xref: /lighttpd1.4/src/buffer.c (revision d958cf32)
1 #include "first.h"
2 
3 #include "buffer.h"
4 
5 #include <stdlib.h>
6 #include <string.h>
7 #include "sys-time.h"   /* strftime() */
8 
9 static const char hex_chars_lc[] = "0123456789abcdef";
10 static const char hex_chars_uc[] = "0123456789ABCDEF";
11 
12 /**
13  * init the buffer
14  *
15  */
16 
17 __attribute_noinline__
18 buffer* buffer_init(void) {
19 	buffer * const b = calloc(1, sizeof(*b));
20 	force_assert(b);
21 	return b;
22 }
23 
24 buffer *buffer_init_buffer(const buffer *src) {
25 	buffer * const b = buffer_init();
26 	buffer_copy_string_len(b, BUF_PTR_LEN(src));
27 	return b;
28 }
29 
30 buffer *buffer_init_string(const char *str) {
31 	buffer * const b = buffer_init();
32 	buffer_copy_string(b, str);
33 	return b;
34 }
35 
36 void buffer_free(buffer *b) {
37 	if (NULL == b) return;
38 
39 	free(b->ptr);
40 	free(b);
41 }
42 
43 void buffer_free_ptr(buffer *b) {
44 	free(b->ptr);
45 	b->ptr = NULL;
46 	b->used = 0;
47 	b->size = 0;
48 }
49 
50 void buffer_move(buffer * restrict b, buffer * restrict src) {
51 	buffer tmp;
52 	buffer_clear(b);
53 	tmp = *src; *src = *b; *b = tmp;
54 }
55 
56 /* make sure buffer is at least "size" big + 1 for '\0'. keep old data */
57 __attribute_cold__
58 __attribute_noinline__
59 __attribute_nonnull__
60 __attribute_returns_nonnull__
61 static char* buffer_realloc(buffer * const restrict b, const size_t len) {
62     #define BUFFER_PIECE_SIZE 64uL  /*(must be power-of-2)*/
63     size_t sz = (len + 1 + BUFFER_PIECE_SIZE-1) & ~(BUFFER_PIECE_SIZE-1);
64     force_assert(sz > len);
65     if ((sz & (sz-1)) && sz < INT_MAX) {/* not power-2; huge val not expected */
66         /*(optimizer should recognize this and use ffs or clz or equivalent)*/
67         const size_t psz = sz;
68         for (sz = 256; sz < psz; sz <<= 1) ;
69     }
70 
71     b->size = sz;
72     b->ptr = realloc(b->ptr, sz);
73 
74     force_assert(NULL != b->ptr);
75     return b->ptr;
76 }
77 
78 __attribute_cold__
79 __attribute_noinline__
80 __attribute_nonnull__
81 __attribute_returns_nonnull__
82 static char* buffer_alloc_replace(buffer * const restrict b, const size_t size) {
83     /*(discard old data so realloc() does not copy)*/
84     if (NULL != b->ptr) {
85         free(b->ptr);
86         b->ptr = NULL;
87     }
88     /*(note: if size larger than one lshift, use size instead of power-2)*/
89     return buffer_realloc(b, (b->size << 1) > size ? (b->size << 1)-1 : size);
90 }
91 
92 char* buffer_string_prepare_copy(buffer * const b, const size_t size) {
93     b->used = 0;
94     return (size < b->size)
95       ? b->ptr
96       : buffer_alloc_replace(b, size);
97 }
98 
99 __attribute_cold__
100 __attribute_noinline__
101 __attribute_nonnull__
102 __attribute_returns_nonnull__
103 static char* buffer_string_prepare_append_resize(buffer * const restrict b, const size_t size) {
104     if (b->used < 2) {  /* buffer_is_blank(b) */
105         char * const s = buffer_string_prepare_copy(b, size);
106         *s = '\0'; /*(for case (1 == b->used))*/
107         return s;
108     }
109 
110     /* not empty, b->used already includes a terminating 0 */
111     /*(note: if size larger than one lshift, use size instead of power-2)*/
112     const size_t req_size = ((b->size << 1) - b->used > size)
113       ? (b->size << 1)-1
114       : b->used + size;
115 
116     /* check for overflow: unsigned overflow is defined to wrap around */
117     force_assert(req_size >= b->used);
118 
119     return buffer_realloc(b, req_size) + b->used - 1;
120 }
121 
122 char* buffer_string_prepare_append(buffer * const b, const size_t size) {
123     const uint32_t len = b->used ? b->used-1 : 0;
124     return (b->size - len >= size + 1)
125       ? b->ptr + len
126       : buffer_string_prepare_append_resize(b, size);
127 }
128 
129 /*(prefer smaller code than inlining buffer_extend in many places in buffer.c)*/
130 __attribute_noinline__
131 char*
132 buffer_extend (buffer * const b, const size_t x)
133 {
134     /* extend buffer to append x (reallocate by power-2 (or larger), if needed)
135      * (combine buffer_string_prepare_append() and buffer_commit())
