items++; \
|
198
|
+
(add)->hh.tbl = (head)->hh.tbl; \
|
199
|
+
HASH_FCN(keyptr,keylen_in, (head)->hh.tbl->num_buckets, \
|
200
|
+
(add)->hh.hashv, _ha_bkt); \
|
201
|
+
HASH_ADD_TO_BKT((head)->hh.tbl->buckets[_ha_bkt],&(add)->hh); \
|
202
|
+
HASH_BLOOM_ADD((head)->hh.tbl,(add)->hh.hashv); \
|
203
|
+
HASH_EMIT_KEY(hh,head,keyptr,keylen_in); \
|
204
|
+
HASH_FSCK(hh,head); \
|
205
|
+
} while(0)
|
206
|
+
|
207
|
+
#define HASH_TO_BKT( hashv, num_bkts, bkt ) \
|
208
|
+
do { \
|
209
|
+
bkt = ((hashv) & ((num_bkts) - 1)); \
|
210
|
+
} while(0)
|
211
|
+
|
212
|
+
/* delete "delptr" from the hash table.
|
213
|
+
* "the usual" patch-up process for the app-order doubly-linked-list.
|
214
|
+
* The use of _hd_hh_del below deserves special explanation.
|
215
|
+
* These used to be expressed using (delptr) but that led to a bug
|
216
|
+
* if someone used the same symbol for the head and deletee, like
|
217
|
+
* HASH_DELETE(hh,users,users);
|
218
|
+
* We want that to work, but by changing the head (users) below
|
219
|
+
* we were forfeiting our ability to further refer to the deletee (users)
|
220
|
+
* in the patch-up process. Solution: use scratch space to
|
221
|
+
* copy the deletee pointer, then the latter references are via that
|
222
|
+
* scratch pointer rather than through the repointed (users) symbol.
|
223
|
+
*/
|
224
|
+
#define HASH_DELETE(hh,head,delptr) \
|
225
|
+
do { \
|
226
|
+
unsigned _hd_bkt; \
|
227
|
+
struct UT_hash_handle *_hd_hh_del; \
|
228
|
+
if ( ((delptr)->hh.prev == NULL) && ((delptr)->hh.next == NULL) ) { \
|
229
|
+
uthash_free((head)->hh.tbl->buckets, \
|
230
|
+
(head)->hh.tbl->num_buckets*sizeof(struct UT_hash_bucket) ); \
|
231
|
+
HASH_BLOOM_FREE((head)->hh.tbl); \
|
232
|
+
uthash_free((head)->hh.tbl, sizeof(UT_hash_table)); \
|
233
|
+
head = NULL; \
|
234
|
+
} else { \
|
235
|
+
_hd_hh_del = &((delptr)->hh); \
|
236
|
+
if ((delptr) == ELMT_FROM_HH((head)->hh.tbl,(head)->hh.tbl->tail)) { \
|
237
|
+
(head)->hh.tbl->tail = \
|
238
|
+
(UT_hash_handle*)((ptrdiff_t)((delptr)->hh.prev) + \
|
239
|
+
(head)->hh.tbl->hho); \
|
240
|
+
} \
|
241
|
+
if ((delptr)->hh.prev) { \
|
242
|
+
((UT_hash_handle*)((ptrdiff_t)((delptr)->hh.prev) + \
|
243
|
+
(head)->hh.tbl->hho))->next = (delptr)->hh.next; \
|
244
|
+
} else { \
|
245
|
+
DECLTYPE_ASSIGN(head,(delptr)->hh.next); \
|
246
|
+
} \
|
247
|
+
if (_hd_hh_del->next) { \
|
248
|
+
((UT_hash_handle*)((ptrdiff_t)_hd_hh_del->next + \
|
249
|
+
(head)->hh.tbl->hho))->prev = \
|
250
|
+
_hd_hh_del->prev; \
|
251
|
+
} \
|
252
|
+
HASH_TO_BKT( _hd_hh_del->hashv, (head)->hh.tbl->num_buckets, _hd_bkt); \
|
253
|
+
HASH_DEL_IN_BKT(hh,(head)->hh.tbl->buckets[_hd_bkt], _hd_hh_del); \
|
254
|
+
(head)->hh.tbl->num_items--; \
|
255
|
+
} \
|
256
|
+
HASH_FSCK(hh,head); \
|
257
|
+
} while (0)
|
258
|
+
|
259
|
+
|
260
|
+
/* convenience forms of HASH_FIND/HASH_ADD/HASH_DEL */
|
261
|
+
#define HASH_FIND_STR(head,findstr,out) \
|
262
|
+
HASH_FIND(hh,head,findstr,strlen(findstr),out)
|
263
|
+
#define HASH_ADD_STR(head,strfield,add) \
|
264
|
+
HASH_ADD(hh,head,strfield[0],strlen(add->strfield),add)
|
265
|
+
#define HASH_REPLACE_STR(head,strfield,add,replaced) \
|
266
|
+
HASH_REPLACE(hh,head,strfield[0],strlen(add->strfield),add,replaced)
|
267
|
+
#define HASH_FIND_INT(head,findint,out) \
|
268
|
+
HASH_FIND(hh,head,findint,sizeof(int),out)
|
269
|
+
#define HASH_ADD_INT(head,intfield,add) \
|
270
|
+
HASH_ADD(hh,head,intfield,sizeof(int),add)
|
271
|
+
#define HASH_REPLACE_INT(head,intfield,add,replaced) \
|
272
|
+
HASH_REPLACE(hh,head,intfield,sizeof(int),add,replaced)
|
273
|
+
#define HASH_FIND_PTR(head,findptr,out) \
|
274
|
+
HASH_FIND(hh,head,findptr,sizeof(void *),out)
|
275
|
+
#define HASH_ADD_PTR(head,ptrfield,add) \
|
276
|
+
HASH_ADD(hh,head,ptrfield,sizeof(void *),add)
|
277
|
+
#define HASH_REPLACE_PTR(head,ptrfield,add,replaced) \
|
278
|
+
HASH_REPLACE(hh,head,ptrfield,sizeof(void *),add,replaced)
|
279
|
+
#define HASH_DEL(head,delptr) \
|
280
|
+
HASH_DELETE(hh,head,delptr)
|
281
|
+
|
282
|
+
/* HASH_FSCK checks hash integrity on every add/delete when HASH_DEBUG is defined.
|
283
|
+
* This is for uthash developer only; it compiles away if HASH_DEBUG isn't defined.
