src/HttpHdrRange.cc

   1
   2 /*
   3  * $Id: HttpHdrRange.cc,v 1.45 2007/08/13 17:20:51 hno Exp $
   4  *
   5  * DEBUG: section 64    HTTP Range Header
   6  * AUTHOR: Alex Rousskov
   7  *
   8  * SQUID Web Proxy Cache          http://www.squid-cache.org/
   9  * ----------------------------------------------------------
  10  *
  11  *  Squid is the result of efforts by numerous individuals from
  12  *  the Internet community; see the CONTRIBUTORS file for full
  13  *  details.   Many organizations have provided support for Squid's
  14  *  development; see the SPONSORS file for full details.  Squid is
  15  *  Copyrighted (C) 2001 by the Regents of the University of
  16  *  California; see the COPYRIGHT file for full details.  Squid
  17  *  incorporates software developed and/or copyrighted by other
  18  *  sources; see the CREDITS file for full details.
  19  *
  20  *  This program is free software; you can redistribute it and/or modify
  21  *  it under the terms of the GNU General Public License as published by
  22  *  the Free Software Foundation; either version 2 of the License, or
  23  *  (at your option) any later version.
  24  *
  25  *  This program is distributed in the hope that it will be useful,
  26  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  27  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  28  *  GNU General Public License for more details.
  29  *
  30  *  You should have received a copy of the GNU General Public License
  31  *  along with this program; if not, write to the Free Software
  32  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
  33  *
  34  */
  35
  36 #include "squid.h"
  37 #include "Store.h"
  38 #include "HttpHeaderRange.h"
  39 #include "client_side_request.h"
  40 #include "HttpReply.h"
  41
  42 /*
  43  *    Currently only byte ranges are supported
  44  *
  45  *    Essentially, there are three types of byte ranges:
  46  *
  47  *      1) first-byte-pos "-" last-byte-pos  // range
  48  *      2) first-byte-pos "-"                // trailer
  49  *      3)                "-" suffix-length  // suffix (last length bytes)
  50  *
  51  *
  52  *    When Range field is parsed, we have no clue about the content
  53  *    length of the document. Thus, we simply code an "absent" part
  54  *    using HttpHdrRangeSpec::UnknownPosition constant.
  55  *
  56  *    Note: when response length becomes known, we convert any range
  57  *    spec into type one above. (Canonization process).
  58  */
  59
  60
  61 /* local routines */
  62 #define known_spec(s) ((s) > HttpHdrRangeSpec::UnknownPosition)
  63
  64 /* globals */
  65 size_t HttpHdrRange::ParsedCount = 0;
  66 int64_t const HttpHdrRangeSpec::UnknownPosition = -1;
  67
  68 /*
  69  * Range-Spec
  70  */
  71
  72 HttpHdrRangeSpec::HttpHdrRangeSpec() : offset(UnknownPosition), length(UnknownPosition){}
  73
  74 /* parses range-spec and returns new object on success */
  75 HttpHdrRangeSpec *
  76 HttpHdrRangeSpec::Create(const char *field, int flen)
  77 {
  78     HttpHdrRangeSpec spec;
  79
  80     if (!spec.parseInit(field, flen))
  81         return NULL;
  82
  83     return new HttpHdrRangeSpec(spec);
  84 }
  85
  86 bool
  87 HttpHdrRangeSpec::parseInit(const char *field, int flen)
  88 {
  89     const char *p;
  90
  91     if (flen < 2)
  92         return false;
  93
  94     /* is it a suffix-byte-range-spec ? */
  95     if (*field == '-') {
  96         if (!httpHeaderParseOffset(field + 1, &length))
  97             return false;
  98     } else
  99         /* must have a '-' somewhere in _this_ field */
 100         if (!((p = strchr(field, '-')) || (p - field >= flen))) {
 101             debugs(64, 2, "ignoring invalid (missing '-') range-spec near: '" << field << "'");
 102             return false;
 103         } else {
