name-hash.c

   1 /*
   2  * name-hash.c
   3  *
   4  * Hashing names in the index state
   5  *
   6  * Copyright (C) 2008 Linus Torvalds
   7  */
   8 #define NO_THE_INDEX_COMPATIBILITY_MACROS
   9 #include "cache.h"
  10
  11 struct dir_entry {
  12         struct hashmap_entry ent;
  13         struct dir_entry *parent;
  14         int nr;
  15         unsigned int namelen;
  16         char name[FLEX_ARRAY];
  17 };
  18
  19 static int dir_entry_cmp(const struct dir_entry *e1,
  20                 const struct dir_entry *e2, const char *name)
  21 {
  22         return e1->namelen != e2->namelen || strncasecmp(e1->name,
  23                         name ? name : e2->name, e1->namelen);
  24 }
  25
  26 static struct dir_entry *find_dir_entry(struct index_state *istate,
  27                 const char *name, unsigned int namelen)
  28 {
  29         struct dir_entry key;
  30         hashmap_entry_init(&key, memihash(name, namelen));
  31         key.namelen = namelen;
  32         return hashmap_get(&istate->dir_hash, &key, name);
  33 }
  34
  35 static struct dir_entry *hash_dir_entry(struct index_state *istate,
  36                 struct cache_entry *ce, int namelen)
  37 {
  38         /*
  39          * Throw each directory component in the hash for quick lookup
  40          * during a git status. Directory components are stored without their
  41          * closing slash.  Despite submodules being a directory, they never
  42          * reach this point, because they are stored
  43          * in index_state.name_hash (as ordinary cache_entries).
  44          */
  45         struct dir_entry *dir;
  46
  47         /* get length of parent directory */
  48         while (namelen > 0 && !is_dir_sep(ce->name[namelen - 1]))
  49                 namelen--;
  50         if (namelen <= 0)
  51                 return NULL;
  52         namelen--;
  53
  54         /* lookup existing entry for that directory */
  55         dir = find_dir_entry(istate, ce->name, namelen);
  56         if (!dir) {
  57                 /* not found, create it and add to hash table */
  58                 dir = xcalloc(1, sizeof(struct dir_entry) + namelen + 1);
  59                 hashmap_entry_init(dir, memihash(ce->name, namelen));
  60                 dir->namelen = namelen;
  61                 strncpy(dir->name, ce->name, namelen);
  62                 hashmap_add(&istate->dir_hash, dir);
  63
  64                 /* recursively add missing parent directories */
  65                 dir->parent = hash_dir_entry(istate, ce, namelen);
  66         }
  67         return dir;
  68 }
  69
  70 static void add_dir_entry(struct index_state *istate, struct cache_entry *ce)
  71 {
  72         /* Add reference to the directory entry (and parents if 0). */
  73         struct dir_entry *dir = hash_dir_entry(istate, ce, ce_namelen(ce));
  74         while (dir && !(dir->nr++))
  75                 dir = dir->parent;
  76 }
  77
  78 static void remove_dir_entry(struct index_state *istate, struct cache_entry *ce)
  79 {
  80         /*
  81          * Release reference to the directory entry. If 0, remove and continue
  82          * with parent directory.
  83          */
  84         struct dir_entry *dir = hash_dir_entry(istate, ce, ce_namelen(ce));
  85         while (dir && !(--dir->nr)) {
  86                 struct dir_entry *parent = dir->parent;
  87                 hashmap_remove(&istate->dir_hash, dir, NULL);
  88                 free(dir);
  89                 dir = parent;
  90         }
  91 }
  92
  93 static void hash_index_entry(struct index_state *istate, struct cache_entry *ce)
  94 {
  95         if (ce->ce_flags & CE_HASHED)
  96                 return;
  97         ce->ce_flags |= CE_HASHED;
  98         hashmap_entry_init(ce, memihash(ce->name, ce_namelen(ce)));
  99         hashmap_add(&istate->name_hash, ce);
 100
 101         if (ignore_case)
 102                 add_dir_entry(istate, ce);
 103 }
 104
 105 static int cache_entry_cmp(const struct cache_entry *ce1,
 106                 const struct cache_entry *ce2, const void *remove)
 107 {
 108         /*
 109          * For remove_name_hash, find the exact entry (pointer equality); for
 110          * index_file_exists, find all entries with matching hash code and
 111          * decide whether the entry matches in same_name.