136      * (future: might make buffer.h static inline func for HTTP/1.1 performance)
137      * pre-sets '\0' byte and b->used (unlike buffer_string_prepare_append())*/
138   #if 0
139     char * const s = buffer_string_prepare_append(b, x);
140     b->used += x + (0 == b->used);
141   #else
142     const uint32_t len = b->used ? b->used-1 : 0;
143     char * const s = (b->size - len >= x + 1)
144       ? b->ptr + len
145       : buffer_string_prepare_append_resize(b, x);
146     b->used = len+x+1;
147   #endif
148     s[x] = '\0';
149     return s;
150 }
151 
152 void buffer_commit(buffer *b, size_t size)
153 {
154 	size_t sz = b->used;
155 	if (0 == sz) sz = 1;
156 
157 	if (size > 0) {
158 		/* check for overflow: unsigned overflow is defined to wrap around */
159 		sz += size;
160 		force_assert(sz > size);
161 	}
162 
163 	b->used = sz;
164 	b->ptr[sz - 1] = '\0';
165 }
166 
167 void buffer_copy_string(buffer * restrict b, const char * restrict s) {
168 	buffer_copy_string_len(b, s, NULL != s ? strlen(s) : 0);
169 }
170 
171 void buffer_copy_string_len(buffer * const restrict b, const char * const restrict s, const size_t len) {
172     b->used = len + 1;
173     char * const restrict d = (len < b->size)
174       ? b->ptr
175       : buffer_alloc_replace(b, len);
176     d[len] = '\0';
177     memcpy(d, s, len);
178 }
179 
180 void buffer_append_string(buffer * restrict b, const char * restrict s) {
181 	buffer_append_string_len(b, s, NULL != s ? strlen(s) : 0);
182 }
183 
184 /**
185  * append a string to the end of the buffer
186  *
187  * the resulting buffer is terminated with a '\0'
188  * s is treated as a un-terminated string (a \0 is handled a normal character)
189  *
190  * @param b a buffer
191  * @param s the string
192  * @param s_len size of the string (without the terminating \0)
193  */
194 
195 void buffer_append_string_len(buffer * const restrict b, const char * const restrict s, const size_t len) {
196     memcpy(buffer_extend(b, len), s, len);
197 }
198 
199 void buffer_append_str2(buffer * const restrict b, const char * const s1, const size_t len1, const char * const s2, const size_t len2) {
200     char * const restrict s = buffer_extend(b, len1+len2);
201   #ifdef HAVE_MEMPCPY
202     mempcpy(mempcpy(s, s1, len1), s2, len2);
203   #else
204     memcpy(s,      s1, len1);
205     memcpy(s+len1, s2, len2);
206   #endif
207 }
208 
209 void buffer_append_str3(buffer * const restrict b, const char * const s1, const size_t len1, const char * const s2, const size_t len2, const char * const s3, const size_t len3) {
210     char * restrict s = buffer_extend(b, len1+len2+len3);
211   #ifdef HAVE_MEMPCPY
212     mempcpy(mempcpy(mempcpy(s, s1, len1), s2, len2), s3, len3);
213   #else
214     memcpy(s,         s1, len1);
215     memcpy((s+=len1), s2, len2);
216     memcpy((s+=len2), s3, len3);
217   #endif
218 }
219 
220 void buffer_append_iovec(buffer * const restrict b, const struct const_iovec * const iov, const size_t n) {
221     size_t len = 0;
222     for (size_t i = 0; i < n; ++i)
223         len += iov[i].iov_len;
224     char *s = buffer_extend(b, len);
225     for (size_t i = 0; i < n; ++i) {
226         if (0 == iov[i].iov_len) continue;
227       #ifdef HAVE_MEMPCPY
228         s = mempcpy(s, iov[i].iov_base, iov[i].iov_len);
229       #else
230         memcpy(s, iov[i].iov_base, iov[i].iov_len);
231         s += iov[i].iov_len;
232       #endif
233     }
234 }
235 
236 void buffer_append_path_len(buffer * restrict b, const char * restrict a, size_t alen) {
237     char * restrict s = buffer_string_prepare_append(b, alen+1);
238     const int aslash = (alen && a[0] == '/');
239     if (b->used > 1 && s[-1] == '/') {
240         if (aslash) {
241             ++a;
242             --alen;
243         }
244     }
245     else {
246         if (0 == b->used) b->used = 1;
247         if (!aslash) {
248             *s++ = '/';
249             ++b->used;
250         }
251     }
252     b->used += alen;
253     s[alen] = '\0';
254     memcpy(s, a, alen);
255 }
256 
257 void
258 buffer_copy_path_len2 (buffer * const restrict b, const char * const restrict s1, size_t len1, const char * const restrict s2, size_t len2)
259 {
260     /*(similar to buffer_copy_string_len(b, s1, len1) but combined allocation)*/
