netwerk/protocol/http/src/nsHttp.cpp

   1 /* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
   2 /* vim:set ts=4 sw=4 sts=4 et cin: */
   3 /* ***** BEGIN LICENSE BLOCK *****
   4  * Version: MPL 1.1/GPL 2.0/LGPL 2.1
   5  *
   6  * The contents of this file are subject to the Mozilla Public License Version
   7  * 1.1 (the "License"); you may not use this file except in compliance with
   8  * the License. You may obtain a copy of the License at
   9  * http://www.mozilla.org/MPL/
  10  *
  11  * Software distributed under the License is distributed on an "AS IS" basis,
  12  * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
  13  * for the specific language governing rights and limitations under the
  14  * License.
  15  *
  16  * The Original Code is Mozilla.
  17  *
  18  * The Initial Developer of the Original Code is
  19  * Netscape Communications.
  20  * Portions created by the Initial Developer are Copyright (C) 2001
  21  * the Initial Developer. All Rights Reserved.
  22  *
  23  * Contributor(s):
  24  *   Darin Fisher <darin@netscape.com> (original author)
  25  *
  26  * Alternatively, the contents of this file may be used under the terms of
  27  * either the GNU General Public License Version 2 or later (the "GPL"), or
  28  * the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
  29  * in which case the provisions of the GPL or the LGPL are applicable instead
  30  * of those above. If you wish to allow use of your version of this file only
  31  * under the terms of either the GPL or the LGPL, and not to allow others to
  32  * use your version of this file under the terms of the MPL, indicate your
  33  * decision by deleting the provisions above and replace them with the notice
  34  * and other provisions required by the GPL or the LGPL. If you do not delete
  35  * the provisions above, a recipient may use your version of this file under
  36  * the terms of any one of the MPL, the GPL or the LGPL.
  37  *
  38  * ***** END LICENSE BLOCK ***** */
  39
  40 #include "nsHttp.h"
  41 #include "nsAutoLock.h"
  42 #include "pldhash.h"
  43 #include "nsCRT.h"
  44 #include "prbit.h"
  45
  46 #if defined(PR_LOGGING)
  47 PRLogModuleInfo *gHttpLog = nsnull;
  48 #endif
  49
  50 // define storage for all atoms
  51 #define HTTP_ATOM(_name, _value) nsHttpAtom nsHttp::_name = { _value };
  52 #include "nsHttpAtomList.h"
  53 #undef HTTP_ATOM
  54
  55 // find out how many atoms we have
  56 #define HTTP_ATOM(_name, _value) Unused_ ## _name,
  57 enum {
  58 #include "nsHttpAtomList.h"
  59     NUM_HTTP_ATOMS
  60 };
  61 #undef HTTP_ATOM
  62
  63 // we keep a linked list of atoms allocated on the heap for easy clean up when
  64 // the atom table is destroyed.  The structure and value string are allocated
  65 // as one contiguous block.
  66
  67 struct HttpHeapAtom {
  68     struct HttpHeapAtom *next;
  69     char                 value[1];
  70 };
  71
  72 static struct PLDHashTable  sAtomTable = {0};
  73 static struct HttpHeapAtom *sHeapAtoms = nsnull;
  74 static PRLock              *sLock = nsnull;
  75
  76 HttpHeapAtom *
  77 NewHeapAtom(const char *value) {
  78     int len = strlen(value);
  79
  80     HttpHeapAtom *a =
  81         reinterpret_cast<HttpHeapAtom *>(malloc(sizeof(*a) + len));
  82     if (!a)
  83         return nsnull;
  84     memcpy(a->value, value, len + 1);
  85
  86     // add this heap atom to the list of all heap atoms
  87     a->next = sHeapAtoms;
  88     sHeapAtoms = a;
  89
  90     return a;
  91 }
  92
  93 // Hash string ignore case, based on PL_HashString
  94 static PLDHashNumber
  95 StringHash(PLDHashTable *table, const void *key)
  96 {
  97     PLDHashNumber h = 0;
  98     for (const char *s = reinterpret_cast<const char*>(key); *s; ++s)
  99         h = PR_ROTATE_LEFT32(h, 4) ^ nsCRT::ToLower(*s);
 100     return h;
 101 }
 102
 103 static PRBool
 104 StringCompare(PLDHashTable *table, const PLDHashEntryHdr *entry,
 105               const void *testKey)
 106 {
 107     const void *entryKey =
 108             reinterpret_cast<const PLDHashEntryStub *>(entry)->key;
 109
 110     return PL_strcasecmp(reinterpret_cast<const char *>(entryKey),
 111                          reinterpret_cast<const char *>(testKey)) == 0;
 112 }
 113
 114 static const PLDHashTableOps ops = {
 115     PL_DHashAllocTable,
 116     PL_DHashFreeTable,
 117     StringHash,
 118     StringCompare,
 119     PL_DHashMoveEntryStub,
 120     PL_DHashClearEntryStub,
 121     PL_DHashFinalizeStub,
 122     nsnull
 123 };
 124
 125 // We put the atoms in a hash table for speedy lookup.. see ResolveAtom.