 112          */
 113         return remove ? !(ce1 == ce2) : 0;
 114 }
 115
 116 static void lazy_init_name_hash(struct index_state *istate)
 117 {
 118         int nr;
 119
 120         if (istate->name_hash_initialized)
 121                 return;
 122         hashmap_init(&istate->name_hash, (hashmap_cmp_fn) cache_entry_cmp,
 123                         istate->cache_nr);
 124         hashmap_init(&istate->dir_hash, (hashmap_cmp_fn) dir_entry_cmp, 0);
 125         for (nr = 0; nr < istate->cache_nr; nr++)
 126                 hash_index_entry(istate, istate->cache[nr]);
 127         istate->name_hash_initialized = 1;
 128 }
 129
 130 void add_name_hash(struct index_state *istate, struct cache_entry *ce)
 131 {
 132         if (istate->name_hash_initialized)
 133                 hash_index_entry(istate, ce);
 134 }
 135
 136 void remove_name_hash(struct index_state *istate, struct cache_entry *ce)
 137 {
 138         if (!istate->name_hash_initialized || !(ce->ce_flags & CE_HASHED))
 139                 return;
 140         ce->ce_flags &= ~CE_HASHED;
 141         hashmap_remove(&istate->name_hash, ce, ce);
 142
 143         if (ignore_case)
 144                 remove_dir_entry(istate, ce);
 145 }
 146
 147 static int slow_same_name(const char *name1, int len1, const char *name2, int len2)
 148 {
 149         if (len1 != len2)
 150                 return 0;
 151
 152         while (len1) {
 153                 unsigned char c1 = *name1++;
 154                 unsigned char c2 = *name2++;
 155                 len1--;
 156                 if (c1 != c2) {
 157                         c1 = toupper(c1);
 158                         c2 = toupper(c2);
 159                         if (c1 != c2)
 160                                 return 0;
 161                 }
 162         }
 163         return 1;
 164 }
 165
 166 static int same_name(const struct cache_entry *ce, const char *name, int namelen, int icase)
 167 {
 168         int len = ce_namelen(ce);
 169
 170         /*
 171          * Always do exact compare, even if we want a case-ignoring comparison;
 172          * we do the quick exact one first, because it will be the common case.
 173          */
 174         if (len == namelen && !memcmp(name, ce->name, len))
 175                 return 1;
 176
 177         if (!icase)
 178                 return 0;
 179
 180         return slow_same_name(name, namelen, ce->name, len);
 181 }
 182
 183 int index_dir_exists(struct index_state *istate, const char *name, int namelen)
 184 {
 185         struct dir_entry *dir;
 186
 187         lazy_init_name_hash(istate);
 188         dir = find_dir_entry(istate, name, namelen);
 189         return dir && dir->nr;
 190 }
 191
 192 void adjust_dirname_case(struct index_state *istate, char *name)
 193 {
 194         const char *startPtr = name;
 195         const char *ptr = startPtr;
 196
 197         lazy_init_name_hash(istate);
 198         while (*ptr) {
 199                 while (*ptr && *ptr != '/')
 200                         ptr++;
 201
 202                 if (*ptr == '/') {
 203                         struct dir_entry *dir;
 204
 205                         ptr++;
 206                         dir = find_dir_entry(istate, name, ptr - name + 1);
 207                         if (dir) {
 208                                 memcpy((void *)startPtr, dir->name + (startPtr - name), ptr - startPtr);
 209                                 startPtr = ptr;
 210                         }
 211                 }
 212         }
 213 }
 214
 215 struct cache_entry *index_file_exists(struct index_state *istate, const char *name, int namelen, int icase)
 216 {
 217         struct cache_entry *ce;
 218
 219         lazy_init_name_hash(istate);
 220
 221         ce = hashmap_get_from_hash(&istate->name_hash,
 222                                    memihash(name, namelen), NULL);
 223         while (ce) {
 224                 if (same_name(ce, name, namelen, icase))
 225                         return ce;
 226                 ce = hashmap_get_next(&istate->name_hash, ce);
 227         }
 228         return NULL;
 229 }
 230
 231 void free_name_hash(struct index_state *istate)
 232 {
 233         if (!istate->name_hash_initialized)
 234                 return;
 235         istate->name_hash_initialized = 0;
 236
 237         hashmap_free(&istate->name_hash, 0);
 238         hashmap_free(&istate->dir_hash, 1);
 239 }