|
284
|
+
*/
|
285
|
+
#ifdef HASH_DEBUG
|
286
|
+
#define HASH_OOPS(...) do { fprintf(stderr,__VA_ARGS__); exit(-1); } while (0)
|
287
|
+
#define HASH_FSCK(hh,head) \
|
288
|
+
do { \
|
289
|
+
unsigned _bkt_i; \
|
290
|
+
unsigned _count, _bkt_count; \
|
291
|
+
char *_prev; \
|
292
|
+
struct UT_hash_handle *_thh; \
|
293
|
+
if (head) { \
|
294
|
+
_count = 0; \
|
295
|
+
for( _bkt_i = 0; _bkt_i < (head)->hh.tbl->num_buckets; _bkt_i++) { \
|
296
|
+
_bkt_count = 0; \
|
297
|
+
_thh = (head)->hh.tbl->buckets[_bkt_i].hh_head; \
|
298
|
+
_prev = NULL; \
|
299
|
+
while (_thh) { \
|
300
|
+
if (_prev != (char*)(_thh->hh_prev)) { \
|
301
|
+
HASH_OOPS("invalid hh_prev %p, actual %p\n", \
|
302
|
+
_thh->hh_prev, _prev ); \
|
303
|
+
} \
|
304
|
+
_bkt_count++; \
|
305
|
+
_prev = (char*)(_thh); \
|
306
|
+
_thh = _thh->hh_next; \
|
307
|
+
} \
|
308
|
+
_count += _bkt_count; \
|
309
|
+
if ((head)->hh.tbl->buckets[_bkt_i].count != _bkt_count) { \
|
310
|
+
HASH_OOPS("invalid bucket count %d, actual %d\n", \
|
311
|
+
(head)->hh.tbl->buckets[_bkt_i].count, _bkt_count); \
|
312
|
+
} \
|
313
|
+
} \
|
314
|
+
if (_count != (head)->hh.tbl->num_items) { \
|
315
|
+
HASH_OOPS("invalid hh item count %d, actual %d\n", \
|
316
|
+
(head)->hh.tbl->num_items, _count ); \
|
317
|
+
} \
|
318
|
+
/* traverse hh in app order; check next/prev integrity, count */ \
|
319
|
+
_count = 0; \
|
320
|
+
_prev = NULL; \
|
321
|
+
_thh = &(head)->hh; \
|
322
|
+
while (_thh) { \
|
323
|
+
_count++; \
|
324
|
+
if (_prev !=(char*)(_thh->prev)) { \
|
325
|
+
HASH_OOPS("invalid prev %p, actual %p\n", \
|
326
|
+
_thh->prev, _prev ); \
|
327
|
+
} \
|
328
|
+
_prev = (char*)ELMT_FROM_HH((head)->hh.tbl, _thh); \
|
329
|
+
_thh = ( _thh->next ? (UT_hash_handle*)((char*)(_thh->next) + \
|
330
|
+
(head)->hh.tbl->hho) : NULL ); \
|
331
|
+
} \
|
332
|
+
if (_count != (head)->hh.tbl->num_items) { \
|
333
|
+
HASH_OOPS("invalid app item count %d, actual %d\n", \
|
334
|
+
(head)->hh.tbl->num_items, _count ); \
|
335
|
+
} \
|
336
|
+
} \
|
337
|
+
} while (0)
|
338
|
+
#else
|
339
|
+
#define HASH_FSCK(hh,head)
|
340
|
+
#endif
|
341
|
+
|
342
|
+
/* When compiled with -DHASH_EMIT_KEYS, length-prefixed keys are emitted to
|
343
|
+
* the descriptor to which this macro is defined for tuning the hash function.
|
344
|
+
* The app can #include <unistd.h> to get the prototype for write(2). */
|
345
|
+
#ifdef HASH_EMIT_KEYS
|
346
|
+
#define HASH_EMIT_KEY(hh,head,keyptr,fieldlen) \
|
347
|
+
do { \
|
348
|
+
unsigned _klen = fieldlen; \
|
349
|
+
write(HASH_EMIT_KEYS, &_klen, sizeof(_klen)); \
|
350
|
+
write(HASH_EMIT_KEYS, keyptr, fieldlen); \
|
351
|
+
} while (0)
|
352
|
+
#else
|
353
|
+
#define HASH_EMIT_KEY(hh,head,keyptr,fieldlen)
|
354
|
+
#endif
|
355
|
+
|
356
|
+
/* default to Jenkin's hash unless overridden e.g. DHASH_FUNCTION=HASH_SAX */
|
357
|
+
#ifdef HASH_FUNCTION
|
358
|
+
#define HASH_FCN HASH_FUNCTION
|
359
|
+
#else
|
360
|
+
#define HASH_FCN HASH_JEN
|
361
|
+
#endif
|
362
|
+
|
363
|
+
/* The Bernstein hash function, used in Perl prior to v5.6. Note (x<<5+x)=x*33. */
|
364
|
+
#define HASH_BER(key,keylen,num_bkts,hashv,bkt) \
|
365
|
+
do { \
|
366
|
+
unsigned _hb_keylen=keylen; \
|
367
|
+
char *_hb_key=(char*)(key); \
|
368
|
+
(hashv) = 0; \
|
369
|
+
while (_hb_keylen--) { (hashv) = (((hashv) << 5) + (hashv)) + *_hb_key++; } \
|
370
|
+
bkt = (hashv) & (num_bkts-1); \
|
371
|
+
} while (0)
|
372
|
+
|
373
|
+
|
374
|
+
/* SAX/FNV/OAT/JEN hash functions are macro variants of those listed at
|
375
|
+
* http://eternallyconfuzzled.com/tuts/algorithms/jsw_tut_hashing.aspx */
|
376
|
+
#define HASH_SAX(key,keylen,num_bkts,hashv,bkt) \
|
377
|
+
do { \
|
378
|
+
unsigned _sx_i; \
|
379
|
+
char *_hs_key=(char*)(key); \
|
380
|
+
hashv = 0; \
|
381
|
+
for(_sx_i=0; _sx_i < keylen; _sx_i++) \
|
382
|
+
hashv ^= (hashv << 5) + (hashv >> 2) + _hs_key[_sx_i]; \
|
383
|
+
bkt = hashv & (num_bkts-1); \
|
384
|
+
} while (0)
|
385
|
+
/* FNV-1a variation */
|
386
|
+
