src/HttpHdrRange.cc

   1
   2 /*
   3  * DEBUG: section 64    HTTP Range Header
   4  * AUTHOR: Alex Rousskov
   5  *
   6  * SQUID Web Proxy Cache          http://www.squid-cache.org/
   7  * ----------------------------------------------------------
   8  *
   9  *  Squid is the result of efforts by numerous individuals from
  10  *  the Internet community; see the CONTRIBUTORS file for full
  11  *  details.   Many organizations have provided support for Squid's
  12  *  development; see the SPONSORS file for full details.  Squid is
  13  *  Copyrighted (C) 2001 by the Regents of the University of
  14  *  California; see the COPYRIGHT file for full details.  Squid
  15  *  incorporates software developed and/or copyrighted by other
  16  *  sources; see the CREDITS file for full details.
  17  *
  18  *  This program is free software; you can redistribute it and/or modify
  19  *  it under the terms of the GNU General Public License as published by
  20  *  the Free Software Foundation; either version 2 of the License, or
  21  *  (at your option) any later version.
  22  *
  23  *  This program is distributed in the hope that it will be useful,
  24  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  25  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  26  *  GNU General Public License for more details.
  27  *
  28  *  You should have received a copy of the GNU General Public License
  29  *  along with this program; if not, write to the Free Software
  30  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
  31  *
  32  */
  33
  34 #include "squid.h"
  35 #include "Store.h"
  36 #include "HttpHeaderRange.h"
  37 #include "client_side_request.h"
  38 #include "HttpReply.h"
  39 #include "HttpHeaderTools.h"
  40 #include "StrList.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 /* local routines */
  61 #define known_spec(s) ((s) > HttpHdrRangeSpec::UnknownPosition)
  62
  63 /* globals */
  64 size_t HttpHdrRange::ParsedCount = 0;
  65 int64_t const HttpHdrRangeSpec::UnknownPosition = -1;
  66
  67 /*
  68  * Range-Spec
  69  */
  70
  71 HttpHdrRangeSpec::HttpHdrRangeSpec() : offset(UnknownPosition), length(UnknownPosition) {}
  72
  73 /* parses range-spec and returns new object on success */
  74 HttpHdrRangeSpec *
  75 HttpHdrRangeSpec::Create(const char *field, int flen)
  76 {
  77     HttpHdrRangeSpec spec;
  78
  79     if (!spec.parseInit(field, flen))
  80         return NULL;
  81
  82     return new HttpHdrRangeSpec(spec);
  83 }
  84
  85 bool
  86 HttpHdrRangeSpec::parseInit(const char *field, int flen)
  87 {
  88     const char *p;
  89
  90     if (flen < 2)
  91         return false;
  92
  93     /* is it a suffix-byte-range-spec ? */
  94     if (*field == '-') {
  95         if (!httpHeaderParseOffset(field + 1, &length))
  96             return false;
  97     } else
  98         /* must have a '-' somewhere in _this_ field */
  99         if (!((p = strchr(field, '-')) && (p - field < flen))) {
 100             debugs(64, 2, "invalid (missing '-') range-spec near: '" << field << "'");
 101             return false;
 102         } else {
 103             if (!httpHeaderParseOffset(field, &offset))
 104                 return false;
 105
 106             ++p;
 107
 108             /* do we have last-pos ? */
 109             if (p - field < flen) {
 110                 int64_t last_pos;
 111
 112                 if (!httpHeaderParseOffset(p, &last_pos))
 113                     return false;
 114
 115                 // RFC 2616 s14.35.1 MUST: last-byte-pos >= first-byte-pos
 116                 if (last_pos < offset) {
 117                     debugs(64, 2, "invalid (last-byte-pos < first-byte-pos) range-spec near: " << field);
 118                     return false;
 119                 }
 120
 121                 HttpHdrRangeSpec::HttpRange aSpec (offset, last_pos + 1);
 122
 123                 length = aSpec.size();
 124             }
 125         }
 126
 127     return true;
 128 }
 129
 130 void
 131 HttpHdrRangeSpec::packInto(Packer * packer) const
 132 {
 133     if (!known_spec(offset))    /* suffix */
 134         packerPrintf(packer, "-%" PRId64,  length);
 135     else if (!known_spec(length))               /* trailer */
 136         packerPrintf(packer, "%" PRId64 "-", offset);
