src/errorpage.cc

   1 /*
   2  * DEBUG: section 04    Error Generation
   3  * AUTHOR: Duane Wessels
   4  *
   5  * SQUID Web Proxy Cache          http://www.squid-cache.org/
   6  * ----------------------------------------------------------
   7  *
   8  *  Squid is the result of efforts by numerous individuals from
   9  *  the Internet community; see the CONTRIBUTORS file for full
  10  *  details.   Many organizations have provided support for Squid's
  11  *  development; see the SPONSORS file for full details.  Squid is
  12  *  Copyrighted (C) 2001 by the Regents of the University of
  13  *  California; see the COPYRIGHT file for full details.  Squid
  14  *  incorporates software developed and/or copyrighted by other
  15  *  sources; see the CREDITS file for full details.
  16  *
  17  *  This program is free software; you can redistribute it and/or modify
  18  *  it under the terms of the GNU General Public License as published by
  19  *  the Free Software Foundation; either version 2 of the License, or
  20  *  (at your option) any later version.
  21  *
  22  *  This program is distributed in the hope that it will be useful,
  23  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  24  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  25  *  GNU General Public License for more details.
  26  *
  27  *  You should have received a copy of the GNU General Public License
  28  *  along with this program; if not, write to the Free Software
  29  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
  30  *
  31  */
  32 #include "squid.h"
  33 #include "cache_cf.h"
  34 #include "comm/Connection.h"
  35 #include "comm/Write.h"
  36 #include "disk.h"
  37 #include "err_detail_type.h"
  38 #include "errorpage.h"
  39 #include "ftp.h"
  40 #include "Store.h"
  41 #include "html_quote.h"
  42 #include "HttpHeaderTools.h"
  43 #include "HttpReply.h"
  44 #include "HttpRequest.h"
  45 #include "MemObject.h"
  46 #include "fde.h"
  47 #include "MemBuf.h"
  48 #include "rfc1738.h"
  49 #include "SquidConfig.h"
  50 #include "URL.h"
  51 #include "URLScheme.h"
  52 #include "URL.h"
  53 #include "tools.h"
  54 #include "wordlist.h"
  55 #if USE_AUTH
  56 #include "auth/UserRequest.h"
  57 #endif
  58 #include "SquidTime.h"
  59 #if USE_SSL
  60 #include "ssl/ErrorDetailManager.h"
  61 #endif
  62
  63 /**
  64  \defgroup ErrorPageInternal Error Page Internals
  65  \ingroup ErrorPageAPI
  66  *
  67  \section Abstract Abstract:
  68  *   These routines are used to generate error messages to be
  69  *   sent to clients.  The error type is used to select between
  70  *   the various message formats. (formats are stored in the
  71  *   Config.errorDirectory)
  72  */
  73
  74 #if !defined(DEFAULT_SQUID_ERROR_DIR)
  75 /** Where to look for errors if config path fails.
  76  \note Please use ./configure --datadir=/path instead of patching
  77  */
  78 #define DEFAULT_SQUID_ERROR_DIR   DEFAULT_SQUID_DATA_DIR"/errors"
  79 #endif
  80
  81 /// \ingroup ErrorPageInternal
  82 CBDATA_CLASS_INIT(ErrorState);
  83
  84 /* local types */
  85
  86 /// \ingroup ErrorPageInternal
  87 typedef struct {
  88     int id;
  89     char *page_name;
  90     http_status page_redirect;
  91 } ErrorDynamicPageInfo;
  92
  93 /* local constant and vars */
  94
  95 /**
  96  \ingroup ErrorPageInternal
  97  *
  98  \note  hard coded error messages are not appended with %S
  99  *      automagically to give you more control on the format
 100  */
 101 static const struct {
 102     int type;                   /* and page_id */
 103     const char *text;
 104 }
 105
 106 error_hard_text[] = {
 107
 108     {
 109         ERR_SQUID_SIGNATURE,
 110         "\n<br>\n"
 111         "<hr>\n"
 112         "<div id=\"footer\">\n"
 113         "Generated %T by %h (%s)\n"
 114         "</div>\n"
 115         "</body></html>\n"
 116     },
 117     {
 118         TCP_RESET,
 119         "reset"
 120     }
 121 };
 122
 123 /// \ingroup ErrorPageInternal
 124 static Vector<ErrorDynamicPageInfo *> ErrorDynamicPages;
 125
 126 /* local prototypes */
 127
 128 /// \ingroup ErrorPageInternal
 129 static const int error_hard_text_count = sizeof(error_hard_text) / sizeof(*error_hard_text);
 130
 131 /// \ingroup ErrorPageInternal
 132 static char **error_text = NULL;
 133
 134 /// \ingroup ErrorPageInternal
 135 static int error_page_count = 0;
 136
 137 /// \ingroup ErrorPageInternal
 138 static MemBuf error_stylesheet;
 139
 140 static const char *errorFindHardText(err_type type);