 104             if (!httpHeaderParseOffset(field, &offset))
 105                 return false;
 106
 107             p++;
 108
 109             /* do we have last-pos ? */
 110             if (p - field < flen) {
 111                 int64_t last_pos;
 112
 113                 if (!httpHeaderParseOffset(p, &last_pos))
 114                     return false;
 115
 116                 HttpHdrRangeSpec::HttpRange aSpec (offset, last_pos + 1);
 117
 118                 length = aSpec.size();
 119             }
 120         }
 121
 122     /* we managed to parse, check if the result makes sence */
 123     if (length == 0) {
 124         debugs(64, 2, "ignoring invalid (zero length) range-spec near: '" << field << "'");
 125         return false;
 126     }
 127
 128     return true;
 129 }
 130
 131 void
 132 HttpHdrRangeSpec::packInto(Packer * packer) const
 133 {
 134     if (!known_spec(offset))    /* suffix */
 135         packerPrintf(packer, "-%"PRId64,  length);
 136     else if (!known_spec(length))               /* trailer */
 137         packerPrintf(packer, "%"PRId64"-", offset);
 138     else                        /* range */
 139         packerPrintf(packer, "%"PRId64"-%"PRId64,
 140                      offset, offset + length - 1);
 141 }
 142
 143 void
 144 HttpHdrRangeSpec::outputInfo( char const *note) const
 145 {
 146     debugs(64, 5, "HttpHdrRangeSpec::canonize: " << note << ": [" <<
 147            offset << ", " << offset + length <<
 148            ") len: " << length);
 149 }
 150
 151 /* fills "absent" positions in range specification based on response body size
 152  * returns true if the range is still valid
 153  * range is valid if its intersection with [0,length-1] is not empty
 154  */
 155 int
 156 HttpHdrRangeSpec::canonize(int64_t clen)
 157 {
 158     outputInfo ("have");
 159     HttpRange object(0, clen);
 160
 161     if (!known_spec(offset))    /* suffix */
 162     {
 163         assert(known_spec(length));
 164         offset = object.intersection(HttpRange (clen - length, clen)).start;
 165     } else if (!known_spec(length))             /* trailer */
 166     {
 167         assert(known_spec(offset));
 168         HttpRange newRange = object.intersection(HttpRange (offset, clen));
 169         length = newRange.size();
 170     }
 171     /* we have a "range" now, adjust length if needed */
 172     assert(known_spec(length));
 173
 174     assert(known_spec(offset));
 175
 176     HttpRange newRange = object.intersection (HttpRange (offset, offset + length));
 177
 178     length = newRange.size();
 179
 180     outputInfo ("done");
 181
 182     return length > 0;
 183 }
 184
 185 /* merges recepient with donor if possible; returns true on success
 186  * both specs must be canonized prior to merger, of course */
 187 bool
 188 HttpHdrRangeSpec::mergeWith(const HttpHdrRangeSpec * donor)
 189 {
 190     bool merged (false);
 191 #if MERGING_BREAKS_NOTHING
 192     /* Note: this code works, but some clients may not like its effects */
 193     int64_t rhs = offset + length;              /* no -1 ! */
 194     const int64_t donor_rhs = donor->offset + donor->length;    /* no -1 ! */
 195     assert(known_spec(offset));
 196     assert(known_spec(donor->offset));
 197     assert(length > 0);
 198     assert(donor->length > 0);
 199     /* do we have a left hand side overlap? */
 200
 201     if (donor->offset < offset && offset <= donor_rhs) {
 202         offset = donor->offset; /* decrease left offset */
 203         merged = 1;
 204     }
 205
 206     /* do we have a right hand side overlap? */
 207     if (donor->offset <= rhs && rhs < donor_rhs) {
 208         rhs = donor_rhs;        /* increase right offset */
 209         merged = 1;
 210     }
 211
 212     /* adjust length if offsets have been changed */
 213     if (merged) {
 214         assert(rhs > offset);
