parsing.c

   1 /* config.c: parsing of config files
   2  *
   3  * Copyright (C) 2006 Lars Hjemli
   4  *
   5  * Licensed under GNU General Public License v2
   6  *   (see COPYING for full license text)
   7  */
   8
   9 #include "cgit.h"
  10
  11 /*
  12  * url syntax: [repo ['/' cmd [ '/' path]]]
  13  *   repo: any valid repo url, may contain '/'
  14  *   cmd:  log | commit | diff | tree | view | blob | snapshot
  15  *   path: any valid path, may contain '/'
  16  *
  17  */
  18 void cgit_parse_url(const char *url)
  19 {
  20         char *cmd, *p;
  21
  22         ctx.repo = NULL;
  23         if (!url || url[0] == '\0')
  24                 return;
  25
  26         ctx.repo = cgit_get_repoinfo(url);
  27         if (ctx.repo) {
  28                 ctx.qry.repo = ctx.repo->url;
  29                 return;
  30         }
  31
  32         cmd = strchr(url, '/');
  33         while (!ctx.repo && cmd) {
  34                 cmd[0] = '\0';
  35                 ctx.repo = cgit_get_repoinfo(url);
  36                 if (ctx.repo == NULL) {
  37                         cmd[0] = '/';
  38                         cmd = strchr(cmd + 1, '/');
  39                         continue;
  40                 }
  41
  42                 ctx.qry.repo = ctx.repo->url;
  43                 p = strchr(cmd + 1, '/');
  44                 if (p) {
  45                         p[0] = '\0';
  46                         if (p[1])
  47                                 ctx.qry.path = trim_end(p + 1, '/');
  48                 }
  49                 if (cmd[1])
  50                         ctx.qry.page = xstrdup(cmd + 1);
  51                 return;
  52         }
  53 }
  54
  55 char *substr(const char *head, const char *tail)
  56 {
  57         char *buf;
  58
  59         buf = xmalloc(tail - head + 1);
  60         strncpy(buf, head, tail - head);
  61         buf[tail - head] = '\0';
  62         return buf;
  63 }
  64
  65 char *parse_user(char *t, char **name, char **email, unsigned long *date)
  66 {
  67         char *p = t;
  68         int mode = 1;
  69
  70         while (p && *p) {
  71                 if (mode == 1 && *p == '<') {
  72                         *name = substr(t, p - 1);
  73                         t = p;
  74                         mode++;
  75                 } else if (mode == 1 && *p == '\n') {
  76                         *name = substr(t, p);
  77                         p++;
  78                         break;
  79                 } else if (mode == 2 && *p == '>') {
  80                         *email = substr(t, p + 1);
  81                         t = p;
  82                         mode++;
  83                 } else if (mode == 2 && *p == '\n') {
  84                         *email = substr(t, p);
  85                         p++;
  86                         break;
  87                 } else if (mode == 3 && isdigit(*p)) {
  88                         *date = atol(p);
  89                         mode++;
  90                 } else if (*p == '\n') {
  91                         p++;
  92                         break;
  93                 }
  94                 p++;
  95         }
  96         return p;
  97 }
  98
  99 #ifdef NO_ICONV
 100 #define reencode(a, b, c)
 101 #else
 102 const char *reencode(char **txt, const char *src_enc, const char *dst_enc)
 103 {
 104         char *tmp;
 105
 106         if (!txt)
 107                 return NULL;
 108
 109         if (!*txt || !src_enc || !dst_enc)
 110                 return *txt;
 111
 112         /* no encoding needed if src_enc equals dst_enc */
 113         if(!strcasecmp(src_enc, dst_enc))
 114                 return *txt;
 115
 116         tmp = reencode_string(*txt, dst_enc, src_enc);
 117         if (tmp) {
 118                 free(*txt);
 119                 *txt = tmp;
 120         }
 121         return *txt;
 122 }
 123 #endif
 124
 125 struct commitinfo *cgit_parse_commit(struct commit *commit)
 126 {
 127         struct commitinfo *ret;
 128         char *p = commit->buffer, *t;
 129
 130         ret = xmalloc(sizeof(*ret));
 131         ret->commit = commit;
 132         ret->author = NULL;
 133         ret->author_email = NULL;