 141 static ErrorDynamicPageInfo *errorDynamicPageInfoCreate(int id, const char *page_name);
 142 static void errorDynamicPageInfoDestroy(ErrorDynamicPageInfo * info);
 143 static IOCB errorSendComplete;
 144
 145 /// \ingroup ErrorPageInternal
 146 /// manages an error page template
 147 class ErrorPageFile: public TemplateFile
 148 {
 149 public:
 150     ErrorPageFile(const char *name, const err_type code): TemplateFile(name,code) { textBuf.init();}
 151
 152     /// The template text data read from disk
 153     const char *text() { return textBuf.content(); }
 154
 155 private:
 156     /// stores the data read from disk to a local buffer
 157     virtual bool parse(const char *buf, int len, bool eof) {
 158         if (len)
 159             textBuf.append(buf, len);
 160         return true;
 161     }
 162
 163     MemBuf textBuf; ///< A buffer to store the error page
 164 };
 165
 166 /// \ingroup ErrorPageInternal
 167 err_type &operator++ (err_type &anErr)
 168 {
 169     int tmp = (int)anErr;
 170     anErr = (err_type)(++tmp);
 171     return anErr;
 172 }
 173
 174 /// \ingroup ErrorPageInternal
 175 int operator - (err_type const &anErr, err_type const &anErr2)
 176 {
 177     return (int)anErr - (int)anErr2;
 178 }
 179
 180 void
 181 errorInitialize(void)
 182 {
 183     err_type i;
 184     const char *text;
 185     error_page_count = ERR_MAX + ErrorDynamicPages.size();
 186     error_text = static_cast<char **>(xcalloc(error_page_count, sizeof(char *)));
 187
 188     for (i = ERR_NONE, ++i; i < error_page_count; ++i) {
 189         safe_free(error_text[i]);
 190
 191         if ((text = errorFindHardText(i))) {
 192             /**\par
 193              * Index any hard-coded error text into defaults.
 194              */
 195             error_text[i] = xstrdup(text);
 196
 197         } else if (i < ERR_MAX) {
 198             /**\par
 199              * Index precompiled fixed template files from one of two sources:
 200              *  (a) default language translation directory (error_default_language)
 201              *  (b) admin specified custom directory (error_directory)
 202              */
 203             ErrorPageFile errTmpl(err_type_str[i], i);
 204             error_text[i] = errTmpl.loadDefault() ? xstrdup(errTmpl.text()) : NULL;
 205         } else {
 206             /** \par
 207              * Index any unknown file names used by deny_info.
 208              */
 209             ErrorDynamicPageInfo *info = ErrorDynamicPages.items[i - ERR_MAX];
 210             assert(info && info->id == i && info->page_name);
 211
 212             const char *pg = info->page_name;
 213             if (info->page_redirect != HTTP_STATUS_NONE)
 214                 pg = info->page_name +4;
 215
 216             if (strchr(pg, ':') == NULL) {
 217                 /** But only if they are not redirection URL. */
 218                 ErrorPageFile errTmpl(pg, ERR_MAX);
 219                 error_text[i] = errTmpl.loadDefault() ? xstrdup(errTmpl.text()) : NULL;
 220             }
 221         }
 222     }
 223
 224     error_stylesheet.reset();
 225
 226     // look for and load stylesheet into global MemBuf for it.
 227     if (Config.errorStylesheet) {
 228         ErrorPageFile tmpl("StylesSheet", ERR_MAX);
 229         tmpl.loadFromFile(Config.errorStylesheet);
 230         error_stylesheet.Printf("%s",tmpl.text());
 231     }
 232
 233 #if USE_SSL
 234     Ssl::errorDetailInitialize();
 235 #endif
 236 }
 237
 238 void
 239 errorClean(void)
 240 {
 241     if (error_text) {
 242         int i;
 243
 244         for (i = ERR_NONE + 1; i < error_page_count; ++i)
 245             safe_free(error_text[i]);
 246
 247         safe_free(error_text);
 248     }
 249
 250     while (ErrorDynamicPages.size())
 251         errorDynamicPageInfoDestroy(ErrorDynamicPages.pop_back());
 252
 253     error_page_count = 0;
 254
 255 #if USE_SSL
 256     Ssl::errorDetailClean();
 257 #endif
 258 }
 259
 260 /// \ingroup ErrorPageInternal
 261 static const char *
 262 errorFindHardText(err_type type)
 263 {
 264     int i;
 265
 266     for (i = 0; i < error_hard_text_count; ++i)
 267         if (error_hard_text[i].type == type)
 268             return error_hard_text[i].text;
 269
 270     return NULL;
 271 }
 272
 273 TemplateFile::TemplateFile(const char *name, const err_type code): silent(false), wasLoaded(false), templateName(name), templateCode(code)
 274 {
 275     assert(name);
 276 }
 277
 278 bool
 279 TemplateFile::loadDefault()
 280 {
 281     if (loaded()) // already loaded?