261     memcpy(buffer_string_prepare_copy(b, len1+len2+1), s1, len1);
262     b->used = len1 + 1;                    /*('\0' byte will be written below)*/
263 
264     buffer_append_path_len(b, s2, len2);/*(choice: not inlined, special-cased)*/
265 }
266 
267 void
268 buffer_copy_string_len_lc (buffer * const restrict b, const char * const restrict s, const size_t len)
269 {
270     char * const restrict d = buffer_string_prepare_copy(b, len);
271     b->used = len+1;
272     d[len] = '\0';
273     for (size_t i = 0; i < len; ++i)
274         d[i] = (!light_isupper(s[i])) ? s[i] : s[i] | 0x20;
275 }
276 
277 void buffer_append_uint_hex_lc(buffer *b, uintmax_t value) {
278 	char *buf;
279 	unsigned int shift = 0;
280 
281 	{
282 		uintmax_t copy = value;
283 		do {
284 			copy >>= 8;
285 			shift += 8; /* counting bits */
286 		} while (0 != copy);
287 	}
288 
289 	buf = buffer_extend(b, shift >> 2); /*nibbles (4 bits)*/
290 
291 	while (shift > 0) {
292 		shift -= 4;
293 		*(buf++) = hex_chars_lc[(value >> shift) & 0x0F];
294 	}
295 }
296 
297 __attribute_nonnull__
298 __attribute_returns_nonnull__
299 static char* utostr(char buf[LI_ITOSTRING_LENGTH], uintmax_t val) {
300 	char *cur = buf+LI_ITOSTRING_LENGTH;
301 	do {
302 		int mod = val % 10;
303 		val /= 10;
304 		/* prepend digit mod */
305 		*(--cur) = (char) ('0' + mod);
306 	} while (0 != val);
307 	return cur;
308 }
309 
310 __attribute_nonnull__
311 __attribute_returns_nonnull__
312 static char* itostr(char buf[LI_ITOSTRING_LENGTH], intmax_t val) {
313 	/* absolute value not defined for INTMAX_MIN, but can take absolute
314 	 * value of any negative number via twos complement cast to unsigned.
315 	 * negative sign is prepended after (now unsigned) value is converted
316 	 * to string */
317 	uintmax_t uval = val >= 0 ? (uintmax_t)val : ((uintmax_t)~val) + 1;
318 	char *cur = utostr(buf, uval);
319 	if (val < 0) *(--cur) = '-';
320 
321 	return cur;
322 }
323 
324 void buffer_append_int(buffer *b, intmax_t val) {
325 	char buf[LI_ITOSTRING_LENGTH];
326 	const char * const str = itostr(buf, val);
327 	buffer_append_string_len(b, str, buf+sizeof(buf) - str);
328 }
329 
330 void buffer_append_strftime(buffer * const restrict b, const char * const restrict format, const struct tm * const restrict tm) {
331     force_assert(NULL != format);
332     /*(localtime_r() or gmtime_r() producing tm should not have failed)*/
333     if (__builtin_expect( (NULL == tm), 0)) return;
334 
335     /*(expecting typical format strings to result in < 64 bytes needed;
336      * skipping buffer_string_space() calculation and providing fixed size)*/
337     size_t rv = strftime(buffer_string_prepare_append(b, 63), 64, format, tm);
338 
339     /* 0 (in some apis) signals the string may have been too small;
340      * but the format could also just have lead to an empty string */
341     if (__builtin_expect( (0 == rv), 0) || __builtin_expect( (rv > 63), 0)) {
342         /* unexpected; give it a second try with a larger string */
343         rv = strftime(buffer_string_prepare_append(b, 4095), 4096, format, tm);
344         if (__builtin_expect( (rv > 4095), 0))/*(input format was ridiculous)*/
345             return;
346     }
347 
348     /*buffer_commit(b, rv);*/
349     b->used += (uint32_t)rv + (0 == b->used);
350 }
351 
352 
353 size_t li_itostrn(char *buf, size_t buf_len, intmax_t val) {
354 	char p_buf[LI_ITOSTRING_LENGTH];
355 	char* const str = itostr(p_buf, val);
356 	size_t len = (size_t)(p_buf+sizeof(p_buf)-str);
357 	force_assert(len <= buf_len);
358 	memcpy(buf, str, len);
359 	return len;
360 }
361 
362 size_t li_utostrn(char *buf, size_t buf_len, uintmax_t val) {
363 	char p_buf[LI_ITOSTRING_LENGTH];
364 	char* const str = utostr(p_buf, val);
365 	size_t len = (size_t)(p_buf+sizeof(p_buf)-str);
366 	force_assert(len <= buf_len);
367 	memcpy(buf, str, len);
368 	return len;
369 }
370 
371 #define li_ntox_lc(n) ((n) <= 9 ? (n) + '0' : (n) + 'a' - 10)
372 
373 /* c (char) and n (nibble) MUST be unsigned integer types */
374 #define li_cton(c,n) \
375   (((n) = (c) - '0') <= 9 || (((n) = ((c)&0xdf) - 'A') <= 5 ? ((n) += 10) : 0))
376 
377 /* converts hex char (0-9, A-Z, a-z) to decimal.
