MIRROR: javascript for 🐜's, a tiny runtime with big ambitions
add proper url parsing for hashables
+97
-1
+97
-1
src/modules/url.c
+97
-1
src/modules/url.c
···
182
182
return strndup(r->first, (size_t)(r->afterLast - r->first));
183
183
}
184
184
185
+
static char *url_escape_brackets_in_query_or_fragment(const char *url_str, bool *changed_out) {
186
+
size_t len = strlen(url_str);
187
+
size_t extra = 0;
188
+
bool in_query = false;
189
+
bool in_fragment = false;
190
+
191
+
for (size_t i = 0; i < len; i++) {
192
+
char c = url_str[i];
193
+
if (c == '#' && !in_fragment) {
194
+
in_query = false;
195
+
in_fragment = true;
196
+
continue;
197
+
}
198
+
if (c == '?' && !in_query && !in_fragment) {
199
+
in_query = true;
200
+
continue;
201
+
}
202
+
if ((in_query || in_fragment) && (c == '[' || c == ']')) extra += 2;
203
+
}
204
+
205
+
if (changed_out) *changed_out = (extra != 0);
206
+
if (extra == 0) return NULL;
207
+
208
+
char *escaped = malloc(len + extra + 1);
209
+
size_t pos = 0;
210
+
in_query = false;
211
+
in_fragment = false;
212
+
213
+
if (!escaped) return NULL;
214
+
215
+
for (size_t i = 0; i < len; i++) {
216
+
char c = url_str[i];
217
+
if (c == '#' && !in_fragment) {
218
+
in_query = false;
219
+
in_fragment = true;
220
+
escaped[pos++] = c;
221
+
continue;
222
+
}
223
+
if (c == '?' && !in_query && !in_fragment) {
224
+
in_query = true;
225
+
escaped[pos++] = c;
226
+
continue;
227
+
}
228
+
if ((in_query || in_fragment) && c == '[') {
229
+
memcpy(escaped + pos, "%5B", 3);
230
+
pos += 3;
231
+
continue;
232
+
}
233
+
if ((in_query || in_fragment) && c == ']') {
234
+
memcpy(escaped + pos, "%5D", 3);
235
+
pos += 3;
236
+
continue;
237
+
}
238
+
escaped[pos++] = c;
239
+
}
240
+
241
+
escaped[pos] = '\0';
242
+
return escaped;
243
+
}
244
+
245
+
static void url_override_search_hash_from_input(url_state_t *s, const char *url_str) {
246
+
const char *hash = strchr(url_str, '#');
247
+
const char *query = strchr(url_str, '?');
248
+
size_t search_len = 0;
249
+
size_t hash_len = 0;
250
+
251
+
if (query && hash && hash < query) query = NULL;
252
+
253
+
if (query) {
254
+
const char *search_end = hash && hash > query ? hash : url_str + strlen(url_str);
255
+
search_len = (size_t)(search_end - query);
256
+
}
257
+
if (hash) hash_len = strlen(hash);
258
+
259
+
free(s->search);
260
+
s->search = search_len > 0 ? strndup(query, search_len) : strdup("");
261
+
262
+
free(s->hash);
263
+
s->hash = hash_len > 0 ? strndup(hash, hash_len) : strdup("");
264
+
}
265
+
185
266
static void uri_to_state(const UriUriA *uri, url_state_t *s) {
186
267
char *scheme = uri_range_dup(&uri->scheme);
187
268
size_t slen = strlen(scheme);
···
307
388
}
308
389
309
390
UriUriA uri;
310
-
if (uriParseSingleUriA(&uri, url_str, &errpos) != URI_SUCCESS) return -1;
391
+
char *escaped_url = NULL;
392
+
bool used_relaxed_query_parse = false;
393
+
394
+
if (uriParseSingleUriA(&uri, url_str, &errpos) != URI_SUCCESS) {
395
+
escaped_url = url_escape_brackets_in_query_or_fragment(url_str, &used_relaxed_query_parse);
396
+
if (!escaped_url) return -1;
397
+
if (uriParseSingleUriA(&uri, escaped_url, &errpos) != URI_SUCCESS) {
398
+
free(escaped_url);
399
+
return -1;
400
+
}
401
+
}
402
+
311
403
if (!uri.scheme.first || uri.scheme.first == uri.scheme.afterLast) {
312
404
uriFreeUriMembersA(&uri);
405
+
free(escaped_url);
313
406
return -1;
314
407
}
315
408
316
409
uriNormalizeSyntaxA(&uri);
317
410
uri_to_state(&uri, s);
411
+
if (used_relaxed_query_parse) url_override_search_hash_from_input(s, url_str);
412
+
318
413
uriFreeUriMembersA(&uri);
414
+
free(escaped_url);
319
415
320
416
return 0;
321
417
}