1.1 --- a/src/string.c Sat Sep 03 15:11:23 2022 +0200
1.2 +++ b/src/string.c Fri Sep 09 20:19:08 2022 +0200
1.3 @@ -72,6 +72,15 @@
1.4 str->length = 0;
1.5 }
1.6
1.7 +void cx_strfree_a(
1.8 + CxAllocator *alloc,
1.9 + cxmutstr *str
1.10 +) {
1.11 + cxFree(alloc, str->ptr);
1.12 + str->ptr = NULL;
1.13 + str->length = 0;
1.14 +}
1.15 +
1.16 size_t cx_strlen(
1.17 size_t count,
1.18 ...
1.19 @@ -235,6 +244,11 @@
1.20 return haystack;
1.21 }
1.22
1.23 + /* optimize for single-char needles */
1.24 + if (needle.length == 1) {
1.25 + return cx_strchr(haystack, *needle.ptr);
1.26 + }
1.27 +
1.28 /*
1.29 * IMPORTANT:
1.30 * Our prefix table contains the prefix length PLUS ONE
1.31 @@ -308,8 +322,55 @@
1.32 size_t limit,
1.33 cxstring *output
1.34 ) {
1.35 - // TODO: implement
1.36 - return 0;
1.37 + /* special case: output limit is zero */
1.38 + if (limit == 0) return 0;
1.39 +
1.40 + /* special case: delimiter is empty */
1.41 + if (delim.length == 0) {
1.42 + output[0] = string;
1.43 + return 1;
1.44 + }
1.45 +
1.46 + /* special cases: delimiter is at least as large as the string */
1.47 + if (delim.length >= string.length) {
1.48 + /* exact match */
1.49 + if (cx_strcmp(string, delim) == 0) {
1.50 + output[0] = cx_strn(string.ptr, 0);
1.51 + output[1] = cx_strn(string.ptr + string.length, 0);
1.52 + return 2;
1.53 + } else /* no match possible */ {
1.54 + output[0] = string;
1.55 + return 1;
1.56 + }
1.57 + }
1.58 +
1.59 + size_t n = 0;
1.60 + cxstring curpos = string;
1.61 + while (1) {
1.62 + ++n;
1.63 + cxstring match = cx_strstr(curpos, delim);
1.64 + if (match.length > 0) {
1.65 + /* is the limit reached? */
1.66 + if (n < limit) {
1.67 + /* copy the current string to the array */
1.68 + cxstring item = cx_strn(curpos.ptr, match.ptr - curpos.ptr);
1.69 + output[n - 1] = item;
1.70 + size_t processed = item.length + delim.length;
1.71 + curpos.ptr += processed;
1.72 + curpos.length -= processed;
1.73 + } else {
1.74 + /* limit reached, copy the _full_ remaining string */
1.75 + output[n - 1] = curpos;
1.76 + break;
1.77 + }
1.78 + } else {
1.79 + /* no more matches, copy last string */
1.80 + output[n - 1] = curpos;
1.81 + break;
1.82 + }
1.83 + }
1.84 +
1.85 + return n;
1.86 }
1.87
1.88 size_t cx_strsplit_a(
1.89 @@ -319,8 +380,29 @@
1.90 size_t limit,
1.91 cxstring **output
1.92 ) {
1.93 - // TODO: implement
1.94 - return 0;
1.95 + /* find out how many splits we're going to make and allocate memory */
1.96 + size_t n = 0;
1.97 + cxstring curpos = string;
1.98 + while (1) {
1.99 + ++n;
1.100 + cxstring match = cx_strstr(curpos, delim);
1.101 + if (match.length > 0) {
1.102 + /* is the limit reached? */
1.103 + if (n < limit) {
1.104 + size_t processed = match.ptr - curpos.ptr + delim.length;
1.105 + curpos.ptr += processed;
1.106 + curpos.length -= processed;
1.107 + } else {
1.108 + /* limit reached */
1.109 + break;
1.110 + }
1.111 + } else {
1.112 + /* no more matches */
1.113 + break;
1.114 + }
1.115 + }
1.116 + *output = cxCalloc(allocator, n, sizeof(cxstring));
1.117 + return cx_strsplit(string, delim, n, *output);
1.118 }
1.119
1.120 size_t cx_strsplit_m(
1.121 @@ -344,7 +426,10 @@
1.122 delim, limit, (cxstring **) output);