 215         length = rhs - offset;
 216     } else {
 217         /* does recepient contain donor? */
 218         merged =
 219             offset <= donor->offset && donor->offset < rhs;
 220     }
 221
 222 #endif
 223     return merged;
 224 }
 225
 226 /*
 227  * Range
 228  */
 229
 230 HttpHdrRange::HttpHdrRange () : clen (HttpHdrRangeSpec::UnknownPosition)
 231 {}
 232
 233 HttpHdrRange *
 234 HttpHdrRange::ParseCreate(const String * range_spec)
 235 {
 236     HttpHdrRange *r = new HttpHdrRange;
 237
 238     if (!r->parseInit(range_spec)) {
 239         delete r;
 240         r = NULL;
 241     }
 242
 243     return r;
 244 }
 245
 246 /* returns true if ranges are valid; inits HttpHdrRange */
 247 bool
 248 HttpHdrRange::parseInit(const String * range_spec)
 249 {
 250     const char *item;
 251     const char *pos = NULL;
 252     int ilen;
 253     int count = 0;
 254     assert(this && range_spec);
 255     ++ParsedCount;
 256     debugs(64, 8, "parsing range field: '" << range_spec->buf() << "'");
 257     /* check range type */
 258
 259     if (range_spec->caseCmp("bytes=", 6))
 260         return 0;
 261
 262     /* skip "bytes="; hack! */
 263     pos = range_spec->buf() + 5;
 264
 265     /* iterate through comma separated list */
 266     while (strListGetItem(range_spec, ',', &item, &ilen, &pos)) {
 267         HttpHdrRangeSpec *spec = HttpHdrRangeSpec::Create(item, ilen);
 268         /*
 269          * HTTP/1.1 draft says we must ignore the whole header field if one spec
 270          * is invalid. However, RFC 2068 just says that we must ignore that spec.
 271          */
 272
 273         if (spec)
 274             specs.push_back(spec);
 275
 276         ++count;
 277     }
 278
 279     debugs(64, 8, "parsed range range count: " << count << ", kept " <<
 280            specs.size());
 281     return specs.count != 0;
 282 }
 283
 284 HttpHdrRange::~HttpHdrRange()
 285 {
 286     while (specs.size())
 287         delete specs.pop_back();
 288 }
 289
 290 HttpHdrRange::HttpHdrRange(HttpHdrRange const &old) : specs()
 291 {
 292     specs.reserve(old.specs.size());
 293
 294     for (const_iterator i = old.begin(); i != old.end(); ++i)
 295         specs.push_back(new HttpHdrRangeSpec ( **i));
 296
 297     assert(old.specs.size() == specs.size());
 298 }
 299
 300 HttpHdrRange::iterator
 301 HttpHdrRange::begin()
 302 {
 303     return specs.begin();
 304 }
 305
 306 HttpHdrRange::iterator
 307 HttpHdrRange::end()
 308 {
 309     return specs.end();
 310 }
 311
 312 HttpHdrRange::const_iterator
 313 HttpHdrRange::begin() const
 314 {
 315     return specs.begin();
 316 }
 317
 318 HttpHdrRange::const_iterator
 319 HttpHdrRange::end() const
 320 {
 321     return specs.end();
 322 }
 323
 324 void
 325 HttpHdrRange::packInto(Packer * packer) const
 326 {
 327     const_iterator pos = begin();
 328     assert(this);
 329
 330     while (pos != end()) {
 331         if (pos != begin())
 332             packerAppend(packer, ",", 1);
 333
 334         (*pos)->packInto(packer);
 335
 336         ++pos;
 337     }
 338 }
 339
 340 void
 341 HttpHdrRange::merge (Vector<HttpHdrRangeSpec *> &basis)
 342 {
 343     /* reset old array */
 344     specs.clean();
 345     /* merge specs:
 346      * take one spec from "goods" and merge it with specs from
 347      * "specs" (if any) until there is no overlap */
 348     iterator i = basis.begin();
 349
 350     while (i != basis.end()) {
 351         if (specs.size() && (*i)->mergeWith(specs.back())) {
 352             /* merged with current so get rid of the prev one */
 353             delete specs.pop_back();
 354             continue;   /* re-iterate */
 355         }
 356
 357         specs.push_back (*i);
 358         ++i;                    /* progress */
 359     }
 360
 361     debugs(64, 3, "HttpHdrRange::merge: had " << basis.size() <<