378  * returns 0xFF on invalid input.
379  */
380 char hex2int(unsigned char hex) {
381 	unsigned char n;
382 	return li_cton(hex,n) ? (char)n : 0xFF;
383 }
384 
385 int li_hex2bin (unsigned char * const bin, const size_t binlen, const char * const hexstr, const size_t len)
386 {
387     /* validate and transform 32-byte MD5 hex string to 16-byte binary MD5,
388      * or 64-byte SHA-256 or SHA-512-256 hex string to 32-byte binary digest */
389     if (len > (binlen << 1)) return -1;
390     for (int i = 0, ilen = (int)len; i < ilen; i+=2) {
391         int hi = hexstr[i];
392         int lo = hexstr[i+1];
393         if ('0' <= hi && hi <= '9')                    hi -= '0';
394         else if ((uint32_t)(hi |= 0x20)-'a' <= 'f'-'a')hi += -'a' + 10;
395         else                                           return -1;
396         if ('0' <= lo && lo <= '9')                    lo -= '0';
397         else if ((uint32_t)(lo |= 0x20)-'a' <= 'f'-'a')lo += -'a' + 10;
398         else                                           return -1;
399         bin[(i >> 1)] = (unsigned char)((hi << 4) | lo);
400     }
401     return 0;
402 }
403 
404 
405 __attribute_noinline__
406 int buffer_eq_icase_ssn(const char * const a, const char * const b, const size_t len) {
407     for (size_t i = 0; i < len; ++i) {
408         unsigned int ca = ((unsigned char *)a)[i];
409         unsigned int cb = ((unsigned char *)b)[i];
410         if (ca != cb) {
411             ca |= 0x20;
412             cb |= 0x20;
413             if (ca != cb) return 0;
414             if (!light_islower(ca)) return 0;
415             if (!light_islower(cb)) return 0;
416         }
417     }
418     return 1;
419 }
420 
421 int buffer_eq_icase_ss(const char * const a, const size_t alen, const char * const b, const size_t blen) {
422     /* 1 = equal; 0 = not equal */ /* short string sizes expected (< INT_MAX) */
423     return (alen == blen) ? buffer_eq_icase_ssn(a, b, blen) : 0;
424 }
425 
426 int buffer_eq_icase_slen(const buffer * const b, const char * const s, const size_t slen) {
427     /* Note: b must be initialized, i.e. 0 != b->used; uninitialized is not eq*/
428     /* 1 = equal; 0 = not equal */ /* short string sizes expected (< INT_MAX) */
429     return (b->used == slen + 1) ? buffer_eq_icase_ssn(b->ptr, s, slen) : 0;
430 }
431 
432 int buffer_eq_slen(const buffer * const b, const char * const s, const size_t slen) {
433     /* Note: b must be initialized, i.e. 0 != b->used; uninitialized is not eq*/
434     /* 1 = equal; 0 = not equal */ /* short string sizes expected (< INT_MAX) */
435     return (b->used == slen + 1 && 0 == memcmp(b->ptr, s, slen));
436 }
437 
438 
439 /**
440  * check if two buffer contain the same data
441  */
442 
443 int buffer_is_equal(const buffer *a, const buffer *b) {
444 	/* 1 = equal; 0 = not equal */
445 	return (a->used == b->used && 0 == memcmp(a->ptr, b->ptr, a->used));
446 }
447 
448 
449 void li_tohex_lc(char * const restrict buf, size_t buf_len, const char * const restrict s, size_t s_len) {
450 	force_assert(2 * s_len > s_len);
451 	force_assert(2 * s_len < buf_len);
452 
453 	for (size_t i = 0; i < s_len; ++i) {
454 		buf[2*i]   = hex_chars_lc[(s[i] >> 4) & 0x0F];
455 		buf[2*i+1] = hex_chars_lc[s[i] & 0x0F];
456 	}
457 	buf[2*s_len] = '\0';
458 }
459 
460 void li_tohex_uc(char * const restrict buf, size_t buf_len, const char * const restrict s, size_t s_len) {
461 	force_assert(2 * s_len > s_len);
462 	force_assert(2 * s_len < buf_len);
463 
464 	for (size_t i = 0; i < s_len; ++i) {
465 		buf[2*i]   = hex_chars_uc[(s[i] >> 4) & 0x0F];
466 		buf[2*i+1] = hex_chars_uc[s[i] & 0x0F];
467 	}
468 	buf[2*s_len] = '\0';
469 }
470 
471 
472 void buffer_substr_replace (buffer * const restrict b, const size_t offset,
473                             const size_t len, const buffer * const restrict replace)
474 {
475     const size_t blen = buffer_clen(b);
476     const size_t rlen = buffer_clen(replace);
477 
478     if (rlen > len) {
479         buffer_extend(b, blen-len+rlen);
480         memmove(b->ptr+offset+rlen, b->ptr+offset+len, blen-offset-len);
481     }
482 
483     memcpy(b->ptr+offset, replace->ptr, rlen);
484 
485     if (rlen < len) {