#define HASH_FNV(key,keylen,num_bkts,hashv,bkt) \
|
387
|
+
do { \
|
388
|
+
unsigned _fn_i; \
|
389
|
+
char *_hf_key=(char*)(key); \
|
390
|
+
hashv = 2166136261UL; \
|
391
|
+
for(_fn_i=0; _fn_i < keylen; _fn_i++) \
|
392
|
+
hashv = hashv ^ _hf_key[_fn_i]; \
|
393
|
+
hashv = hashv * 16777619; \
|
394
|
+
bkt = hashv & (num_bkts-1); \
|
395
|
+
} while(0)
|
396
|
+
|
397
|
+
#define HASH_OAT(key,keylen,num_bkts,hashv,bkt) \
|
398
|
+
do { \
|
399
|
+
unsigned _ho_i; \
|
400
|
+
char *_ho_key=(char*)(key); \
|
401
|
+
hashv = 0; \
|
402
|
+
for(_ho_i=0; _ho_i < keylen; _ho_i++) { \
|
403
|
+
hashv += _ho_key[_ho_i]; \
|
404
|
+
hashv += (hashv << 10); \
|
405
|
+
hashv ^= (hashv >> 6); \
|
406
|
+
} \
|
407
|
+
hashv += (hashv << 3); \
|
408
|
+
hashv ^= (hashv >> 11); \
|
409
|
+
hashv += (hashv << 15); \
|
410
|
+
bkt = hashv & (num_bkts-1); \
|
411
|
+
} while(0)
|
412
|
+
|
413
|
+
#define HASH_JEN_MIX(a,b,c) \
|
414
|
+
do { \
|
415
|
+
a -= b; a -= c; a ^= ( c >> 13 ); \
|
416
|
+
b -= c; b -= a; b ^= ( a << 8 ); \
|
417
|
+
c -= a; c -= b; c ^= ( b >> 13 ); \
|
418
|
+
a -= b; a -= c; a ^= ( c >> 12 ); \
|
419
|
+
b -= c; b -= a; b ^= ( a << 16 ); \
|
420
|
+
c -= a; c -= b; c ^= ( b >> 5 ); \
|
421
|
+
a -= b; a -= c; a ^= ( c >> 3 ); \
|
422
|
+
b -= c; b -= a; b ^= ( a << 10 ); \
|
423
|
+
c -= a; c -= b; c ^= ( b >> 15 ); \
|
424
|
+
} while (0)
|
425
|
+
|
426
|
+
#define HASH_JEN(key,keylen,num_bkts,hashv,bkt) \
|
427
|
+
do { \
|
428
|
+
unsigned _hj_i,_hj_j,_hj_k; \
|
429
|
+
unsigned char *_hj_key=(unsigned char*)(key); \
|
430
|
+
hashv = 0xfeedbeef; \
|
431
|
+
_hj_i = _hj_j = 0x9e3779b9; \
|
432
|
+
_hj_k = (unsigned)(keylen); \
|
433
|
+
while (_hj_k >= 12) { \
|
434
|
+
_hj_i += (_hj_key[0] + ( (unsigned)_hj_key[1] << 8 ) \
|
435
|
+
+ ( (unsigned)_hj_key[2] << 16 ) \
|
436
|
+
+ ( (unsigned)_hj_key[3] << 24 ) ); \
|
437
|
+
_hj_j += (_hj_key[4] + ( (unsigned)_hj_key[5] << 8 ) \
|
438
|
+
+ ( (unsigned)_hj_key[6] << 16 ) \
|
439
|
+
+ ( (unsigned)_hj_key[7] << 24 ) ); \
|
440
|
+
hashv += (_hj_key[8] + ( (unsigned)_hj_key[9] << 8 ) \
|
441
|
+
+ ( (unsigned)_hj_key[10] << 16 ) \
|
442
|
+
+ ( (unsigned)_hj_key[11] << 24 ) ); \
|
443
|
+
\
|
444
|
+
HASH_JEN_MIX(_hj_i, _hj_j, hashv); \
|
445
|
+
\
|
446
|
+
_hj_key += 12; \
|
447
|
+
_hj_k -= 12; \
|
448
|
+
} \
|
449
|
+
hashv += keylen; \
|
450
|
+
switch ( _hj_k ) { \
|
451
|
+
case 11: hashv += ( (unsigned)_hj_key[10] << 24 ); \
|
452
|
+
case 10: hashv += ( (unsigned)_hj_key[9] << 16 ); \
|
453
|
+
case 9: hashv += ( (unsigned)_hj_key[8] << 8 ); \
|
454
|
+
case 8: _hj_j += ( (unsigned)_hj_key[7] << 24 ); \
|
455
|
+
case 7: _hj_j += ( (unsigned)_hj_key[6] << 16 ); \
|
456
|
+
case 6: _hj_j += ( (unsigned)_hj_key[5] << 8 ); \
|
457
|
+
case 5: _hj_j += _hj_key[4]; \
|
458
|
+
case 4: _hj_i += ( (unsigned)_hj_key[3] << 24 ); \
|
459
|
+
case 3: _hj_i += ( (unsigned)_hj_key[2] << 16 ); \
|
460
|
+
case 2: _hj_i += ( (unsigned)_hj_key[1] << 8 ); \
|
461
|
+
case 1: _hj_i += _hj_key[0]; \
|
462
|
+
} \
|
463
|
+
HASH_JEN_MIX(_hj_i, _hj_j, hashv); \
|
464
|
+
bkt = hashv & (num_bkts-1); \
|
465
|
+
} while(0)
|
466
|
+
|
467
|
+
/* The Paul Hsieh hash function */
|
468
|
+
#undef get16bits
|
469
|
+
#if (defined(__GNUC__) && defined(__i386__)) || defined(__WATCOMC__) \
|
470
|
+
|| defined(_MSC_VER) || defined (__BORLANDC__) || defined (__TURBOC__)
|
471
|
+
#define get16bits(d) (*((const uint16_t *) (d)))
|
472
|
+
#endif
|
473
|
+
|
474
|
+
#if !defined (get16bits)
|
475
|
+
#define get16bits(d) ((((uint32_t)(((const uint8_t *)(d))[1])) << 8) \
|
476
|
+
+(uint32_t)(((const uint8_t *)(d))[0]) )
|
477
|
+
#endif
|
478
|
+
#define HASH_SFH(key,keylen,num_bkts,hashv,bkt) \
|
479
|
+
do { \
|
480
|
+
unsigned char *_sfh_key=(unsigned char*)(key); \
|
481
|
+
uint32_t _sfh_tmp, _sfh_len = keylen; \
|
482
|
+
\
|
483
|
+
int _sfh_rem = _sfh_len & 3; \
|
484
|
+
_sfh_len >>= 2; \
|
485
|
+
hashv = 0xcafebabe; \
|
486
|
+
\
|
487
|
+
/* Main loop */ \
|
488
|
+
for (;_sfh_len > 0; _sfh_len--) { \
|
489
|
+