 362            " specs, merged " << basis.size() - specs.size() << " specs");
 363 }
 364
 365
 366 void
 367 HttpHdrRange::getCanonizedSpecs (Vector<HttpHdrRangeSpec *> &copy)
 368 {
 369     /* canonize each entry and destroy bad ones if any */
 370
 371     for (iterator pos (begin()); pos != end(); ++pos) {
 372         if ((*pos)->canonize(clen))
 373             copy.push_back (*pos);
 374         else
 375             delete (*pos);
 376     }
 377
 378     debugs(64, 3, "HttpHdrRange::getCanonizedSpecs: found " <<
 379            specs.size() - copy.size() << " bad specs");
 380 }
 381
 382 #include "HttpHdrContRange.h"
 383
 384 /*
 385  * canonizes all range specs within a set preserving the order
 386  * returns true if the set is valid after canonization;
 387  * the set is valid if
 388  *   - all range specs are valid and
 389  *   - there is at least one range spec
 390  */
 391 int
 392 HttpHdrRange::canonize(HttpReply *rep)
 393 {
 394     assert(this && rep);
 395
 396     if (rep->content_range)
 397         clen = rep->content_range->elength;
 398     else
 399         clen = rep->content_length;
 400
 401     return canonize (clen);
 402 }
 403
 404 int
 405 HttpHdrRange::canonize (int64_t newClen)
 406 {
 407     clen = newClen;
 408     debugs(64, 3, "HttpHdrRange::canonize: started with " << specs.count <<
 409            " specs, clen: " << clen);
 410     Vector<HttpHdrRangeSpec*> goods;
 411     getCanonizedSpecs(goods);
 412     merge (goods);
 413     debugs(64, 3, "HttpHdrRange::canonize: finished with " << specs.count <<
 414            " specs");
 415     return specs.count > 0;
 416 }
 417
 418 /* hack: returns true if range specs are too "complex" for Squid to handle */
 419 /* requires that specs are "canonized" first! */
 420 bool
 421 HttpHdrRange::isComplex() const
 422 {
 423     int64_t offset = 0;
 424     assert(this);
 425     /* check that all rangers are in "strong" order */
 426     const_iterator pos (begin());
 427
 428     while (pos != end()) {
 429         /* Ensure typecasts is safe */
 430         assert ((*pos)->offset >= 0);
 431
 432         if ((*pos)->offset < offset)
 433             return 1;
 434
 435         offset = (*pos)->offset + (*pos)->length;
 436
 437         ++pos;
 438     }
 439
 440     return 0;
 441 }
 442
 443 /*
 444  * hack: returns true if range specs may be too "complex" when "canonized".
 445  * see also: HttpHdrRange::isComplex.
 446  */
 447 bool
 448 HttpHdrRange::willBeComplex() const
 449 {
 450     assert(this);
 451     /* check that all rangers are in "strong" order, */
 452     /* as far as we can tell without the content length */
 453     int64_t offset = 0;
 454
 455     for (const_iterator pos (begin()); pos != end(); ++pos) {
 456         if (!known_spec((*pos)->offset))        /* ignore unknowns */
 457             continue;
 458
 459         /* Ensure typecasts is safe */
 460         assert ((*pos)->offset >= 0);
 461
 462         if ((*pos)->offset < offset)
 463             return true;
 464
 465         offset = (*pos)->offset;
 466
 467         if (known_spec((*pos)->length)) /* avoid  unknowns */
 468             offset += (*pos)->length;
 469     }
 470
 471     return false;
 472 }
 473
 474 /*
 475  * Returns lowest known offset in range spec(s),
 476  * or HttpHdrRangeSpec::UnknownPosition
 477  * this is used for size limiting
 478  */
 479 int64_t
 480 HttpHdrRange::firstOffset() const
 481 {
 482     int64_t offset = HttpHdrRangeSpec::UnknownPosition;
 483     assert(this);
 484     const_iterator pos = begin();
 485
 486     while (pos != end()) {
 487         if ((*pos)->offset < offset || !known_spec(offset))
 488             offset = (*pos)->offset;
 489
 490         ++pos;
 491     }
 492
 493     return offset;
 494 }
 495
 496 /*
 497  * Returns lowest offset in range spec(s), 0 if unknown.