486         memmove(b->ptr+offset+rlen, b->ptr+offset+len, blen-offset-len);
487         buffer_truncate(b, blen-len+rlen);
488     }
489 }
490 
491 
492 void buffer_append_string_encoded_hex_lc(buffer * const restrict b, const char * const restrict s, size_t len) {
493     unsigned char * const p = (unsigned char *)buffer_extend(b, len*2);
494     for (size_t i = 0; i < len; ++i) {
495         p[(i<<1)]   = hex_chars_lc[(s[i] >> 4) & 0x0F];
496         p[(i<<1)+1] = hex_chars_lc[(s[i])      & 0x0F];
497     }
498 }
499 
500 void buffer_append_string_encoded_hex_uc(buffer * const restrict b, const char * const restrict s, size_t len) {
501     unsigned char * const p = (unsigned char *)buffer_extend(b, len*2);
502     for (size_t i = 0; i < len; ++i) {
503         p[(i<<1)]   = hex_chars_uc[(s[i] >> 4) & 0x0F];
504         p[(i<<1)+1] = hex_chars_uc[(s[i])      & 0x0F];
505     }
506 }
507 
508 
509 /* everything except: ! ( ) * - . 0-9 A-Z _ a-z */
510 static const char encoded_chars_rel_uri_part[] = {
511 	/*
512 	0  1  2  3  4  5  6  7  8  9  A  B  C  D  E  F
513 	*/
514 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  00 -  0F control chars */
515 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  10 -  1F */
516 	1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 0, 0, 1,  /*  20 -  2F space " # $ % & ' + , / */
517 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1,  /*  30 -  3F : ; < = > ? */
518 	1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  40 -  4F @ */
519 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0,  /*  50 -  5F [ \ ] ^ */
520 	1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  60 -  6F ` */
521 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1,  /*  70 -  7F { | } DEL */
522 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  80 -  8F */
523 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  90 -  9F */
524 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  A0 -  AF */
525 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  B0 -  BF */
526 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  C0 -  CF */
527 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  D0 -  DF */
528 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  E0 -  EF */
529 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  F0 -  FF */
530 };
531 
532 /* everything except: ! ( ) * - . / 0-9 A-Z _ a-z */
533 static const char encoded_chars_rel_uri[] = {
534 	/*
535 	0  1  2  3  4  5  6  7  8  9  A  B  C  D  E  F
536 	*/
537 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  00 -  0F control chars */
538 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  10 -  1F */
539 	1, 0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 0, 0, 0,  /*  20 -  2F space " # $ % & ' + , */
540 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1,  /*  30 -  3F : ; < = > ? */
541 	1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  40 -  4F @ */
542 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0,  /*  50 -  5F [ \ ] ^ */
543 	1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  60 -  6F ` */
544 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1,  /*  70 -  7F { | } DEL */
545 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  80 -  8F */
546 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  90 -  9F */
547 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  A0 -  AF */
548 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  B0 -  BF */
549 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  C0 -  CF */
550 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  D0 -  DF */
551 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  E0 -  EF */
552 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  F0 -  FF */
553 };
554 
555 static const char encoded_chars_html[] = {
556 	/*
557 	0  1  2  3  4  5  6  7  8  9  A  B  C  D  E  F
558 	*/
559 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  00 -  0F control chars */
560 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  10 -  1F */
561 	0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0,  /*  20 -  2F " & ' */
562 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0,  /*  30 -  3F < > */