hashv += get16bits (_sfh_key); \
|
490
|
+
_sfh_tmp = (uint32_t)(get16bits (_sfh_key+2)) << 11 ^ hashv; \
|
491
|
+
hashv = (hashv << 16) ^ _sfh_tmp; \
|
492
|
+
_sfh_key += 2*sizeof (uint16_t); \
|
493
|
+
hashv += hashv >> 11; \
|
494
|
+
} \
|
495
|
+
\
|
496
|
+
/* Handle end cases */ \
|
497
|
+
switch (_sfh_rem) { \
|
498
|
+
case 3: hashv += get16bits (_sfh_key); \
|
499
|
+
hashv ^= hashv << 16; \
|
500
|
+
hashv ^= (uint32_t)(_sfh_key[sizeof (uint16_t)] << 18); \
|
501
|
+
hashv += hashv >> 11; \
|
502
|
+
break; \
|
503
|
+
case 2: hashv += get16bits (_sfh_key); \
|
504
|
+
hashv ^= hashv << 11; \
|
505
|
+
hashv += hashv >> 17; \
|
506
|
+
break; \
|
507
|
+
case 1: hashv += *_sfh_key; \
|
508
|
+
hashv ^= hashv << 10; \
|
509
|
+
hashv += hashv >> 1; \
|
510
|
+
} \
|
511
|
+
\
|
512
|
+
/* Force "avalanching" of final 127 bits */ \
|
513
|
+
hashv ^= hashv << 3; \
|
514
|
+
hashv += hashv >> 5; \
|
515
|
+
hashv ^= hashv << 4; \
|
516
|
+
hashv += hashv >> 17; \
|
517
|
+
hashv ^= hashv << 25; \
|
518
|
+
hashv += hashv >> 6; \
|
519
|
+
bkt = hashv & (num_bkts-1); \
|
520
|
+
} while(0)
|
521
|
+
|
522
|
+
#ifdef HASH_USING_NO_STRICT_ALIASING
|
523
|
+
/* The MurmurHash exploits some CPU's (x86,x86_64) tolerance for unaligned reads.
|
524
|
+
* For other types of CPU's (e.g. Sparc) an unaligned read causes a bus error.
|
525
|
+
* MurmurHash uses the faster approach only on CPU's where we know it's safe.
|
526
|
+
*
|
527
|
+
* Note the preprocessor built-in defines can be emitted using:
|
528
|
+
*
|
529
|
+
* gcc -m64 -dM -E - < /dev/null (on gcc)
|
530
|
+
* cc -## a.c (where a.c is a simple test file) (Sun Studio)
|
531
|
+
*/
|
532
|
+
#if (defined(__i386__) || defined(__x86_64__) || defined(_M_IX86))
|
533
|
+
#define MUR_GETBLOCK(p,i) p[i]
|
534
|
+
#else /* non intel */
|
535
|
+
#define MUR_PLUS0_ALIGNED(p) (((unsigned long)p & 0x3) == 0)
|
536
|
+
#define MUR_PLUS1_ALIGNED(p) (((unsigned long)p & 0x3) == 1)
|
537
|
+
#define MUR_PLUS2_ALIGNED(p) (((unsigned long)p & 0x3) == 2)
|
538
|
+
#define MUR_PLUS3_ALIGNED(p) (((unsigned long)p & 0x3) == 3)
|
539
|
+
#define WP(p) ((uint32_t*)((unsigned long)(p) & ~3UL))
|
540
|
+
#if (defined(__BIG_ENDIAN__) || defined(SPARC) || defined(__ppc__) || defined(__ppc64__))
|
541
|
+
#define MUR_THREE_ONE(p) ((((*WP(p))&0x00ffffff) << 8) | (((*(WP(p)+1))&0xff000000) >> 24))
|
542
|
+
#define MUR_TWO_TWO(p) ((((*WP(p))&0x0000ffff) <<16) | (((*(WP(p)+1))&0xffff0000) >> 16))
|
543
|
+
#define MUR_ONE_THREE(p) ((((*WP(p))&0x000000ff) <<24) | (((*(WP(p)+1))&0xffffff00) >> 8))
|
544
|
+
#else /* assume little endian non-intel */
|
545
|
+
#define MUR_THREE_ONE(p) ((((*WP(p))&0xffffff00) >> 8) | (((*(WP(p)+1))&0x000000ff) << 24))
|
546
|
+
#define MUR_TWO_TWO(p) ((((*WP(p))&0xffff0000) >>16) | (((*(WP(p)+1))&0x0000ffff) << 16))
|
547
|
+
#define MUR_ONE_THREE(p) ((((*WP(p))&0xff000000) >>24) | (((*(WP(p)+1))&0x00ffffff) << 8))
|
548
|
+
#endif
|
549
|
+
#define MUR_GETBLOCK(p,i) (MUR_PLUS0_ALIGNED(p) ? ((p)[i]) : \
|
550
|
+
(MUR_PLUS1_ALIGNED(p) ? MUR_THREE_ONE(p) : \
|
551
|
+
(MUR_PLUS2_ALIGNED(p) ? MUR_TWO_TWO(p) : \
|
552
|
+
MUR_ONE_THREE(p))))
|
553
|
+
#endif
|
554
|
+
#define MUR_ROTL32(x,r) (((x) << (r)) | ((x) >> (32 - (r))))
|
555
|
+
#define MUR_FMIX(_h) \
|
556
|
+
do { \
|
557
|
+
_h ^= _h >> 16; \
|
558
|
+
_h *= 0x85ebca6b; \
|
559
|
+
_h ^= _h >> 13; \
|
560
|
+
_h *= 0xc2b2ae35l; \
|
561
|
+
_h ^= _h >> 16; \
|
562
|
+
} while(0)
|
563
|
+
|
564
|
+
#define HASH_MUR(key,keylen,num_bkts,hashv,bkt) \
|
565
|
+
do { \
|
566
|
+
const uint8_t *_mur_data = (const uint8_t*)(key); \
|
567
|
+
const int _mur_nblocks = (keylen) / 4; \
|
568
|
+
uint32_t _mur_h1 = 0xf88D5353; \
|
569
|
+
uint32_t _mur_c1 = 0xcc9e2d51; \
|
570
|
+
uint32_t _mur_c2 = 0x1b873593; \
|
571
|
+
uint32_t _mur_k1 = 0; \
|
572
|
+
const uint8_t *_mur_tail; \
|
573
|
+
const uint32_t *_mur_blocks = (const uint32_t*)(_mur_data+_mur_nblocks*4); \
|
574
|
+
int _mur_i; \
|
575
|
+
for(_mur_i = -_mur_nblocks; _mur_i; _mur_i++) { \
|
576
|
+
