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                 FLEX_ALLOC_MEM(dir, name, ce->name, namelen);
  59                 hashmap_entry_init(dir, memihash(ce->name, namelen));
  60                 dir->namelen = namelen;
  61                 hashmap_add(&istate->dir_hash, dir);
  62
  63                 /* recursively add missing parent directories */
  64                 dir->parent = hash_dir_entry(istate, ce, namelen);
  65         }
  66         return dir;
  67 }
  68
  69 static void add_dir_entry(struct index_state *istate, struct cache_entry *ce)
  70 {
  71         /* Add reference to the directory entry (and parents if 0). */
  72         struct dir_entry *dir = hash_dir_entry(istate, ce, ce_namelen(ce));
  73         while (dir && !(dir->nr++))
  74                 dir = dir->parent;
  75 }
  76
  77 static void remove_dir_entry(struct index_state *istate, struct cache_entry *ce)
  78 {
  79         /*
  80          * Release reference to the directory entry. If 0, remove and continue
  81          * with parent directory.
  82          */
  83         struct dir_entry *dir = hash_dir_entry(istate, ce, ce_namelen(ce));
  84         while (dir && !(--dir->nr)) {
  85                 struct dir_entry *parent = dir->parent;
  86                 hashmap_remove(&istate->dir_hash, dir, NULL);
  87                 free(dir);
  88                 dir = parent;
  89         }
  90 }
  91
  92 static void hash_index_entry(struct index_state *istate, struct cache_entry *ce)
  93 {
  94         if (ce->ce_flags & CE_HASHED)
  95                 return;
  96         ce->ce_flags |= CE_HASHED;
  97         hashmap_entry_init(ce, memihash(ce->name, ce_namelen(ce)));
  98         hashmap_add(&istate->name_hash, ce);
  99
 100         if (ignore_case)
 101                 add_dir_entry(istate, ce);
 102 }
 103
 104 static int cache_entry_cmp(const struct cache_entry *ce1,
 105                 const struct cache_entry *ce2, const void *remove)
 106 {
 107         /*
 108          * For remove_name_hash, find the exact entry (pointer equality); for
 109          * index_file_exists, find all entries with matching hash code and
 110          * decide whether the entry matches in same_name.
 111          */
 112         return remove ? !(ce1 == ce2) : 0;
 113 }
 114
 115 static void lazy_init_name_hash(struct index_state *istate)
 116 {
 117         int nr;
 118
 119         if (istate->name_hash_initialized)
 120                 return;
 121         hashmap_init(&istate->name_hash, (hashmap_cmp_fn) cache_entry_cmp,
 122                         istate->cache_nr);
 123         hashmap_init(&istate->dir_hash, (hashmap_cmp_fn) dir_entry_cmp, 0);
 124         for (nr = 0; nr < istate->cache_nr; nr++)
 125                 hash_index_entry(istate, istate->cache[nr]);
 126         istate->name_hash_initialized = 1;
 127 }
 128
 129 void add_name_hash(struct index_state *istate, struct cache_entry *ce)
 130 {
 131         if (istate->name_hash_initialized)
 132                 hash_index_entry(istate, ce);
 133 }
 134
 135 void remove_name_hash(struct index_state *istate, struct cache_entry *ce)
 136 {
 137         if (!istate->name_hash_initialized || !(ce->ce_flags & CE_HASHED))
 138                 return;
 139         ce->ce_flags &= ~CE_HASHED;
 140         hashmap_remove(&istate->name_hash, ce, ce);
 141
 142         if (ignore_case)
 143                 remove_dir_entry(istate, ce);
 144 }
 145
 146 static int slow_same_name(const char *name1, int len1, const char *name2, int len2)
 147 {
 148         if (len1 != len2)
 149                 return 0;
 150
 151         while (len1) {
 152                 unsigned char c1 = *name1++;
 153                 unsigned char c2 = *name2++;
 154                 len1--;
 155                 if (c1 != c2) {
 156                         c1 = toupper(c1);
 157                         c2 = toupper(c2);
 158                         if (c1 != c2)
 159                                 return 0;
 160                 }
 161         }
 162         return 1;
 163 }
 164
 165 static int same_name(const struct cache_entry *ce, const char *name, int namelen, int icase)
 166 {
 167         int len = ce_namelen(ce);
 168
 169         /*
 170          * Always do exact compare, even if we want a case-ignoring comparison;
 171          * we do the quick exact one first, because it will be the common case.
 172          */
 173         if (len == namelen && !memcmp(name, ce->name, len))
 174                 return 1;
 175
 176         if (!icase)
 177                 return 0;
 178
 179         return slow_same_name(name, namelen, ce->name, len);
 180 }
 181
 182 int index_dir_exists(struct index_state *istate, const char *name, int namelen)
 183 {
 184         struct dir_entry *dir;
 185
 186         lazy_init_name_hash(istate);
 187         dir = find_dir_entry(istate, name, namelen);
 188         return dir && dir->nr;
 189 }
 190
 191 void adjust_dirname_case(struct index_state *istate, char *name)
 192 {
 193         const char *startPtr = name;
 194         const char *ptr = startPtr;
 195
 196         lazy_init_name_hash(istate);
 197         while (*ptr) {
 198                 while (*ptr && *ptr != '/')
 199                         ptr++;
 200
 201                 if (*ptr == '/') {
 202                         struct dir_entry *dir;
 203
 204                         ptr++;
 205                         dir = find_dir_entry(istate, name, ptr - name + 1);
 206                         if (dir) {
 207                                 memcpy((void *)startPtr, dir->name + (startPtr - name), ptr - startPtr);
 208                                 startPtr = ptr;
 209                         }
 210                 }
 211         }
 212 }
 213
 214 struct cache_entry *index_file_exists(struct index_state *istate, const char *name, int namelen, int icase)
 215 {
 216         struct cache_entry *ce;
 217
 218         lazy_init_name_hash(istate);
 219
 220         ce = hashmap_get_from_hash(&istate->name_hash,
 221                                    memihash(name, namelen), NULL);
 222         while (ce) {
 223                 if (same_name(ce, name, namelen, icase))
 224                         return ce;
 225                 ce = hashmap_get_next(&istate->name_hash, ce);
 226         }
 227         return NULL;
 228 }
 229
 230 void free_name_hash(struct index_state *istate)
 231 {
 232         if (!istate->name_hash_initialized)
 233                 return;
 234         istate->name_hash_initialized = 0;
 235
 236         hashmap_free(&istate->name_hash, 0);
 237         hashmap_free(&istate->dir_hash, 1);
 238 }