 126 nsresult
 127 nsHttp::CreateAtomTable()
 128 {
 129     NS_ASSERTION(!sAtomTable.ops, "atom table already initialized");
 130
 131     if (!sLock) {
 132         sLock = PR_NewLock();
 133         if (!sLock)
 134             return NS_ERROR_OUT_OF_MEMORY;
 135     }
 136
 137     // The capacity for this table is initialized to a value greater than the
 138     // number of known atoms (NUM_HTTP_ATOMS) because we expect to encounter a
 139     // few random headers right off the bat.
 140     if (!PL_DHashTableInit(&sAtomTable, &ops, nsnull, sizeof(PLDHashEntryStub),
 141                            NUM_HTTP_ATOMS + 10)) {
 142         sAtomTable.ops = nsnull;
 143         return NS_ERROR_OUT_OF_MEMORY;
 144     }
 145
 146     // fill the table with our known atoms
 147     const char *const atoms[] = {
 148 #define HTTP_ATOM(_name, _value) nsHttp::_name._val,
 149 #include "nsHttpAtomList.h"
 150 #undef HTTP_ATOM
 151         nsnull
 152     };
 153
 154     for (int i = 0; atoms[i]; ++i) {
 155         PLDHashEntryStub *stub = reinterpret_cast<PLDHashEntryStub *>
 156                                                  (PL_DHashTableOperate(&sAtomTable, atoms[i], PL_DHASH_ADD));
 157         if (!stub)
 158             return NS_ERROR_OUT_OF_MEMORY;
 159
 160         NS_ASSERTION(!stub->key, "duplicate static atom");
 161         stub->key = atoms[i];
 162     }
 163
 164     return NS_OK;
 165 }
 166
 167 void
 168 nsHttp::DestroyAtomTable()
 169 {
 170     if (sAtomTable.ops) {
 171         PL_DHashTableFinish(&sAtomTable);
 172         sAtomTable.ops = nsnull;
 173     }
 174
 175     while (sHeapAtoms) {
 176         HttpHeapAtom *next = sHeapAtoms->next;
 177         free(sHeapAtoms);
 178         sHeapAtoms = next;
 179     }
 180
 181     if (sLock) {
 182         PR_DestroyLock(sLock);
 183         sLock = nsnull;
 184     }
 185 }
 186
 187 // this function may be called from multiple threads
 188 nsHttpAtom
 189 nsHttp::ResolveAtom(const char *str)
 190 {
 191     nsHttpAtom atom = { nsnull };
 192
 193     if (!str || !sAtomTable.ops)
 194         return atom;
 195
 196     nsAutoLock lock(sLock);
 197
 198     PLDHashEntryStub *stub = reinterpret_cast<PLDHashEntryStub *>
 199                                              (PL_DHashTableOperate(&sAtomTable, str, PL_DHASH_ADD));
 200     if (!stub)
 201         return atom;  // out of memory
 202
 203     if (stub->key) {
 204         atom._val = reinterpret_cast<const char *>(stub->key);
 205         return atom;
 206     }
 207
 208     // if the atom could not be found in the atom table, then we'll go
 209     // and allocate a new atom on the heap.