_mur_k1 = MUR_GETBLOCK(_mur_blocks,_mur_i); \
|
577
|
+
_mur_k1 *= _mur_c1; \
|
578
|
+
_mur_k1 = MUR_ROTL32(_mur_k1,15); \
|
579
|
+
_mur_k1 *= _mur_c2; \
|
580
|
+
\
|
581
|
+
_mur_h1 ^= _mur_k1; \
|
582
|
+
_mur_h1 = MUR_ROTL32(_mur_h1,13); \
|
583
|
+
_mur_h1 = _mur_h1*5+0xe6546b64; \
|
584
|
+
} \
|
585
|
+
_mur_tail = (const uint8_t*)(_mur_data + _mur_nblocks*4); \
|
586
|
+
_mur_k1=0; \
|
587
|
+
switch((keylen) & 3) { \
|
588
|
+
case 3: _mur_k1 ^= _mur_tail[2] << 16; \
|
589
|
+
case 2: _mur_k1 ^= _mur_tail[1] << 8; \
|
590
|
+
case 1: _mur_k1 ^= _mur_tail[0]; \
|
591
|
+
_mur_k1 *= _mur_c1; \
|
592
|
+
_mur_k1 = MUR_ROTL32(_mur_k1,15); \
|
593
|
+
_mur_k1 *= _mur_c2; \
|
594
|
+
_mur_h1 ^= _mur_k1; \
|
595
|
+
} \
|
596
|
+
_mur_h1 ^= (keylen); \
|
597
|
+
MUR_FMIX(_mur_h1); \
|
598
|
+
hashv = _mur_h1; \
|
599
|
+
bkt = hashv & (num_bkts-1); \
|
600
|
+
} while(0)
|
601
|
+
#endif /* HASH_USING_NO_STRICT_ALIASING */
|
602
|
+
|
603
|
+
/* key comparison function; return 0 if keys equal */
|
604
|
+
#define HASH_KEYCMP(a,b,len) memcmp(a,b,len)
|
605
|
+
|
606
|
+
/* iterate over items in a known bucket to find desired item */
|
607
|
+
#define HASH_FIND_IN_BKT(tbl,hh,head,keyptr,keylen_in,out) \
|
608
|
+
do { \
|
609
|
+
if (head.hh_head) DECLTYPE_ASSIGN(out,ELMT_FROM_HH(tbl,head.hh_head)); \
|
610
|
+
else out=NULL; \
|
611
|
+
while (out) { \
|
612
|
+
if ((out)->hh.keylen == keylen_in) { \
|
613
|
+
if ((HASH_KEYCMP((out)->hh.key,keyptr,keylen_in)) == 0) break; \
|
614
|
+
} \
|
615
|
+
if ((out)->hh.hh_next) DECLTYPE_ASSIGN(out,ELMT_FROM_HH(tbl,(out)->hh.hh_next)); \
|
616
|
+
else out = NULL; \
|
617
|
+
} \
|
618
|
+
} while(0)
|
619
|
+
|
620
|
+
/* add an item to a bucket */
|
621
|
+
#define HASH_ADD_TO_BKT(head,addhh) \
|
622
|
+
do { \
|
623
|
+
head.count++; \
|
624
|
+
(addhh)->hh_next = head.hh_head; \
|
625
|
+
(addhh)->hh_prev = NULL; \
|
626
|
+
if (head.hh_head) { (head).hh_head->hh_prev = (addhh); } \
|
627
|
+
(head).hh_head=addhh; \
|
628
|
+
if (head.count >= ((head.expand_mult+1) * HASH_BKT_CAPACITY_THRESH) \
|
629
|
+
&& (addhh)->tbl->noexpand != 1) { \
|
630
|
+
HASH_EXPAND_BUCKETS((addhh)->tbl); \
|
631
|
+
} \
|
632
|
+
} while(0)
|
633
|
+
|
634
|
+
/* remove an item from a given bucket */
|
635
|
+
#define HASH_DEL_IN_BKT(hh,head,hh_del) \
|
636
|
+
(head).count--; \
|
637
|
+
if ((head).hh_head == hh_del) { \
|
638
|
+
(head).hh_head = hh_del->hh_next; \
|
639
|
+
} \
|
640
|
+
if (hh_del->hh_prev) { \
|
641
|
+
hh_del->hh_prev->hh_next = hh_del->hh_next; \
|
642
|
+
} \
|
643
|
+
if (hh_del->hh_next) { \
|
644
|
+
hh_del->hh_next->hh_prev = hh_del->hh_prev; \
|
645
|
+
}
|
646
|
+
|
647
|
+
/* Bucket expansion has the effect of doubling the number of buckets
|
648
|
+
* and redistributing the items into the new buckets. Ideally the
|
649
|
+
* items will distribute more or less evenly into the new buckets
|
650
|
+
* (the extent to which this is true is a measure of the quality of
|
651
|
+
* the hash function as it applies to the key domain).
|
652
|
+
*
|
653
|
+
* With the items distributed into more buckets, the chain length
|
654
|
+
* (item count) in each bucket is reduced. Thus by expanding buckets
|
655
|
+
* the hash keeps a bound on the chain length. This bounded chain
|
656
|
+
* length is the essence of how a hash provides constant time lookup.
|
657
|
+
*
|
658
|
+
* The calculation of tbl->ideal_chain_maxlen below deserves some
|
659
|
+
* explanation. First, keep in mind that we're calculating the ideal
|
660
|
+
* maximum chain length based on the *new* (doubled) bucket count.
|
661
|
+
* In fractions this is just n/b (n=number of items,b=new num buckets).
|
662
|
+
* Since the ideal chain length is an integer, we want to calculate
|
663
|
+
* ceil(n/b). We don't depend on floating point arithmetic in this
|
664
|
+
* hash, so to calculate ceil(n/b) with integers we could write
|
665
|
+
*
|
666
|
+
* ceil(n/b) = (n/b) + ((n%b)?1:0)
|
667
|
+
*
|
668
|
+
* and in fact a previous version of this hash did just that.