 282         return true;
 283
 284     /** test error_directory configured location */
 285     if (Config.errorDirectory) {
 286         char path[MAXPATHLEN];
 287         snprintf(path, sizeof(path), "%s/%s", Config.errorDirectory, templateName.termedBuf());
 288         loadFromFile(path);
 289     }
 290
 291 #if USE_ERR_LOCALES
 292     /** test error_default_language location */
 293     if (!loaded() && Config.errorDefaultLanguage) {
 294         if (!tryLoadTemplate(Config.errorDefaultLanguage)) {
 295             debugs(1, (templateCode < TCP_RESET ? DBG_CRITICAL : 3), "Unable to load default error language files. Reset to backups.");
 296         }
 297     }
 298 #endif
 299
 300     /* test default location if failed (templates == English translation base templates) */
 301     if (!loaded()) {
 302         tryLoadTemplate("templates");
 303     }
 304
 305     /* giving up if failed */
 306     if (!loaded()) {
 307         debugs(1, (templateCode < TCP_RESET ? DBG_CRITICAL : 3), "WARNING: failed to find or read error text file " << templateName);
 308         parse("Internal Error: Missing Template ", 33, '\0');
 309         parse(templateName.termedBuf(), templateName.size(), '\0');
 310     }
 311
 312     return true;
 313 }
 314
 315 bool
 316 TemplateFile::tryLoadTemplate(const char *lang)
 317 {
 318     assert(lang);
 319
 320     char path[MAXPATHLEN];
 321     /* TODO: prep the directory path string to prevent snprintf ... */
 322     snprintf(path, sizeof(path), "%s/%s/%s",
 323              DEFAULT_SQUID_ERROR_DIR, lang, templateName.termedBuf());
 324     path[MAXPATHLEN-1] = '\0';
 325
 326     if (loadFromFile(path))
 327         return true;
 328
 329 #if HAVE_GLOB
 330     if ( strlen(lang) == 2) {
 331         /* TODO glob the error directory for sub-dirs matching: <tag> '-*'   */
 332         /* use first result. */
 333         debugs(4,2, HERE << "wildcard fallback errors not coded yet.");
 334     }
 335 #endif
 336
 337     return false;
 338 }
 339
 340 bool
 341 TemplateFile::loadFromFile(const char *path)
 342 {
 343     int fd;
 344     char buf[4096];
 345     ssize_t len;
 346
 347     if (loaded()) // already loaded?
 348         return true;
 349
 350     fd = file_open(path, O_RDONLY | O_TEXT);
 351
 352     if (fd < 0) {
 353         /* with dynamic locale negotiation we may see some failures before a success. */
 354         if (!silent && templateCode < TCP_RESET)
 355             debugs(4, DBG_CRITICAL, HERE << "'" << path << "': " << xstrerror());
 356         wasLoaded = false;
 357         return wasLoaded;
 358     }
 359
 360     while ((len = FD_READ_METHOD(fd, buf, sizeof(buf))) > 0) {
 361         if (!parse(buf, len, false)) {
 362             debugs(4, DBG_CRITICAL, HERE << " parse error while reading template file: " << path);
 363             wasLoaded = false;
 364             return wasLoaded;
 365         }
 366     }
 367     parse(buf, 0, true);
 368
 369     if (len < 0) {
 370         debugs(4, DBG_CRITICAL, HERE << "failed to fully read: '" << path << "': " << xstrerror());
 371     }
 372
 373     file_close(fd);
 374
 375     wasLoaded = true;
 376     return wasLoaded;
 377 }
 378
 379 bool strHdrAcptLangGetItem(const String &hdr, char *lang, int langLen, size_t &pos)
 380 {
 381     while (pos < hdr.size()) {
 382         char *dt = lang;
 383
 384         if (!pos) {
 385             /* skip any initial whitespace. */
 386             while (pos < hdr.size() && xisspace(hdr[pos]))
 387                 ++pos;
 388         } else {
 389             // IFF we terminated the tag on whitespace or ';' we need to skip to the next ',' or end of header.
 390             while (pos < hdr.size() && hdr[pos] != ',')
 391                 ++pos;
 392             if (hdr[pos] == ',')
 393                 ++pos;
 394         }
 395
 396         /*
 397          * Header value format:
 398          *  - sequence of whitespace delimited tags
 399          *  - each tag may suffix with ';'.* which we can ignore.