 137     else                        /* range */
 138         packerPrintf(packer, "%" PRId64 "-%" PRId64,
 139                      offset, offset + length - 1);
 140 }
 141
 142 void
 143 HttpHdrRangeSpec::outputInfo( char const *note) const
 144 {
 145     debugs(64, 5, "HttpHdrRangeSpec::canonize: " << note << ": [" <<
 146            offset << ", " << offset + length <<
 147            ") len: " << length);
 148 }
 149
 150 /* fills "absent" positions in range specification based on response body size
 151  * returns true if the range is still valid
 152  * range is valid if its intersection with [0,length-1] is not empty
 153  */
 154 int
 155 HttpHdrRangeSpec::canonize(int64_t clen)
 156 {
 157     outputInfo ("have");
 158     HttpRange object(0, clen);
 159
 160     if (!known_spec(offset)) {  /* suffix */
 161         assert(known_spec(length));
 162         offset = object.intersection(HttpRange (clen - length, clen)).start;
 163     } else if (!known_spec(length)) {   /* trailer */
 164         assert(known_spec(offset));
 165         HttpRange newRange = object.intersection(HttpRange (offset, clen));
 166         length = newRange.size();
 167     }
 168     /* we have a "range" now, adjust length if needed */
 169     assert(known_spec(length));
 170
 171     assert(known_spec(offset));
 172
 173     HttpRange newRange = object.intersection (HttpRange (offset, offset + length));
 174
 175     length = newRange.size();
 176
 177     outputInfo ("done");
 178
 179     return length > 0;
 180 }
 181
 182 /* merges recepient with donor if possible; returns true on success
 183  * both specs must be canonized prior to merger, of course */
 184 bool
 185 HttpHdrRangeSpec::mergeWith(const HttpHdrRangeSpec * donor)
 186 {
 187     bool merged (false);
 188 #if MERGING_BREAKS_NOTHING
 189     /* Note: this code works, but some clients may not like its effects */
 190     int64_t rhs = offset + length;              /* no -1 ! */
 191     const int64_t donor_rhs = donor->offset + donor->length;    /* no -1 ! */
 192     assert(known_spec(offset));
 193     assert(known_spec(donor->offset));
 194     assert(length > 0);
 195     assert(donor->length > 0);
 196     /* do we have a left hand side overlap? */
 197
 198     if (donor->offset < offset && offset <= donor_rhs) {
 199         offset = donor->offset; /* decrease left offset */
 200         merged = 1;
 201     }
 202
 203     /* do we have a right hand side overlap? */
 204     if (donor->offset <= rhs && rhs < donor_rhs) {
 205         rhs = donor_rhs;        /* increase right offset */
 206         merged = 1;
 207     }
 208
 209     /* adjust length if offsets have been changed */
 210     if (merged) {
 211         assert(rhs > offset);
 212         length = rhs - offset;
 213     } else {
 214         /* does recepient contain donor? */
 215         merged =
 216             offset <= donor->offset && donor->offset < rhs;
 217     }
 218
 219 #endif
 220     return merged;
 221 }
 222
 223 /*
 224  * Range
 225  */
 226
 227 HttpHdrRange::HttpHdrRange() : clen(HttpHdrRangeSpec::UnknownPosition)
 228 {}
 229
 230 HttpHdrRange *
 231 HttpHdrRange::ParseCreate(const String * range_spec)
 232 {
 233     HttpHdrRange *r = new HttpHdrRange;
 234
 235     if (!r->parseInit(range_spec)) {
 236         delete r;
 237         r = NULL;
 238     }
 239
 240     return r;
 241 }
 242
 243 /* returns true if ranges are valid; inits HttpHdrRange */
 244 bool
 245 HttpHdrRange::parseInit(const String * range_spec)
 246 {
 247     const char *item;
 248     const char *pos = NULL;
 249     int ilen;
 250     assert(this && range_spec);
 251     ++ParsedCount;
 252     debugs(64, 8, "parsing range field: '" << range_spec << "'");
 253     /* check range type */
 254
 255     if (range_spec->caseCmp("bytes=", 6))
 256         return 0;
 257
 258     /* skip "bytes="; hack! */
 259     pos = range_spec->termedBuf() + 6;
 260
 261     /* iterate through comma separated list */
 262     while (strListGetItem(range_spec, ',', &item, &ilen, &pos)) {
 263         HttpHdrRangeSpec *spec = HttpHdrRangeSpec::Create(item, ilen);
 264         /*
 265          * RFC 2616 section 14.35.1: MUST ignore Range with
 266          * at least one syntactically invalid byte-range-specs.