|
669
|
+
* But now we have improved things a bit by recognizing that b is
|
670
|
+
* always a power of two. We keep its base 2 log handy (call it lb),
|
671
|
+
* so now we can write this with a bit shift and logical AND:
|
672
|
+
*
|
673
|
+
* ceil(n/b) = (n>>lb) + ( (n & (b-1)) ? 1:0)
|
674
|
+
*
|
675
|
+
*/
|
676
|
+
#define HASH_EXPAND_BUCKETS(tbl) \
|
677
|
+
do { \
|
678
|
+
unsigned _he_bkt; \
|
679
|
+
unsigned _he_bkt_i; \
|
680
|
+
struct UT_hash_handle *_he_thh, *_he_hh_nxt; \
|
681
|
+
UT_hash_bucket *_he_new_buckets, *_he_newbkt; \
|
682
|
+
_he_new_buckets = (UT_hash_bucket*)uthash_malloc( \
|
683
|
+
2 * tbl->num_buckets * sizeof(struct UT_hash_bucket)); \
|
684
|
+
if (!_he_new_buckets) { uthash_fatal( "out of memory"); } \
|
685
|
+
memset(_he_new_buckets, 0, \
|
686
|
+
2 * tbl->num_buckets * sizeof(struct UT_hash_bucket)); \
|
687
|
+
tbl->ideal_chain_maxlen = \
|
688
|
+
(tbl->num_items >> (tbl->log2_num_buckets+1)) + \
|
689
|
+
((tbl->num_items & ((tbl->num_buckets*2)-1)) ? 1 : 0); \
|
690
|
+
tbl->nonideal_items = 0; \
|
691
|
+
for(_he_bkt_i = 0; _he_bkt_i < tbl->num_buckets; _he_bkt_i++) \
|
692
|
+
{ \
|
693
|
+
_he_thh = tbl->buckets[ _he_bkt_i ].hh_head; \
|
694
|
+
while (_he_thh) { \
|
695
|
+
_he_hh_nxt = _he_thh->hh_next; \
|
696
|
+
HASH_TO_BKT( _he_thh->hashv, tbl->num_buckets*2, _he_bkt); \
|
697
|
+
_he_newbkt = &(_he_new_buckets[ _he_bkt ]); \
|
698
|
+
if (++(_he_newbkt->count) > tbl->ideal_chain_maxlen) { \
|
699
|
+
tbl->nonideal_items++; \
|
700
|
+
_he_newbkt->expand_mult = _he_newbkt->count / \
|
701
|
+
tbl->ideal_chain_maxlen; \
|
702
|
+
} \
|
703
|
+
_he_thh->hh_prev = NULL; \
|
704
|
+
_he_thh->hh_next = _he_newbkt->hh_head; \
|
705
|
+
if (_he_newbkt->hh_head) _he_newbkt->hh_head->hh_prev = \
|
706
|
+
_he_thh; \
|
707
|
+
_he_newbkt->hh_head = _he_thh; \
|
708
|
+
_he_thh = _he_hh_nxt; \
|
709
|
+
} \
|
710
|
+
} \
|
711
|
+
uthash_free( tbl->buckets, tbl->num_buckets*sizeof(struct UT_hash_bucket) ); \
|
712
|
+
tbl->num_buckets *= 2; \
|
713
|
+
tbl->log2_num_buckets++; \
|
714
|
+
tbl->buckets = _he_new_buckets; \
|
715
|
+
tbl->ineff_expands = (tbl->nonideal_items > (tbl->num_items >> 1)) ? \
|
716
|
+
(tbl->ineff_expands+1) : 0; \
|
717
|
+
if (tbl->ineff_expands > 1) { \
|
718
|
+
tbl->noexpand=1; \
|
719
|
+
uthash_noexpand_fyi(tbl); \
|
720
|
+
} \
|
721
|
+
uthash_expand_fyi(tbl); \
|
722
|
+
} while(0)
|
723
|
+
|
724
|
+
|
725
|
+
/* This is an adaptation of Simon Tatham's O(n log(n)) mergesort */
|
726
|
+
/* Note that HASH_SORT assumes the hash handle name to be hh.
|
727
|
+
* HASH_SRT was added to allow the hash handle name to be passed in. */
|
728
|
+
#define HASH_SORT(head,cmpfcn) HASH_SRT(hh,head,cmpfcn)
|
729
|
+
#define HASH_SRT(hh,head,cmpfcn) \
|
730
|
+
do { \
|
731
|
+
unsigned _hs_i; \
|
732
|
+
unsigned _hs_looping,_hs_nmerges,_hs_insize,_hs_psize,_hs_qsize; \
|
733
|
+
struct UT_hash_handle *_hs_p, *_hs_q, *_hs_e, *_hs_list, *_hs_tail; \
|
734
|
+
if (head) { \
|
735
|
+
_hs_insize = 1; \
|
736
|
+
_hs_looping = 1; \
|
737
|
+
_hs_list = &((head)->hh); \
|
738
|
+
while (_hs_looping) { \
|
739
|
+
_hs_p = _hs_list; \
|
740
|
+
_hs_list = NULL; \
|
741
|
+
_hs_tail = NULL; \
|
742
|
+
_hs_nmerges = 0; \
|
743
|
+
while (_hs_p) { \
|
744
|
+
_hs_nmerges++; \
|
745
|
+
_hs_q = _hs_p; \
|
746
|
+
_hs_psize = 0; \
|
747
|
+
for ( _hs_i = 0; _hs_i < _hs_insize; _hs_i++ ) { \
|
748
|
+
_hs_psize++; \
|
749
|
+
_hs_q = (UT_hash_handle*)((_hs_q->next) ? \
|
750
|
+
((void*)((char*)(_hs_q->next) + \
|
751
|
+
(head)->hh.tbl->hho)) : NULL); \
|
752
|
+
if (! (_hs_q) ) break; \
|
753
|
+
} \
|
754
|
+
_hs_qsize = _hs_insize; \
|
755
|
+
while ((_hs_psize > 0) || ((_hs_qsize > 0) && _hs_q )) { \
|
756
|
+
if (_hs_psize == 0) { \
|
757
|
+
_hs_e = _hs_q; \
|
758
|
+
_hs_q = (UT_hash_handle*)((_hs_q->next) ? \
|
759
|
+
((void*)((char*)(_hs_q->next) + \
|
760
|
+
(head)->hh.tbl->hho)) : NULL); \
|
761
|
+
_hs_qsize--; \
|
762
|
+
} else if ( (_hs_qsize == 0) || !(_hs_q) ) { \
|
763
|
+
_hs_e = _hs_p; \
|
764
|
+
if (_hs_p){ \
|
765
|
+
_hs_p = (UT_hash_handle*)((_hs_p->next) ? \
|
766
|
+
((void*)((char*)(_hs_p->next) + \
|
767