 134         ret->committer = NULL;
 135         ret->committer_email = NULL;
 136         ret->subject = NULL;
 137         ret->msg = NULL;
 138         ret->msg_encoding = NULL;
 139
 140         if (p == NULL)
 141                 return ret;
 142
 143         if (strncmp(p, "tree ", 5))
 144                 die("Bad commit: %s", sha1_to_hex(commit->object.sha1));
 145         else
 146                 p += 46; // "tree " + hex[40] + "\n"
 147
 148         while (!strncmp(p, "parent ", 7))
 149                 p += 48; // "parent " + hex[40] + "\n"
 150
 151         if (p && !strncmp(p, "author ", 7)) {
 152                 p = parse_user(p + 7, &ret->author, &ret->author_email,
 153                         &ret->author_date);
 154         }
 155
 156         if (p && !strncmp(p, "committer ", 9)) {
 157                 p = parse_user(p + 9, &ret->committer, &ret->committer_email,
 158                         &ret->committer_date);
 159         }
 160
 161         if (p && !strncmp(p, "encoding ", 9)) {
 162                 p += 9;
 163                 t = strchr(p, '\n');
 164                 if (t) {
 165                         ret->msg_encoding = substr(p, t + 1);
 166                         p = t + 1;
 167                 }
 168         }
 169
 170         /* if no special encoding is found, assume UTF-8 */
 171         if(!ret->msg_encoding)
 172                 ret->msg_encoding = xstrdup("UTF-8");
 173
 174         // skip unknown header fields
 175         while (p && *p && (*p != '\n')) {
 176                 p = strchr(p, '\n');
 177                 if (p)
 178                         p++;
 179         }
 180
 181         // skip empty lines between headers and message
 182         while (p && *p == '\n')
 183                 p++;
 184
 185         if (!p)
 186                 return ret;
 187
 188         t = strchr(p, '\n');
 189         if (t) {
 190                 ret->subject = substr(p, t);
 191                 p = t + 1;
 192
 193                 while (p && *p == '\n') {
 194                         p = strchr(p, '\n');
 195                         if (p)
 196                                 p++;
 197                 }
 198                 if (p)
 199                         ret->msg = xstrdup(p);
 200         } else
 201                 ret->subject = xstrdup(p);
 202
 203         reencode(&ret->author, ret->msg_encoding, PAGE_ENCODING);
 204         reencode(&ret->author_email, ret->msg_encoding, PAGE_ENCODING);
 205         reencode(&ret->committer, ret->msg_encoding, PAGE_ENCODING);
 206         reencode(&ret->committer_email, ret->msg_encoding, PAGE_ENCODING);
 207         reencode(&ret->subject, ret->msg_encoding, PAGE_ENCODING);
 208         reencode(&ret->msg, ret->msg_encoding, PAGE_ENCODING);
 209
 210         return ret;
 211 }
 212
 213
 214 struct taginfo *cgit_parse_tag(struct tag *tag)
 215 {
 216         void *data;
 217         enum object_type type;
 218         unsigned long size;
 219         char *p;
 220         struct taginfo *ret;
 221
 222         data = read_sha1_file(tag->object.sha1, &type, &size);
 223         if (!data || type != OBJ_TAG) {
 224                 free(data);
 225                 return 0;
 226         }
 227
 228         ret = xmalloc(sizeof(*ret));
 229         ret->tagger = NULL;
 230         ret->tagger_email = NULL;
 231         ret->tagger_date = 0;
 232         ret->msg = NULL;
 233
 234         p = data;
 235
 236         while (p && *p) {
 237                 if (*p == '\n')
 238                         break;
 239
 240                 if (!strncmp(p, "tagger ", 7)) {
 241                         p = parse_user(p + 7, &ret->tagger, &ret->tagger_email,
 242                                 &ret->tagger_date);
 243                 } else {
 244                         p = strchr(p, '\n');
 245                         if (p)
 246                                 p++;
 247                 }
 248         }
 249
 250         // skip empty lines between headers and message
 251         while (p && *p == '\n')
 252                 p++;
 253
 254         if (p && *p)
 255                 ret->msg = xstrdup(p);
 256         free(data);
 257         return ret;
 258 }