1.123 }
1.124
1.125 -int cx_strcmp(cxstring s1, cxstring s2) {
1.126 +int cx_strcmp(
1.127 + cxstring s1,
1.128 + cxstring s2
1.129 +) {
1.130 if (s1.length == s2.length) {
1.131 return memcmp(s1.ptr, s2.ptr, s1.length);
1.132 } else if (s1.length > s2.length) {
1.133 @@ -354,7 +439,10 @@
1.134 }
1.135 }
1.136
1.137 -int cx_strcasecmp(cxstring s1, cxstring s2) {
1.138 +int cx_strcasecmp(
1.139 + cxstring s1,
1.140 + cxstring s2
1.141 +) {
1.142 if (s1.length == s2.length) {
1.143 #ifdef _WIN32
1.144 return _strnicmp(s1.ptr, s2.ptr, s1.length);
1.145 @@ -368,7 +456,10 @@
1.146 }
1.147 }
1.148
1.149 -cxmutstr cx_strdup_a(CxAllocator *allocator, cxstring string) {
1.150 +cxmutstr cx_strdup_a(
1.151 + CxAllocator *allocator,
1.152 + cxstring string
1.153 +) {
1.154 cxmutstr result = {
1.155 cxMalloc(allocator, string.length + 1),
1.156 string.length
1.157 @@ -400,18 +491,27 @@
1.158 return (cxmutstr) {(char *) result.ptr, result.length};
1.159 }
1.160
1.161 -bool cx_strprefix(cxstring string, cxstring prefix) {
1.162 +bool cx_strprefix(
1.163 + cxstring string,
1.164 + cxstring prefix
1.165 +) {
1.166 if (string.length < prefix.length) return false;
1.167 return memcmp(string.ptr, prefix.ptr, prefix.length) == 0;
1.168 }
1.169
1.170 -bool cx_strsuffix(cxstring string, cxstring suffix) {
1.171 +bool cx_strsuffix(
1.172 + cxstring string,
1.173 + cxstring suffix
1.174 +) {
1.175 if (string.length < suffix.length) return false;
1.176 return memcmp(string.ptr + string.length - suffix.length,
1.177 suffix.ptr, suffix.length) == 0;
1.178 }
1.179
1.180 -bool cx_casestrprefix(cxstring string, cxstring prefix) {
1.181 +bool cx_strcaseprefix(
1.182 + cxstring string,
1.183 + cxstring prefix
1.184 +) {
1.185 if (string.length < prefix.length) return false;
1.186 #ifdef _WIN32
1.187 return _strnicmp(string.ptr, prefix.ptr, prefix.length) == 0;
1.188 @@ -420,7 +520,10 @@
1.189 #endif
1.190 }
1.191
1.192 -bool cx_casestrsuffix(cxstring string, cxstring suffix) {
1.193 +bool cx_strcasesuffix(
1.194 + cxstring string,
1.195 + cxstring suffix
1.196 +) {
1.197 if (string.length < suffix.length) return false;
1.198 #ifdef _WIN32
1.199 return _strnicmp(string.ptr+string.length-suffix.length,
1.200 @@ -442,3 +545,133 @@
1.201 string.ptr[i] = toupper(string.ptr[i]);
1.202 }
1.203 }
1.204 +
1.205 +#define REPLACE_INDEX_BUFFER_MAX 100
1.206 +
1.207 +struct cx_strreplace_ibuf {
1.208 + size_t *buf;
1.209 + unsigned int len; /* small indices */
1.210 + struct cx_strreplace_ibuf *next;
1.211 +};
1.212 +
1.213 +static void cx_strrepl_free_ibuf(struct cx_strreplace_ibuf *buf) {
1.214 + while (buf) {
1.215 + struct cx_strreplace_ibuf *next = buf->next;
1.216 + free(buf->buf);
1.217 + free(buf);
1.218 + buf = next;
1.219 + }
1.220 +}
1.221 +
1.222 +cxmutstr cx_strreplacen_a(
1.223 + CxAllocator *allocator,
1.224 + cxstring str,
1.225 + cxstring pattern,
1.226 + cxstring replacement,
1.227 + size_t replmax
1.228 +) {
1.229 +
1.230 + if (pattern.length == 0 || pattern.length > str.length || replmax == 0)
1.231 + return cx_strdup_a(allocator, str);
1.232 +
1.233 + /* Compute expected buffer length */
1.234 + size_t ibufmax = str.length / pattern.length;