563 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  40 -  4F */
564 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  50 -  5F */
565 	1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  60 -  6F ` */
566 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,  /*  70 -  7F DEL */
567 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  80 -  8F */
568 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  90 -  9F */
569 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  A0 -  AF */
570 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  B0 -  BF */
571 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  C0 -  CF */
572 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  D0 -  DF */
573 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  E0 -  EF */
574 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  F0 -  FF */
575 };
576 
577 static const char encoded_chars_minimal_xml[] = {
578 	/*
579 	0  1  2  3  4  5  6  7  8  9  A  B  C  D  E  F
580 	*/
581 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  00 -  0F control chars */
582 	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /*  10 -  1F */
583 	0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0,  /*  20 -  2F " & ' */
584 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0,  /*  30 -  3F < > */
585 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  40 -  4F */
586 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  50 -  5F */
587 	1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  60 -  6F ` */
588 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,  /*  70 -  7F DEL */
589 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  80 -  8F */
590 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  90 -  9F */
591 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  A0 -  AF */
592 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  B0 -  BF */
593 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  C0 -  CF */
594 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  D0 -  DF */
595 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  E0 -  EF */
596 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  /*  F0 -  FF */
597 };
598 
599 
600 
601 void buffer_append_string_encoded(buffer * const restrict b, const char * const restrict s, size_t s_len, buffer_encoding_t encoding) {
602 	unsigned char *ds, *d;
603 	size_t d_len, ndx;
604 	const char *map = NULL;
605 
606 	if (0 == s_len) return;
607 
608 	force_assert(NULL != s);
609 
610 	switch(encoding) {
611 	case ENCODING_REL_URI:
612 		map = encoded_chars_rel_uri;
613 		break;
614 	case ENCODING_REL_URI_PART:
615 		map = encoded_chars_rel_uri_part;
616 		break;
617 	case ENCODING_HTML:
618 		map = encoded_chars_html;
619 		break;
620 	case ENCODING_MINIMAL_XML:
621 		map = encoded_chars_minimal_xml;
622 		break;
623 	}
624 
625 	force_assert(NULL != map);
626 
627 	/* count to-be-encoded-characters */
628 	for (ds = (unsigned char *)s, d_len = 0, ndx = 0; ndx < s_len; ds++, ndx++) {
629 		if (map[*ds & 0xFF]) {
630 			switch(encoding) {
631 			case ENCODING_REL_URI:
632 			case ENCODING_REL_URI_PART:
633 				d_len += 3;
634 				break;
635 			case ENCODING_HTML:
636 			case ENCODING_MINIMAL_XML:
637 				d_len += 6;
638 				break;
639 			}
640 		} else {
641 			d_len++;
642 		}
643 	}
644 
645 	d = (unsigned char*) buffer_extend(b, d_len);
646 
647 	if (d_len == s_len) { /*(short-circuit; nothing to encoded)*/
648 		memcpy(d, s, s_len);
649 		return;
650 	}
651 
652 	for (ds = (unsigned char *)s, d_len = 0, ndx = 0; ndx < s_len; ds++, ndx++) {
653 		if (map[*ds & 0xFF]) {
654 			switch(encoding) {
655 			case ENCODING_REL_URI:
656 			case ENCODING_REL_URI_PART:
657 				d[d_len++] = '%';
658 				d[d_len++] = hex_chars_uc[((*ds) >> 4) & 0x0F];
659 				d[d_len++] = hex_chars_uc[(*ds) & 0x0F];
660 				break;
661 			case ENCODING_HTML:
662 			case ENCODING_MINIMAL_XML:
663 				d[d_len++] = '&';
664 				d[d_len++] = '#';
665 				d[d_len++] = 'x';
666 				d[d_len++] = hex_chars_uc[((*ds) >> 4) & 0x0F];
667 				d[d_len++] = hex_chars_uc[(*ds) & 0x0F];
668 				d[d_len++] = ';';
669 				break;
670 			}
671 		} else {
672 			d[d_len++] = *ds;
673 		}
674 	}
675 }
676 