 498  * This is used for finding out where we need to start if all
 499  * ranges are combined into one, for example FTP REST.
 500  * Use 0 for size if unknown
 501  */
 502 int64_t
 503 HttpHdrRange::lowestOffset(int64_t size) const
 504 {
 505     int64_t offset = HttpHdrRangeSpec::UnknownPosition;
 506     const_iterator pos = begin();
 507     assert(this);
 508
 509     while (pos != end()) {
 510         int64_t current = (*pos)->offset;
 511
 512         if (!known_spec(current)) {
 513             if ((*pos)->length > size || !known_spec((*pos)->length))
 514                 return 0;       /* Unknown. Assume start of file */
 515
 516             current = size - (*pos)->length;
 517         }
 518
 519         if (current < offset || !known_spec(offset))
 520             offset = current;
 521
 522         ++pos;
 523     }
 524
 525     return known_spec(offset) ? offset : 0;
 526 }
 527
 528 /*
 529  * Return true if the first range offset is larger than the configured
 530  * limit.
 531  * Note that exceeding the limit (returning true) results in only
 532  * grabbing the needed range elements from the origin.
 533  */
 534 bool
 535 HttpHdrRange::offsetLimitExceeded() const
 536 {
 537     if (NULL == this)
 538         /* not a range request */
 539         return false;
 540
 541     if (-1 == Config.rangeOffsetLimit)
 542         /* disabled */
 543         return false;
 544
 545     if (firstOffset() == -1)
 546         /* tail request */
 547         return true;
 548
 549     if (Config.rangeOffsetLimit >= firstOffset())
 550         /* below the limit */
 551         return false;
 552
 553     return true;
 554 }
 555
 556 bool
 557 HttpHdrRange::contains(HttpHdrRangeSpec& r) const
 558 {
 559     assert(r.length >= 0);
 560     HttpHdrRangeSpec::HttpRange rrange(r.offset, r.offset + r.length);
 561
 562     for (const_iterator i = begin(); i != end(); ++i) {
 563         HttpHdrRangeSpec::HttpRange irange((*i)->offset, (*i)->offset + (*i)->length);
 564         HttpHdrRangeSpec::HttpRange intersection = rrange.intersection(irange);
 565
 566         if (intersection.start == irange.start && intersection.size() == irange.size())
 567             return true;
 568     }
 569
 570     return false;
 571 }
 572
 573 const HttpHdrRangeSpec *
 574 HttpHdrRangeIter::currentSpec() const
 575 {
 576     if (pos.incrementable())
 577         return *pos;
 578
 579     return NULL;
 580 }
 581
 582 void
 583 HttpHdrRangeIter::updateSpec()
 584 {
 585     assert (debt_size == 0);
 586     assert (valid);
 587
 588     if (pos.incrementable()) {
 589         debt(currentSpec()->length);
 590     }
 591 }
 592
 593 int64_t
 594 HttpHdrRangeIter::debt() const
 595 {
 596     debugs(64, 3, "HttpHdrRangeIter::debt: debt is " << debt_size);
 597     return debt_size;
 598 }
 599
 600 void HttpHdrRangeIter::debt(int64_t newDebt)
 601 {
 602     debugs(64, 3, "HttpHdrRangeIter::debt: was " << debt_size << " now " << newDebt);
 603     debt_size = newDebt;
 604 }