|
+
(head)->hh.tbl->hho)) : NULL); \
|
768
|
+
} \
|
769
|
+
_hs_psize--; \
|
770
|
+
} else if (( \
|
771
|
+
cmpfcn(DECLTYPE(head)(ELMT_FROM_HH((head)->hh.tbl,_hs_p)), \
|
772
|
+
DECLTYPE(head)(ELMT_FROM_HH((head)->hh.tbl,_hs_q))) \
|
773
|
+
) <= 0) { \
|
774
|
+
_hs_e = _hs_p; \
|
775
|
+
if (_hs_p){ \
|
776
|
+
_hs_p = (UT_hash_handle*)((_hs_p->next) ? \
|
777
|
+
((void*)((char*)(_hs_p->next) + \
|
778
|
+
(head)->hh.tbl->hho)) : NULL); \
|
779
|
+
} \
|
780
|
+
_hs_psize--; \
|
781
|
+
} else { \
|
782
|
+
_hs_e = _hs_q; \
|
783
|
+
_hs_q = (UT_hash_handle*)((_hs_q->next) ? \
|
784
|
+
((void*)((char*)(_hs_q->next) + \
|
785
|
+
(head)->hh.tbl->hho)) : NULL); \
|
786
|
+
_hs_qsize--; \
|
787
|
+
} \
|
788
|
+
if ( _hs_tail ) { \
|
789
|
+
_hs_tail->next = ((_hs_e) ? \
|
790
|
+
ELMT_FROM_HH((head)->hh.tbl,_hs_e) : NULL); \
|
791
|
+
} else { \
|
792
|
+
_hs_list = _hs_e; \
|
793
|
+
} \
|
794
|
+
if (_hs_e) { \
|
795
|
+
_hs_e->prev = ((_hs_tail) ? \
|
796
|
+
ELMT_FROM_HH((head)->hh.tbl,_hs_tail) : NULL); \
|
797
|
+
} \
|
798
|
+
_hs_tail = _hs_e; \
|
799
|
+
} \
|
800
|
+
_hs_p = _hs_q; \
|
801
|
+
} \
|
802
|
+
if (_hs_tail){ \
|
803
|
+
_hs_tail->next = NULL; \
|
804
|
+
} \
|
805
|
+
if ( _hs_nmerges <= 1 ) { \
|
806
|
+
_hs_looping=0; \
|
807
|
+
(head)->hh.tbl->tail = _hs_tail; \
|
808
|
+
DECLTYPE_ASSIGN(head,ELMT_FROM_HH((head)->hh.tbl, _hs_list)); \
|
809
|
+
} \
|
810
|
+
_hs_insize *= 2; \
|
811
|
+
} \
|
812
|
+
HASH_FSCK(hh,head); \
|
813
|
+
} \
|
814
|
+
} while (0)
|
815
|
+
|
816
|
+
/* This function selects items from one hash into another hash.
|
817
|
+
* The end result is that the selected items have dual presence
|
818
|
+
* in both hashes. There is no copy of the items made; rather
|
819
|
+
* they are added into the new hash through a secondary hash
|
820
|
+
* hash handle that must be present in the structure. */
|
821
|
+
#define HASH_SELECT(hh_dst, dst, hh_src, src, cond) \
|
822
|
+
do { \
|
823
|
+
unsigned _src_bkt, _dst_bkt; \
|
824
|
+
void *_last_elt=NULL, *_elt; \
|
825
|
+
UT_hash_handle *_src_hh, *_dst_hh, *_last_elt_hh=NULL; \
|
826
|
+
ptrdiff_t _dst_hho = ((char*)(&(dst)->hh_dst) - (char*)(dst)); \
|
827
|
+
if (src) { \
|
828
|
+
for(_src_bkt=0; _src_bkt < (src)->hh_src.tbl->num_buckets; _src_bkt++) { \
|
829
|
+
for(_src_hh = (src)->hh_src.tbl->buckets[_src_bkt].hh_head; \
|
830
|
+
_src_hh; \
|
831
|
+
_src_hh = _src_hh->hh_next) { \
|
832
|
+
_elt = ELMT_FROM_HH((src)->hh_src.tbl, _src_hh); \
|
833
|
+
if (cond(_elt)) { \
|
834
|
+
_dst_hh = (UT_hash_handle*)(((char*)_elt) + _dst_hho); \
|
835
|
+
_dst_hh->key = _src_hh->key; \
|
836
|
+
_dst_hh->keylen = _src_hh->keylen; \
|
837
|
+
_dst_hh->hashv = _src_hh->hashv; \
|
838
|
+
_dst_hh->prev = _last_elt; \
|
839
|
+
_dst_hh->next = NULL; \
|
840
|
+
if (_last_elt_hh) { _last_elt_hh->next = _elt; } \
|
841
|
+
if (!dst) { \
|
842
|
+
DECLTYPE_ASSIGN(dst,_elt); \
|
843
|
+
HASH_MAKE_TABLE(hh_dst,dst); \
|
844
|
+
} else { \
|
845
|
+
_dst_hh->tbl = (dst)->hh_dst.tbl; \
|
846
|
+
} \
|
847
|
+
HASH_TO_BKT(_dst_hh->hashv, _dst_hh->tbl->num_buckets, _dst_bkt); \
|
848
|
+
HASH_ADD_TO_BKT(_dst_hh->tbl->buckets[_dst_bkt],_dst_hh); \
|
849
|
+
(dst)->hh_dst.tbl->num_items++; \
|
850
|
+
_last_elt = _elt; \
|
851
|
+
_last_elt_hh = _dst_hh; \
|
852
|
+
} \
|
853
|
+
} \
|
854
|
+
} \
|
855
|
+
} \
|
856
|
+
HASH_FSCK(hh_dst,dst); \
|
857
|
+
} while (0)
|
858
|
+
|
859
|
+
#define HASH_CLEAR(hh,head) \
|
860
|
+
do { \
|
861
|
+
if (head) { \
|
862
|
+
uthash_free((head)->hh.tbl->buckets, \
|
863
|
+
(head)->hh.tbl->num_buckets*sizeof(struct UT_hash_bucket)); \
|
864
|
+
HASH_BLOOM_FREE((head)->hh.tbl); \
|
865
|
+
uthash_free((head)->hh.tbl, sizeof(UT_hash_table)); \
|
866
|
+
(head)=NULL; \
|
867
|
+
} \
|
868
|
+
} while(0)
|
869
|
+
|
870
|
+
#define HASH_OVERHEAD(hh,head) \
|
871
|
+
(size_t)((((head)->hh.tbl->num_items * sizeof(UT_hash_handle)) + \
|
872
|
+
((head)->hh.tbl->num_buckets * sizeof(UT_hash_bucket)) + \
|
873
|
+
(sizeof(UT_hash_table)) + \
|
874
|
+
(HASH_BLOOM_BYTELEN)))
|
875
|
+
|
876
|
+
#ifdef NO_DECLTYPE
|
877
|
+
#define HASH_ITER(hh,head,el,tmp) \
|
878
|
+