 267          */
 268         if (!spec) {
 269             while (!specs.empty())
 270                 delete specs.pop_back();
 271             debugs(64, 2, "ignoring invalid range field: '" << range_spec << "'");
 272             break;
 273         }
 274
 275         specs.push_back(spec);
 276     }
 277
 278     debugs(64, 8, "got range specs: " << specs.size());
 279     return !specs.empty();
 280 }
 281
 282 HttpHdrRange::~HttpHdrRange()
 283 {
 284     while (specs.size())
 285         delete specs.pop_back();
 286 }
 287
 288 HttpHdrRange::HttpHdrRange(HttpHdrRange const &old) :
 289         specs(),
 290         clen(HttpHdrRangeSpec::UnknownPosition)
 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 void
 366 HttpHdrRange::getCanonizedSpecs (Vector<HttpHdrRangeSpec *> &copy)
 367 {
 368     /* canonize each entry and destroy bad ones if any */
 369
 370     for (iterator pos (begin()); pos != end(); ++pos) {
 371         if ((*pos)->canonize(clen))
 372             copy.push_back (*pos);
 373         else
 374             delete (*pos);
 375     }
 376
 377     debugs(64, 3, "HttpHdrRange::getCanonizedSpecs: found " <<
 378            specs.size() - copy.size() << " bad specs");
 379 }
 380
 381 #include "HttpHdrContRange.h"
 382
 383 /*
 384  * canonizes all range specs within a set preserving the order
 385  * returns true if the set is valid after canonization;
 386  * the set is valid if
 387  *   - all range specs are valid and
 388  *   - there is at least one range spec
 389  */
 390 int
 391 HttpHdrRange::canonize(HttpReply *rep)
 392 {
 393     assert(this && rep);
 394
 395     if (rep->content_range)
 396         clen = rep->content_range->elength;
 397     else
 398         clen = rep->content_length;
 399
 400     return canonize (clen);
 401 }
 402
 403 int
 404 HttpHdrRange::canonize (int64_t newClen)
 405 {
 406     clen = newClen;
 407     debugs(64, 3, "HttpHdrRange::canonize: started with " << specs.count <<
 408            " specs, clen: " << clen);
 409     Vector<HttpHdrRangeSpec*> goods;
 410     getCanonizedSpecs(goods);
 411     merge (goods);
 412     debugs(64, 3, "HttpHdrRange::canonize: finished with " << specs.count <<
 413            " specs");
 414     return specs.count > 0;
 415 }
 416
 417 /* hack: returns true if range specs are too "complex" for Squid to handle */
 418 /* requires that specs are "canonized" first! */
 419 bool
 420 HttpHdrRange::isComplex() const
 421 {
 422     int64_t offset = 0;
 423     assert(this);
 424     /* check that all rangers are in "strong" order */
 425     const_iterator pos (begin());
 426
 427     while (pos != end()) {
 428         /* Ensure typecasts is safe */
 429         assert ((*pos)->offset >= 0);
 430
 431         if ((*pos)->offset < offset)
 432             return 1;
 433
 434         offset = (*pos)->offset + (*pos)->length;
 435
 436         ++pos;
 437     }
 438
 439     return 0;
 440 }
 441
 442 /*
 443  * hack: returns true if range specs may be too "complex" when "canonized".
 444  * see also: HttpHdrRange::isComplex.
