/home/andy/git/oilshell/oil/mycpp/gc_builtins.cc
Line | Count | Source (jump to first uncovered line) |
1 | | #include <ctype.h> // isspace() |
2 | | #include <errno.h> // errno |
3 | | |
4 | | #include "_build/detected-cpp-config.h" |
5 | | |
6 | | #ifdef HAVE_READLINE |
7 | | #include <readline/readline.h> |
8 | | #endif |
9 | | |
10 | | #include "mycpp/runtime.h" |
11 | | |
12 | | // forward decl |
13 | | namespace py_readline { |
14 | | Str* readline(Str*); |
15 | | } |
16 | | |
17 | | // Translation of Python's print(). |
18 | 118 | void print(Str* s) { |
19 | 118 | fputs(s->data_, stdout); // print until first NUL |
20 | 118 | fputc('\n', stdout); |
21 | 118 | } |
22 | | |
23 | 23 | Str* str(int i) { |
24 | 23 | Str* s = OverAllocatedStr(kIntBufSize); |
25 | 23 | int length = snprintf(s->data(), kIntBufSize, "%d", i); |
26 | 23 | s->MaybeShrink(length); |
27 | 23 | return s; |
28 | 23 | } |
29 | | |
30 | | // Print quoted string. TODO: use C-style strings (YSTR) |
31 | 56 | Str* repr(Str* s) { |
32 | | // Worst case: \0 becomes 4 bytes as '\\x00', and then two quote bytes. |
33 | 56 | int n = len(s); |
34 | 56 | int upper_bound = n * 4 + 2; |
35 | | |
36 | 56 | Str* result = OverAllocatedStr(upper_bound); |
37 | | |
38 | | // Single quote by default. |
39 | 56 | char quote = '\''; |
40 | 56 | if (memchr(s->data_, '\'', n) && !memchr(s->data_, '"', n)) { |
41 | 10 | quote = '"'; |
42 | 10 | } |
43 | 56 | char* p = result->data_; |
44 | | |
45 | | // From PyString_Repr() |
46 | 56 | *p++ = quote; |
47 | 474 | for (int i = 0; i < n; ++i) { |
48 | 418 | char c = s->data_[i]; |
49 | 418 | if (c == quote || c == '\\') { |
50 | 0 | *p++ = '\\'; |
51 | 0 | *p++ = c; |
52 | 418 | } else if (c == '\t') { |
53 | 7 | *p++ = '\\'; |
54 | 7 | *p++ = 't'; |
55 | 411 | } else if (c == '\n') { |
56 | 14 | *p++ = '\\'; |
57 | 14 | *p++ = 'n'; |
58 | 397 | } else if (c == '\r') { |
59 | 7 | *p++ = '\\'; |
60 | 7 | *p++ = 'r'; |
61 | 390 | } else if (isprint(c)) { |
62 | 372 | *p++ = c; |
63 | 372 | } else { // Unprintable is \xff |
64 | 18 | sprintf(p, "\\x%02x", c & 0xff); |
65 | 18 | p += 4; |
66 | 18 | } |
67 | 418 | } |
68 | 56 | *p++ = quote; |
69 | 56 | *p = '\0'; |
70 | | |
71 | 56 | int length = p - result->data_; |
72 | 56 | result->MaybeShrink(length); |
73 | 56 | return result; |
74 | 56 | } |
75 | | |
76 | | // Helper for str_to_int() that doesn't use exceptions. |
77 | 84 | bool StringToInteger(const char* s, int length, int base, int* result) { |
78 | 84 | if (length == 0) { |
79 | 2 | return false; // empty string isn't a valid integer |
80 | 2 | } |
81 | | |
82 | 82 | char* pos; // mutated by strtol |
83 | 82 | long v = strtol(s, &pos, base); |
84 | | |
85 | 82 | switch (v) { |
86 | 2 | case LONG_MIN: |
87 | 2 | return false; // underflow |
88 | 2 | case LONG_MAX: |
89 | 2 | return false; // overflow |
90 | 82 | } |
91 | | |
92 | 78 | const char* end = s + length; |