677 void buffer_append_string_c_escaped(buffer * const restrict b, const char * const restrict s, size_t s_len) {
678 	unsigned char *ds, *d;
679 	size_t d_len, ndx;
680 
681 	if (0 == s_len) return;
682 
683 	/* count to-be-encoded-characters */
684 	for (ds = (unsigned char *)s, d_len = 0, ndx = 0; ndx < s_len; ds++, ndx++) {
685 		if ((*ds < 0x20) /* control character */
686 				|| (*ds >= 0x7f)) { /* DEL + non-ASCII characters */
687 			switch (*ds) {
688 			case '\t':
689 			case '\r':
690 			case '\n':
691 				d_len += 2;
692 				break;
693 			default:
694 				d_len += 4; /* \xCC */
695 				break;
696 			}
697 		} else {
698 			d_len++;
699 		}
700 	}
701 
702 	d = (unsigned char*) buffer_extend(b, d_len);
703 
704 	if (d_len == s_len) { /*(short-circuit; nothing to encoded)*/
705 		memcpy(d, s, s_len);
706 		return;
707 	}
708 
709 	for (ds = (unsigned char *)s, d_len = 0, ndx = 0; ndx < s_len; ds++, ndx++) {
710 		if ((*ds < 0x20) /* control character */
711 				|| (*ds >= 0x7f)) { /* DEL + non-ASCII characters */
712 			d[d_len++] = '\\';
713 			switch (*ds) {
714 			case '\t':
715 				d[d_len++] = 't';
716 				break;
717 			case '\r':
718 				d[d_len++] = 'r';
719 				break;
720 			case '\n':
721 				d[d_len++] = 'n';
722 				break;
723 			default:
724 				d[d_len++] = 'x';
725 				d[d_len++] = hex_chars_lc[((*ds) >> 4) & 0x0F];
726 				d[d_len++] = hex_chars_lc[(*ds) & 0x0F];
727 				break;
728 			}
729 		} else {
730 			d[d_len++] = *ds;
731 		}
732 	}
733 }
734 
735 
736 /* decodes url-special-chars inplace.
737  * replaces non-printable characters with '_'
738  * (If this is used on a portion of query string, then query string should be
739  *  split on '&', and '+' replaced with ' ' before calling this routine)
740  */
741 
742 void buffer_urldecode_path(buffer * const b) {
743     const size_t len = buffer_clen(b);
744     char *src = len ? memchr(b->ptr, '%', len) : NULL;
745     if (NULL == src) return;
746 
747     char *dst = src;
748     do {
749         /* *src == '%' */
750         unsigned char high = hex2int(*(src + 1));
751         unsigned char low = hex2int(*(src + 2));
752         if (0xFF != high && 0xFF != low) {
753             high = (high << 4) | low;   /* map ctrls to '_' */
754             *dst = (high >= 32 && high != 127) ? high : '_';
755             src += 2;
756         } /* else ignore this '%'; leave as-is and move on */
757 
758         while ((*++dst = *++src) != '%' && *src) ;
759     } while (*src);
760     b->used = (dst - b->ptr) + 1;
761 }
762 
763 int buffer_is_valid_UTF8(const buffer *b) {
764     /* https://www.w3.org/International/questions/qa-forms-utf-8 */
765     const unsigned char *c = (unsigned char *)b->ptr;
766     while (*c) {
767 
768         /*(note: includes ctrls)*/
769         if (                         c[0] <  0x80 ) { ++c;  continue; }
770 
771         if (         0xc2 <= c[0] && c[0] <= 0xdf
772             &&       0x80 <= c[1] && c[1] <= 0xbf ) { c+=2; continue; }
773 
774         if ( (   (   0xe0 == c[0]
775                   && 0xa0 <= c[1] && c[1] <= 0xbf)
776               || (   0xe1 <= c[0] && c[0] <= 0xef && c[0] != 0xed
777                   && 0x80 <= c[1] && c[1] <= 0xbf)
778               || (   0xed == c[0]
779                   && 0x80 <= c[1] && c[1] <= 0x9f)   )
780             &&       0x80 <= c[2] && c[2] <= 0xbf ) { c+=3; continue; }
781 
782         if ( (   (   0xf0 == c[0]
783                   && 0x90 <= c[1] && c[1] <= 0xbf)
784               || (   0xf1 <= c[0] && c[0] <= 0xf3
785                   && 0x80 <= c[1] && c[1] <= 0xbf)
786               || (   0xf4 == c[0]
787                   && 0x80 <= c[1] && c[1] <= 0x8f)   )
788             &&       0x80 <= c[2] && c[2] <= 0xbf
789             &&       0x80 <= c[3] && c[3] <= 0xbf ) { c+=4; continue; }
790 
791         return 0; /* invalid */
792     }
793     return 1; /* valid */
794 }
795 
796 /* - special case: empty string returns empty string
797  * - on windows or cygwin: replace \ with /
798  * - strip leading spaces
799  * - prepends "/" if not present already
800  * - resolve "/../", "//" and "/./" the usual way:
801  *   the first one removes a preceding component, the other two
802  *   get compressed to "/".