 445  */
 446 bool
 447 HttpHdrRange::willBeComplex() const
 448 {
 449     assert(this);
 450     /* check that all rangers are in "strong" order, */
 451     /* as far as we can tell without the content length */
 452     int64_t offset = 0;
 453
 454     for (const_iterator pos (begin()); pos != end(); ++pos) {
 455         if (!known_spec((*pos)->offset))        /* ignore unknowns */
 456             continue;
 457
 458         /* Ensure typecasts is safe */
 459         assert ((*pos)->offset >= 0);
 460
 461         if ((*pos)->offset < offset)
 462             return true;
 463
 464         offset = (*pos)->offset;
 465
 466         if (known_spec((*pos)->length)) /* avoid  unknowns */
 467             offset += (*pos)->length;
 468     }
 469
 470     return false;
 471 }
 472
 473 /*
 474  * Returns lowest known offset in range spec(s),
 475  * or HttpHdrRangeSpec::UnknownPosition
 476  * this is used for size limiting
 477  */
 478 int64_t
 479 HttpHdrRange::firstOffset() const
 480 {
 481     int64_t offset = HttpHdrRangeSpec::UnknownPosition;
 482     assert(this);
 483     const_iterator pos = begin();
 484
 485     while (pos != end()) {
 486         if ((*pos)->offset < offset || !known_spec(offset))
 487             offset = (*pos)->offset;
 488
 489         ++pos;
 490     }
 491
 492     return offset;
 493 }
 494
 495 /*
 496  * Returns lowest offset in range spec(s), 0 if unknown.
 497  * This is used for finding out where we need to start if all
 498  * ranges are combined into one, for example FTP REST.
 499  * Use 0 for size if unknown
 500  */
 501 int64_t
 502 HttpHdrRange::lowestOffset(int64_t size) const
 503 {
 504     int64_t offset = HttpHdrRangeSpec::UnknownPosition;
 505     const_iterator pos = begin();
 506     assert(this);
 507
 508     while (pos != end()) {
 509         int64_t current = (*pos)->offset;
 510
 511         if (!known_spec(current)) {
 512             if ((*pos)->length > size || !known_spec((*pos)->length))
 513                 return 0;       /* Unknown. Assume start of file */
 514
 515             current = size - (*pos)->length;
 516         }
 517
 518         if (current < offset || !known_spec(offset))
 519             offset = current;
 520
 521         ++pos;
 522     }
 523
 524     return known_spec(offset) ? offset : 0;
 525 }
 526
 527 /*
 528  * \retval true   Fetch only requested ranges. The first range is larger that configured limit.
 529  * \retval false  Full download. Not a range request, no limit, or the limit is not yet reached.
 530  */
 531 bool
 532 HttpHdrRange::offsetLimitExceeded(const int64_t limit) const
 533 {
 534     if (NULL == this)
 535         /* not a range request */
 536         return false;
 537
 538     if (limit == 0)
 539         /* 0 == disabled */
 540         return true;
 541
 542     if (-1 == limit)
 543         /* 'none' == forced */
 544         return false;
 545
 546     if (firstOffset() == -1)
 547         /* tail request */
 548         return true;
 549
 550     if (limit >= firstOffset())
 551         /* below the limit */
 552         return false;
 553
 554     return true;
 555 }
 556
 557 bool
 558 HttpHdrRange::contains(HttpHdrRangeSpec& r) const
 559 {
 560     assert(r.length >= 0);
 561     HttpHdrRangeSpec::HttpRange rrange(r.offset, r.offset + r.length);
 562
 563     for (const_iterator i = begin(); i != end(); ++i) {
 564         HttpHdrRangeSpec::HttpRange irange((*i)->offset, (*i)->offset + (*i)->length);
 565         HttpHdrRangeSpec::HttpRange intersection = rrange.intersection(irange);
 566
 567         if (intersection.start == irange.start && intersection.size() == irange.size())
 568             return true;
 569     }
 570
 571     return false;
 572 }
 573
 574 const HttpHdrRangeSpec *
 575 HttpHdrRangeIter::currentSpec() const
 576 {
 577     if (pos.incrementable())
 578         return *pos;
 579
 580     return NULL;
 581 }
 582
 583 void
 584 HttpHdrRangeIter::updateSpec()
 585 {
 586     assert (debt_size == 0);
 587     assert (valid);
 588
 589     if (pos.incrementable()) {
 590         debt(currentSpec()->length);
 591     }
 592 }
 593
 594 int64_t
 595 HttpHdrRangeIter::debt() const
 596 {
 597     debugs(64, 3, "HttpHdrRangeIter::debt: debt is " << debt_size);
 598     return debt_size;
 599 }
 600
 601 void HttpHdrRangeIter::debt(int64_t newDebt)
 602 {
 603     debugs(64, 3, "HttpHdrRangeIter::debt: was " << debt_size << " now " << newDebt);
 604     debt_size = newDebt;
 605 }