93 | 78 | if (pos == end) { |
94 | 69 | *result = v; |
95 | 69 | return true; // strtol() consumed ALL characters. |
96 | 69 | } |
97 | | |
98 | 13 | while (pos < end) { |
99 | 11 | if (!isspace(*pos)) { |
100 | 7 | return false; // Trailing non-space |
101 | 7 | } |
102 | 4 | pos++; |
103 | 4 | } |
104 | | |
105 | 2 | *result = v; |
106 | 2 | return true; // Trailing space is OK |
107 | 9 | } |
108 | | |
109 | 8 | int to_int(Str* s, int base) { |
110 | 8 | int i; |
111 | 8 | if (StringToInteger(s->data_, len(s), base, &i)) { |
112 | 8 | return i; |
113 | 8 | } else { |
114 | 0 | throw Alloc<ValueError>(); |
115 | 0 | } |
116 | 8 | } |
117 | | |
118 | 31 | int to_int(Str* s) { |
119 | 31 | int i; |
120 | 31 | if (StringToInteger(s->data_, len(s), 10, &i)) { |
121 | 28 | return i; |
122 | 28 | } else { |
123 | 3 | throw Alloc<ValueError>(); |
124 | 3 | } |
125 | 31 | } |
126 | | |
127 | 832 | Str* chr(int i) { |
128 | | // NOTE: i should be less than 256, in which we could return an object from |
129 | | // GLOBAL_STR() pool, like StrIter |
130 | 832 | auto result = NewStr(1); |
131 | 832 | result->data_[0] = i; |
132 | 832 | return result; |
133 | 832 | } |
134 | | |
135 | 836 | int ord(Str* s) { |
136 | 836 | assert(len(s) == 1); |
137 | | // signed to unsigned conversion, so we don't get values like -127 |
138 | 0 | uint8_t c = static_cast<uint8_t>(s->data_[0]); |
139 | 836 | return c; |
140 | 836 | } |
141 | | |
142 | 4 | bool to_bool(Str* s) { |
143 | 4 | return len(s) != 0; |
144 | 4 | } |
145 | | |
146 | 22 | double to_float(Str* s) { |
147 | 22 | char* begin = s->data_; |
148 | 22 | char* end = begin + len(s); |
149 | | |
150 | 22 | errno = 0; |
151 | 22 | double result = strtod(begin, &end); |
152 | | |
153 | 22 | if (errno == ERANGE) { // error: overflow or underflow |
154 | | // log("OVERFLOW or UNDERFLOW %s", s->data_); |
155 | | // log("result %f", result); |
156 | 4 | throw Alloc<ValueError>(); |
157 | 4 | } |
158 | 18 | if (end == begin) { // error: not a floating point number |
159 | 4 | throw Alloc<ValueError>(); |
160 | 4 | } |
161 | | |
162 | 14 | return result; |
163 | 18 | } |
164 | | |
165 | | // e.g. ('a' in 'abc') |
166 | 84 | bool str_contains(Str* haystack, Str* needle) { |
167 | | // Common case |
168 | 84 | if (len(needle) == 1) { |
169 | 72 | return memchr(haystack->data_, needle->data_[0], len(haystack)); |
170 | 72 | } |
171 | | |
172 | 12 | if (len(needle) > len(haystack)) { |
173 | 2 | return false; |
174 | 2 | } |
175 | | |
176 | | // General case. TODO: We could use a smarter substring algorithm. |
177 | | |
178 | 10 | const char* end = haystack->data_ + len(haystack); |
179 | 10 | const char* last_possible = end - len(needle); |
180 | 10 | const char* p = haystack->data_; |
181 | | |
182 | 22 | while (p <= last_possible) { |
183 | 20 | if (memcmp(p, needle->data_, len(needle)) == 0) { |
184 | 8 | return true; |
185 | 8 | } |
186 | 12 | p++; |
