1 | #include <string.h> |
2 | |
3 | #include "wren_utils.h" |
4 | #include "wren_vm.h" |
5 | |
6 | DEFINE_BUFFER(Byte, uint8_t); |
7 | DEFINE_BUFFER(Int, int); |
8 | DEFINE_BUFFER(String, ObjString*); |
9 | |
10 | void wrenSymbolTableInit(SymbolTable* symbols) |
11 | { |
12 | wrenStringBufferInit(symbols); |
13 | } |
14 | |
15 | void wrenSymbolTableClear(WrenVM* vm, SymbolTable* symbols) |
16 | { |
17 | wrenStringBufferClear(vm, symbols); |
18 | } |
19 | |
20 | int wrenSymbolTableAdd(WrenVM* vm, SymbolTable* symbols, |
21 | const char* name, size_t length) |
22 | { |
23 | ObjString* symbol = AS_STRING(wrenNewStringLength(vm, name, length)); |
24 | |
25 | wrenPushRoot(vm, &symbol->obj); |
26 | wrenStringBufferWrite(vm, symbols, symbol); |
27 | wrenPopRoot(vm); |
28 | |
29 | return symbols->count - 1; |
30 | } |
31 | |
32 | int wrenSymbolTableEnsure(WrenVM* vm, SymbolTable* symbols, |
33 | const char* name, size_t length) |
34 | { |
35 | // See if the symbol is already defined. |
36 | int existing = wrenSymbolTableFind(symbols, name, length); |
37 | if (existing != -1) return existing; |
38 | |
39 | // New symbol, so add it. |
40 | return wrenSymbolTableAdd(vm, symbols, name, length); |
41 | } |
42 | |
43 | int wrenSymbolTableFind(const SymbolTable* symbols, |
44 | const char* name, size_t length) |
45 | { |
46 | // See if the symbol is already defined. |
47 | // TODO: O(n). Do something better. |
48 | for (int i = 0; i < symbols->count; i++) |
49 | { |
50 | if (wrenStringEqualsCString(symbols->data[i], name, length)) return i; |
51 | } |
52 | |
53 | return -1; |
54 | } |
55 | |
56 | void wrenBlackenSymbolTable(WrenVM* vm, SymbolTable* symbolTable) |
57 | { |
58 | for (int i = 0; i < symbolTable->count; i++) |
59 | { |
60 | wrenGrayObj(vm, &symbolTable->data[i]->obj); |
61 | } |
62 | |
63 | // Keep track of how much memory is still in use. |
64 | vm->bytesAllocated += symbolTable->capacity * sizeof(*symbolTable->data); |
65 | } |
66 | |
67 | int wrenUtf8EncodeNumBytes(int value) |
68 | { |
69 | ASSERT(value >= 0, "Cannot encode a negative value." ); |
70 | |
71 | if (value <= 0x7f) return 1; |
72 | if (value <= 0x7ff) return 2; |
73 | if (value <= 0xffff) return 3; |
74 | if (value <= 0x10ffff) return 4; |
75 | return 0; |
76 | } |
77 | |
78 | int wrenUtf8Encode(int value, uint8_t* bytes) |
79 | { |
80 | if (value <= 0x7f) |
81 | { |
82 | // Single byte (i.e. fits in ASCII). |
83 | *bytes = value & 0x7f; |
84 | return 1; |
85 | } |
86 | else if (value <= 0x7ff) |
87 | { |
88 | // Two byte sequence: 110xxxxx 10xxxxxx. |
89 | *bytes = 0xc0 | ((value & 0x7c0) >> 6); |
90 | bytes++; |
91 | *bytes = 0x80 | (value & 0x3f); |
92 | return 2; |
93 | } |
94 | else if (value <= 0xffff) |
95 | { |
96 | // Three byte sequence: 1110xxxx 10xxxxxx 10xxxxxx. |
97 | *bytes = 0xe0 | ((value & 0xf000) >> 12); |
98 | bytes++; |
99 | *bytes = 0x80 | ((value & 0xfc0) >> 6); |
100 | bytes++; |
101 | *bytes = 0x80 | (value & 0x3f); |
102 | return 3; |
103 | } |