 210     HttpHeapAtom *heapAtom = NewHeapAtom(str);
 211     if (!heapAtom)
 212         return atom;  // out of memory
 213
 214     stub->key = atom._val = heapAtom->value;
 215     return atom;
 216 }
 217
 218 //
 219 // From section 2.2 of RFC 2616, a token is defined as:
 220 //
 221 //   token          = 1*<any CHAR except CTLs or separators>
 222 //   CHAR           = <any US-ASCII character (octets 0 - 127)>
 223 //   separators     = "(" | ")" | "<" | ">" | "@"
 224 //                  | "," | ";" | ":" | "\" | <">
 225 //                  | "/" | "[" | "]" | "?" | "="
 226 //                  | "{" | "}" | SP | HT
 227 //   CTL            = <any US-ASCII control character
 228 //                    (octets 0 - 31) and DEL (127)>
 229 //   SP             = <US-ASCII SP, space (32)>
 230 //   HT             = <US-ASCII HT, horizontal-tab (9)>
 231 //
 232 static const char kValidTokenMap[128] = {
 233     0, 0, 0, 0, 0, 0, 0, 0, //   0
 234     0, 0, 0, 0, 0, 0, 0, 0, //   8
 235     0, 0, 0, 0, 0, 0, 0, 0, //  16
 236     0, 0, 0, 0, 0, 0, 0, 0, //  24
 237
 238     0, 1, 0, 1, 1, 1, 1, 1, //  32
 239     0, 0, 1, 1, 0, 1, 1, 0, //  40
 240     1, 1, 1, 1, 1, 1, 1, 1, //  48
 241     1, 1, 0, 0, 0, 0, 0, 0, //  56
 242
 243     0, 1, 1, 1, 1, 1, 1, 1, //  64
 244     1, 1, 1, 1, 1, 1, 1, 1, //  72
 245     1, 1, 1, 1, 1, 1, 1, 1, //  80
 246     1, 1, 1, 0, 0, 0, 1, 1, //  88
 247
 248     1, 1, 1, 1, 1, 1, 1, 1, //  96
 249     1, 1, 1, 1, 1, 1, 1, 1, // 104
 250     1, 1, 1, 1, 1, 1, 1, 1, // 112
 251     1, 1, 1, 0, 1, 0, 1, 0  // 120
 252 };
 253 PRBool
 254 nsHttp::IsValidToken(const char *start, const char *end)
 255 {
 256     if (start == end)
 257         return PR_FALSE;
 258
 259     for (; start != end; ++start) {
 260         const unsigned char idx = *start;
 261         if (idx > 127 || !kValidTokenMap[idx])
 262             return PR_FALSE;
 263     }
 264
 265     return PR_TRUE;
 266 }
 267
 268 const char *
 269 nsHttp::FindToken(const char *input, const char *token, const char *seps)
 270 {
 271     if (!input)
 272         return nsnull;
 273
 274     int inputLen = strlen(input);
 275     int tokenLen = strlen(token);
 276
 277     if (inputLen < tokenLen)
 278         return nsnull;
 279
 280     const char *inputTop = input;
 281     const char *inputEnd = input + inputLen - tokenLen;
 282     for (; input <= inputEnd; ++input) {
 283         if (PL_strncasecmp(input, token, tokenLen) == 0) {
 284             if (input > inputTop && !strchr(seps, *(input - 1)))
 285                 continue;
 286             if (input < inputEnd && !strchr(seps, *(input + tokenLen)))
 287                 continue;
 288             return input;
 289         }
 290     }
 291
 292     return nsnull;
 293 }
 294
 295 PRBool
 296 nsHttp::ParseInt64(const char *input, const char **next, PRInt64 *r)
 297 {
 298     const char *start = input;
 299     *r = 0;
 300     while (*input >= '0' && *input <= '9') {
 301         PRInt64 next = 10 * (*r) + (*input - '0');
 302         if (next < *r) // overflow?
 303             return PR_FALSE;
 304         *r = next;
 305         ++input;
 306     }
 307     if (input == start) // nothing parsed?
 308         return PR_FALSE;
 309     if (next)
 310         *next = input;
 311     return PR_TRUE;
 312 }