187 | 12 | } |
188 | 2 | return false; |
189 | 10 | } |
190 | | |
191 | 52 | Str* str_repeat(Str* s, int times) { |
192 | | // Python allows -1 too, and Oil used that |
193 | 52 | if (times <= 0) { |
194 | 18 | return kEmptyString; |
195 | 18 | } |
196 | 34 | int len_ = len(s); |
197 | 34 | int new_len = len_ * times; |
198 | 34 | Str* result = NewStr(new_len); |
199 | | |
200 | 34 | char* dest = result->data_; |
201 | 527 | for (int i = 0; i < times; i++) { |
202 | 493 | memcpy(dest, s->data_, len_); |
203 | 493 | dest += len_; |
204 | 493 | } |
205 | 34 | return result; |
206 | 52 | } |
207 | | |
208 | | // for os_path.join() |
209 | | // NOTE(Jesse): Perfect candidate for BoundedBuffer |
210 | 20 | Str* str_concat3(Str* a, Str* b, Str* c) { |
211 | 20 | int a_len = len(a); |
212 | 20 | int b_len = len(b); |
213 | 20 | int c_len = len(c); |
214 | | |
215 | 20 | int new_len = a_len + b_len + c_len; |
216 | 20 | Str* result = NewStr(new_len); |
217 | 20 | char* pos = result->data_; |
218 | | |
219 | 20 | memcpy(pos, a->data_, a_len); |
220 | 20 | pos += a_len; |
221 | | |
222 | 20 | memcpy(pos, b->data_, b_len); |
223 | 20 | pos += b_len; |
224 | | |
225 | 20 | memcpy(pos, c->data_, c_len); |
226 | | |
227 | 20 | assert(pos + c_len == result->data_ + new_len); |
228 | | |
229 | 0 | return result; |
230 | 20 | } |
231 | | |
232 | 71 | Str* str_concat(Str* a, Str* b) { |
233 | 71 | int a_len = len(a); |
234 | 71 | int b_len = len(b); |
235 | 71 | int new_len = a_len + b_len; |
236 | 71 | Str* result = NewStr(new_len); |
237 | 71 | char* buf = result->data_; |
238 | | |
239 | 71 | memcpy(buf, a->data_, a_len); |
240 | 71 | memcpy(buf + a_len, b->data_, b_len); |
241 | | |
242 | 71 | return result; |
243 | 71 | } |
244 | | |
245 | | // |
246 | | // Comparators |
247 | | // |
248 | | |
249 | 4.58k | bool str_equals(Str* left, Str* right) { |
250 | | // Fast path for identical strings. String deduplication during GC could |
251 | | // make this more likely. String interning could guarantee it, allowing us |
252 | | // to remove memcmp(). |
253 | 4.58k | if (left == right) { |
254 | 157 | return true; |
255 | 157 | } |
256 | | |
257 | 4.42k | if (left == nullptr || right == nullptr) { |
258 | 0 | return false; |
259 | 0 | } |
260 | | |
261 | | // obj_len equal implies string lengths are equal |
262 | | |
263 | 4.42k | if (left->len_ == right->len_) { |
264 | | // assert(len(left) == len(right)); |
265 | 806 | return memcmp(left->data_, right->data_, left->len_) == 0; |
266 | 806 | } |
267 | | |
268 | 3.62k | return false; |
269 | 4.42k | } |
270 | | |
271 | 10 | bool maybe_str_equals(Str* left, Str* right) { |
272 | 10 | if (left && right) { |
273 | 4 | return str_equals(left, right); |
274 | 4 | } |
275 | | |
276 | 6 | if (!left && !right) { |
277 | 2 | return true; // None == None |
278 | 2 | } |
279 | | |
280 | 4 | return false; // one is None and one is a Str* |
281 | 6 | } |