1.235 + size_t ibuflen = replmax < ibufmax ? replmax : ibufmax;
1.236 + if (ibuflen > REPLACE_INDEX_BUFFER_MAX) {
1.237 + ibuflen = REPLACE_INDEX_BUFFER_MAX;
1.238 + }
1.239 +
1.240 + /* Allocate first index buffer */
1.241 + struct cx_strreplace_ibuf *firstbuf, *curbuf;
1.242 + firstbuf = curbuf = calloc(1, sizeof(struct cx_strreplace_ibuf));
1.243 + if (!firstbuf) return cx_mutstrn(NULL, 0);
1.244 + firstbuf->buf = calloc(ibuflen, sizeof(size_t));
1.245 + if (!firstbuf->buf) {
1.246 + free(firstbuf);
1.247 + return cx_mutstrn(NULL, 0);
1.248 + }
1.249 +
1.250 + /* Search occurrences */
1.251 + cxstring searchstr = str;
1.252 + size_t found = 0;
1.253 + do {
1.254 + cxstring match = cx_strstr(searchstr, pattern);
1.255 + if (match.length > 0) {
1.256 + /* Allocate next buffer in chain, if required */
1.257 + if (curbuf->len == ibuflen) {
1.258 + struct cx_strreplace_ibuf *nextbuf =
1.259 + calloc(1, sizeof(struct cx_strreplace_ibuf));
1.260 + if (!nextbuf) {
1.261 + cx_strrepl_free_ibuf(firstbuf);
1.262 + return cx_mutstrn(NULL, 0);
1.263 + }
1.264 + nextbuf->buf = calloc(ibuflen, sizeof(size_t));
1.265 + if (!nextbuf->buf) {
1.266 + free(nextbuf);
1.267 + cx_strrepl_free_ibuf(firstbuf);
1.268 + return cx_mutstrn(NULL, 0);
1.269 + }
1.270 + curbuf->next = nextbuf;
1.271 + curbuf = nextbuf;
1.272 + }
1.273 +
1.274 + /* Record match index */
1.275 + found++;
1.276 + size_t idx = match.ptr - str.ptr;
1.277 + curbuf->buf[curbuf->len++] = idx;
1.278 + searchstr.ptr = match.ptr + pattern.length;
1.279 + searchstr.length = str.length - idx - pattern.length;
1.280 + } else {
1.281 + break;
1.282 + }
1.283 + } while (searchstr.length > 0 && found < replmax);
1.284 +
1.285 + /* Allocate result string */
1.286 + cxmutstr result;
1.287 + {
1.288 + ssize_t adjlen = (ssize_t) replacement.length - (ssize_t) pattern.length;
1.289 + size_t rcount = 0;
1.290 + curbuf = firstbuf;
1.291 + do {
1.292 + rcount += curbuf->len;
1.293 + curbuf = curbuf->next;
1.294 + } while (curbuf);
1.295 + result.length = str.length + rcount * adjlen;
1.296 + result.ptr = cxMalloc(allocator, result.length);
1.297 + if (!result.ptr) {
1.298 + cx_strrepl_free_ibuf(firstbuf);
1.299 + return cx_mutstrn(NULL, 0);
1.300 + }
1.301 + }
1.302 +
1.303 + /* Build result string */
1.304 + curbuf = firstbuf;
1.305 + size_t srcidx = 0;
1.306 + char *destptr = result.ptr;
1.307 + do {
1.308 + for (size_t i = 0; i < curbuf->len; i++) {
1.309 + /* Copy source part up to next match*/
1.310 + size_t idx = curbuf->buf[i];
1.311 + size_t srclen = idx - srcidx;
1.312 + if (srclen > 0) {
1.313 + memcpy(destptr, str.ptr + srcidx, srclen);
1.314 + destptr += srclen;
1.315 + srcidx += srclen;
1.316 + }
1.317 +
1.318 + /* Copy the replacement and skip the source pattern */
1.319 + srcidx += pattern.length;
1.320 + memcpy(destptr, replacement.ptr, replacement.length);
1.321 + destptr += replacement.length;
1.322 + }
1.323 + curbuf = curbuf->next;
1.324 + } while (curbuf);
1.325 + memcpy(destptr, str.ptr + srcidx, str.length - srcidx);
1.326 +
1.327 + /* Free index buffer */
1.328 + cx_strrepl_free_ibuf(firstbuf);
1.329 +
1.330 + return result;
1.331 +}
1.332 +
1.333 +