803  * - "/." and "/.." at the end are similar, but always leave a trailing
804  *   "/"
805  *
806  * /blah/..         gets  /
807  * /blah/../foo     gets  /foo
808  * /abc/./xyz       gets  /abc/xyz
809  * /abc//xyz        gets  /abc/xyz
810  */
811 
812 void buffer_path_simplify(buffer *b)
813 {
814     char *out = b->ptr;
815     char * const end = b->ptr + b->used - 1;
816 
817     if (__builtin_expect( (buffer_is_blank(b)), 0)) {
818         buffer_blank(b);
819         return;
820     }
821 
822   #if defined(__WIN32) || defined(__CYGWIN__)
823     /* cygwin is treating \ and / the same, so we have to that too */
824     for (char *p = b->ptr; *p; p++) {
825         if (*p == '\\') *p = '/';
826     }
827   #endif
828 
829     *end = '/'; /*(end of path modified to avoid need to check '\0')*/
830 
831     char *walk = out;
832     if (__builtin_expect( (*walk == '/'), 1)) {
833         /* scan to detect (potential) need for path simplification
834          * (repeated '/' or "/.") */
835         do {
836             if (*++walk == '.' || *walk == '/')
837                 break;
838             do { ++walk; } while (*walk != '/');
839         } while (walk != end);
840         if (__builtin_expect( (walk == end), 1)) {
841             /* common case: no repeated '/' or "/." */
842             *end = '\0'; /* overwrite extra '/' added to end of path */
843             return;
844         }
845         out = walk-1;
846     }
847     else {
848         if (walk[0] == '.' && walk[1] == '/')
849             *out = *++walk;
850         else if (walk[0] == '.' && walk[1] == '.' && walk[2] == '/')
851             *out = *(walk += 2);
852         else {
853             while (*++walk != '/') ;
854             out = walk;
855         }
856         ++walk;
857     }
858 
859     while (walk <= end) {
860         /* previous char is '/' at this point (or start of string w/o '/') */
861         if (__builtin_expect( (walk[0] == '/'), 0)) {
862             /* skip repeated '/' (e.g. "///" -> "/") */
863             if (++walk < end)
864                 continue;
865             else {
866                 ++out;
867                 break;
868             }
869         }
870         else if (__builtin_expect( (walk[0] == '.'), 0)) {
871             /* handle "./" and "../" */
872             if (walk[1] == '.' && walk[2] == '/') {
873                 /* handle "../" */
874                 while (out > b->ptr && *--out != '/') ;
875                 *out = '/'; /*(in case path had not started with '/')*/
876                 if ((walk += 3) >= end) {
877                     ++out;
878                     break;
879                 }
880                 else
881                 continue;
882             }
883             else if (walk[1] == '/') {
884                 /* handle "./" */
885                 if ((walk += 2) >= end) {
886                     ++out;
887                     break;
888                 }
889                 continue;
890             }
891             else {
892                 /* accept "." if not part of "../" or "./" */
893                 *++out = '.';
894                 ++walk;
895             }
896         }
897 
898         while ((*++out = *walk++) != '/') ;
899     }
900     *out = *end = '\0'; /* overwrite extra '/' added to end of path */
901     b->used = (out - b->ptr) + 1;
902     /*buffer_truncate(b, out - b->ptr);*/
903 }
904 
905 void buffer_to_lower(buffer * const b) {
906     unsigned char * const restrict s = (unsigned char *)b->ptr;
907     for (uint32_t i = 0; i < b->used; ++i) {
908         if (light_isupper(s[i])) s[i] |= 0x20;
909     }
910 }
911 
912 
913 void buffer_to_upper(buffer * const b) {
914     unsigned char * const restrict s = (unsigned char *)b->ptr;
915     for (uint32_t i = 0; i < b->used; ++i) {
916         if (light_islower(s[i])) s[i] &= 0xdf;
917     }
918 }
919