282 | | |
283 | | // TODO(Jesse): Make an inline version of this |
284 | 3.99k | bool are_equal(Str* left, Str* right) { |
285 | 3.99k | return str_equals(left, right); |
286 | 3.99k | } |
287 | | |
288 | | // TODO(Jesse): Make an inline version of this |
289 | 35 | bool are_equal(int left, int right) { |
290 | 35 | return left == right; |
291 | 35 | } |
292 | | |
293 | | // TODO(Jesse): Make an inline version of this |
294 | 356 | bool keys_equal(int left, int right) { |
295 | 356 | return left == right; |
296 | 356 | } |
297 | | |
298 | | // TODO(Jesse): Make an inline version of this |
299 | 3.89k | bool keys_equal(Str* left, Str* right) { |
300 | 3.89k | return are_equal(left, right); |
301 | 3.89k | } |
302 | | |
303 | 12 | bool are_equal(Tuple2<Str*, int>* t1, Tuple2<Str*, int>* t2) { |
304 | 12 | bool result = are_equal(t1->at0(), t2->at0()); |
305 | 12 | result = result && (t1->at1() == t2->at1()); |
306 | 12 | return result; |
307 | 12 | } |
308 | | |
309 | 8 | bool are_equal(Tuple2<int, int>* t1, Tuple2<int, int>* t2) { |
310 | 8 | return t1->at0() == t2->at0() && t1->at1() == t2->at1(); |
311 | 8 | } |
312 | | |
313 | 8 | bool keys_equal(Tuple2<int, int>* t1, Tuple2<int, int>* t2) { |
314 | 8 | return are_equal(t1, t2); |
315 | 8 | } |
316 | | |
317 | 8 | bool keys_equal(Tuple2<Str*, int>* t1, Tuple2<Str*, int>* t2) { |
318 | 8 | return are_equal(t1, t2); |
319 | 8 | } |
320 | | |
321 | 237 | bool str_equals0(const char* c_string, Str* s) { |
322 | 237 | int n = strlen(c_string); |
323 | 237 | if (len(s) == n) { |
324 | 148 | return memcmp(s->data_, c_string, n) == 0; |
325 | 148 | } else { |
326 | 89 | return false; |
327 | 89 | } |
328 | 237 | } |
329 | | |
330 | 4 | int hash(Str* s) { |
331 | | // FNV-1 from http://www.isthe.com/chongo/tech/comp/fnv/#FNV-1 |
332 | 4 | int h = 2166136261; // 32-bit FNV-1 offset basis |
333 | 4 | constexpr int p = 16777619; // 32-bit FNV-1 prime |
334 | 40 | for (int i = 0; i < len(s); i++) { |
335 | 36 | h *= s->data()[i]; |
336 | 36 | h ^= p; |
337 | 36 | } |
338 | 4 | return h; |
339 | 4 | } |
340 | | |
341 | 8 | int max(int a, int b) { |
342 | 8 | return std::max(a, b); |
343 | 8 | } |
344 | | |
345 | 2 | int max(List<int>* elems) { |
346 | 2 | int n = len(elems); |
347 | 2 | if (n < 1) { |
348 | 0 | throw Alloc<ValueError>(); |
349 | 0 | } |
350 | | |
351 | 2 | int ret = elems->index_(0); |
352 | 10 | for (int i = 0; i < n; ++i) { |
353 | 8 | int cand = elems->index_(i); |
354 | 8 | if (cand > ret) { |
355 | 2 | ret = cand; |
356 | 2 | } |
357 | 8 | } |
358 | | |
359 | 2 | return ret; |
360 | 2 | } |
361 | | |
362 | 0 | Str* raw_input(Str* prompt) { |
363 | 0 | #ifdef HAVE_READLINE |
364 | 0 | Str* ret = py_readline::readline(prompt); |
365 | 0 | if (ret == nullptr) { |
366 | 0 | throw Alloc<EOFError>(); |
367 | 0 | } |
368 | 0 | return ret; |
369 | | #else |
370 | | assert(0); // not implemented |
371 | | #endif |
372 | 0 | } |