104 | else if (value <= 0x10ffff) |
105 | { |
106 | // Four byte sequence: 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx. |
107 | *bytes = 0xf0 | ((value & 0x1c0000) >> 18); |
108 | bytes++; |
109 | *bytes = 0x80 | ((value & 0x3f000) >> 12); |
110 | bytes++; |
111 | *bytes = 0x80 | ((value & 0xfc0) >> 6); |
112 | bytes++; |
113 | *bytes = 0x80 | (value & 0x3f); |
114 | return 4; |
115 | } |
116 | |
117 | // Invalid Unicode value. See: http://tools.ietf.org/html/rfc3629 |
118 | UNREACHABLE(); |
119 | return 0; |
120 | } |
121 | |
122 | int wrenUtf8Decode(const uint8_t* bytes, uint32_t length) |
123 | { |
124 | // Single byte (i.e. fits in ASCII). |
125 | if (*bytes <= 0x7f) return *bytes; |
126 | |
127 | int value; |
128 | uint32_t remainingBytes; |
129 | if ((*bytes & 0xe0) == 0xc0) |
130 | { |
131 | // Two byte sequence: 110xxxxx 10xxxxxx. |
132 | value = *bytes & 0x1f; |
133 | remainingBytes = 1; |
134 | } |
135 | else if ((*bytes & 0xf0) == 0xe0) |
136 | { |
137 | // Three byte sequence: 1110xxxx 10xxxxxx 10xxxxxx. |
138 | value = *bytes & 0x0f; |
139 | remainingBytes = 2; |
140 | } |
141 | else if ((*bytes & 0xf8) == 0xf0) |
142 | { |
143 | // Four byte sequence: 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx. |
144 | value = *bytes & 0x07; |
145 | remainingBytes = 3; |
146 | } |
147 | else |
148 | { |
149 | // Invalid UTF-8 sequence. |
150 | return -1; |
151 | } |
152 | |
153 | // Don't read past the end of the buffer on truncated UTF-8. |
154 | if (remainingBytes > length - 1) return -1; |
155 | |
156 | while (remainingBytes > 0) |
157 | { |
158 | bytes++; |
159 | remainingBytes--; |
160 | |
161 | // Remaining bytes must be of form 10xxxxxx. |
162 | if ((*bytes & 0xc0) != 0x80) return -1; |
163 | |
164 | value = value << 6 | (*bytes & 0x3f); |
165 | } |
166 | |
167 | return value; |
168 | } |
169 | |
170 | int wrenUtf8DecodeNumBytes(uint8_t byte) |
171 | { |
172 | // If the byte starts with 10xxxxx, it's the middle of a UTF-8 sequence, so |
173 | // don't count it at all. |
174 | if ((byte & 0xc0) == 0x80) return 0; |
175 | |
176 | // The first byte's high bits tell us how many bytes are in the UTF-8 |
177 | // sequence. |
178 | if ((byte & 0xf8) == 0xf0) return 4; |
179 | if ((byte & 0xf0) == 0xe0) return 3; |
180 | if ((byte & 0xe0) == 0xc0) return 2; |
181 | return 1; |
182 | } |
183 | |
184 | // From: http://graphics.stanford.edu/~seander/bithacks.html#RoundUpPowerOf2Float |
185 | int wrenPowerOf2Ceil(int n) |
186 | { |
187 | n--; |
188 | n |= n >> 1; |
189 | n |= n >> 2; |
190 | n |= n >> 4; |
191 | n |= n >> 8; |
192 | n |= n >> 16; |
193 | n++; |
194 | |
195 | return n; |
196 | } |
197 | |
198 | uint32_t wrenValidateIndex(uint32_t count, int64_t value) |
199 | { |
200 | // Negative indices count from the end. |
201 | if (value < 0) value = count + value; |
202 | |
203 | // Check bounds. |
204 | if (value >= 0 && value < count) return (uint32_t)value; |
205 | |
206 | return UINT32_MAX; |
207 | } |
208 | |