for((el)=(head), (*(char**)(&(tmp)))=(char*)((head)?(head)->hh.next:NULL); \
|
879
|
+
el; (el)=(tmp),(*(char**)(&(tmp)))=(char*)((tmp)?(tmp)->hh.next:NULL))
|
880
|
+
#else
|
881
|
+
#define HASH_ITER(hh,head,el,tmp) \
|
882
|
+
for((el)=(head),(tmp)=DECLTYPE(el)((head)?(head)->hh.next:NULL); \
|
883
|
+
el; (el)=(tmp),(tmp)=DECLTYPE(el)((tmp)?(tmp)->hh.next:NULL))
|
884
|
+
#endif
|
885
|
+
|
886
|
+
/* obtain a count of items in the hash */
|
887
|
+
#define HASH_COUNT(head) HASH_CNT(hh,head)
|
888
|
+
#define HASH_CNT(hh,head) ((head)?((head)->hh.tbl->num_items):0)
|
889
|
+
|
890
|
+
typedef struct UT_hash_bucket {
|
891
|
+
struct UT_hash_handle *hh_head;
|
892
|
+
unsigned count;
|
893
|
+
|
894
|
+
/* expand_mult is normally set to 0. In this situation, the max chain length
|
895
|
+
* threshold is enforced at its default value, HASH_BKT_CAPACITY_THRESH. (If
|
896
|
+
* the bucket's chain exceeds this length, bucket expansion is triggered).
|
897
|
+
* However, setting expand_mult to a non-zero value delays bucket expansion
|
898
|
+
* (that would be triggered by additions to this particular bucket)
|
899
|
+
* until its chain length reaches a *multiple* of HASH_BKT_CAPACITY_THRESH.
|
900
|
+
* (The multiplier is simply expand_mult+1). The whole idea of this
|
901
|
+
* multiplier is to reduce bucket expansions, since they are expensive, in
|
902
|
+
* situations where we know that a particular bucket tends to be overused.
|
903
|
+
* It is better to let its chain length grow to a longer yet-still-bounded
|
904
|
+
* value, than to do an O(n) bucket expansion too often.
|
905
|
+
*/
|
906
|
+
unsigned expand_mult;
|
907
|
+
|
908
|
+
} UT_hash_bucket;
|
909
|
+
|
910
|
+
/* random signature used only to find hash tables in external analysis */
|
911
|
+
#define HASH_SIGNATURE 0xa0111fe1
|
912
|
+
#define HASH_BLOOM_SIGNATURE 0xb12220f2
|
913
|
+
|
914
|
+
typedef struct UT_hash_table {
|
915
|
+
UT_hash_bucket *buckets;
|
916
|
+
unsigned num_buckets, log2_num_buckets;
|
917
|
+
unsigned num_items;
|
918
|
+
struct UT_hash_handle *tail; /* tail hh in app order, for fast append */
|
919
|
+
ptrdiff_t hho; /* hash handle offset (byte pos of hash handle in element */
|
920
|
+
|
921
|
+
/* in an ideal situation (all buckets used equally), no bucket would have
|
922
|
+
* more than ceil(#items/#buckets) items. that's the ideal chain length. */
|
923
|
+
unsigned ideal_chain_maxlen;
|
924
|
+
|
925
|
+
/* nonideal_items is the number of items in the hash whose chain position
|
926
|
+
* exceeds the ideal chain maxlen. these items pay the penalty for an uneven
|
927
|
+
* hash distribution; reaching them in a chain traversal takes >ideal steps */
|
928
|
+
unsigned nonideal_items;
|
929
|
+
|
930
|
+
/* ineffective expands occur when a bucket doubling was performed, but
|
931
|
+
* afterward, more than half the items in the hash had nonideal chain
|
932
|
+
* positions. If this happens on two consecutive expansions we inhibit any
|
933
|
+
* further expansion, as it's not helping; this happens when the hash
|
934
|
+
* function isn't a good fit for the key domain. When expansion is inhibited
|
935
|
+
* the hash will still work, albeit no longer in constant time. */
|
936
|
+
unsigned ineff_expands, noexpand;
|
937
|
+
|
938
|
+
uint32_t signature; /* used only to find hash tables in external analysis */
|
939
|
+
#ifdef HASH_BLOOM
|
940
|
+
uint32_t bloom_sig; /* used only to test bloom exists in external analysis */
|
941
|
+
uint8_t *bloom_bv;
|
942
|
+
char bloom_nbits;
|
943
|
+
#endif
|
944
|
+
|
945
|
+
} UT_hash_table;
|
946
|
+
|
947
|
+
typedef struct UT_hash_handle {
|
948
|
+
struct UT_hash_table *tbl;
|
949
|
+
void *prev; /* prev element in app order */
|
950
|
+
void *next; /* next element in app order */
|
951
|
+
struct UT_hash_handle *hh_prev; /* previous hh in bucket order */
|
952
|
+
struct UT_hash_handle *hh_next; /* next hh in bucket order */
|
953
|
+
void *key; /* ptr to enclosing struct's key */
|
954
|
+
unsigned keylen; /* enclosing struct's key len */
|
955
|
+
unsigned hashv; /* result of hash-fcn(key) */
|
956
|
+
} UT_hash_handle;
|
957
|
+
|
958
|
+
#endif /* UTHASH_H */
|