 400          *  - IFF a tag contains only two characters we can wildcard ANY translations matching: <it> '-'? .*
 401          *    with preference given to an exact match.
 402          */
 403         bool invalid_byte = false;
 404         while (pos < hdr.size() && hdr[pos] != ';' && hdr[pos] != ',' && !xisspace(hdr[pos]) && dt < (lang + (langLen -1)) ) {
 405             if (!invalid_byte) {
 406 #if USE_HTTP_VIOLATIONS
 407                 // if accepting violations we may as well accept some broken browsers
 408                 //  which may send us the right code, wrong ISO formatting.
 409                 if (hdr[pos] == '_')
 410                     *dt = '-';
 411                 else
 412 #endif
 413                     *dt = xtolower(hdr[pos]);
 414                 // valid codes only contain A-Z, hyphen (-) and *
 415                 if (*dt != '-' && *dt != '*' && (*dt < 'a' || *dt > 'z') )
 416                     invalid_byte = true;
 417                 else
 418                     ++dt; // move to next destination byte.
 419             }
 420             ++pos;
 421         }
 422         *dt = '\0'; // nul-terminated the filename content string before system use.
 423         ++dt;
 424
 425         debugs(4, 9, HERE << "STATE: dt='" << dt << "', lang='" << lang << "', pos=" << pos << ", buf='" << ((pos < hdr.size()) ? hdr.substr(pos,hdr.size()) : "") << "'");
 426
 427         /* if we found anything we might use, try it. */
 428         if (*lang != '\0' && !invalid_byte)
 429             return true;
 430     }
 431     return false;
 432 }
 433
 434 bool
 435 TemplateFile::loadFor(HttpRequest *request)
 436 {
 437     String hdr;
 438
 439 #if USE_ERR_LOCALES
 440     if (loaded()) // already loaded?
 441         return true;
 442
 443     if (!request || !request->header.getList(HDR_ACCEPT_LANGUAGE, &hdr) )
 444         return false;
 445
 446     char lang[256];
 447     size_t pos = 0; // current parsing position in header string
 448
 449     debugs(4, 6, HERE << "Testing Header: '" << hdr << "'");
 450
 451     while ( strHdrAcptLangGetItem(hdr, lang, 256, pos) ) {
 452
 453         /* wildcard uses the configured default language */
 454         if (lang[0] == '*' && lang[1] == '\0') {
 455             debugs(4, 6, HERE << "Found language '" << lang << "'. Using configured default.");
 456             return false;
 457         }
 458
 459         debugs(4, 6, HERE << "Found language '" << lang << "', testing for available template");
 460
 461         if (tryLoadTemplate(lang)) {
 462             /* store the language we found for the Content-Language reply header */
 463             errLanguage = lang;
 464             break;
 465         } else if (Config.errorLogMissingLanguages) {
 466             debugs(4, DBG_IMPORTANT, "WARNING: Error Pages Missing Language: " << lang);
 467         }
 468     }
 469 #endif
 470
 471     return loaded();
 472 }
 473
 474 /// \ingroup ErrorPageInternal
 475 static ErrorDynamicPageInfo *
 476 errorDynamicPageInfoCreate(int id, const char *page_name)
 477 {
 478     ErrorDynamicPageInfo *info = new ErrorDynamicPageInfo;
 479     info->id = id;
 480     info->page_name = xstrdup(page_name);
 481     info->page_redirect = static_cast<http_status>(atoi(page_name));
 482
 483     /* WARNING on redirection status:
 484      * 2xx are permitted, but not documented officially.
 485      * - might be useful for serving static files (PAC etc) in special cases
 486      * 3xx require a URL suitable for Location: header.
 487      * - the current design does not allow for a Location: URI as well as a local file template
 488      *   although this possibility is explicitly permitted in the specs.
 489      * 4xx-5xx require a local file template.
 490      * - sending Location: on these codes with no body is invalid by the specs.
 491      * - current result is Squid crashing or XSS problems as dynamic deny_info load random disk files.
 492      * - a future redesign of the file loading may result in loading remote objects sent inline as local body.
 493      */
 494     if (info->page_redirect == HTTP_STATUS_NONE)
 495         ; // special case okay.
 496     else if (info->page_redirect < 200 || info->page_redirect > 599) {
 497         // out of range
 498         debugs(0, DBG_CRITICAL, "FATAL: status " << info->page_redirect << " is not valid on '" << page_name << "'");
 499         self_destruct();
 500     } else if ( /* >= 200 && */ info->page_redirect < 300 && strchr(&(page_name[4]), ':')) {
 501         // 2xx require a local template file
 502         debugs(0, DBG_CRITICAL, "FATAL: status " << info->page_redirect << " requires a template on '" << page_name << "'");
 503         self_destruct();
 504     } else if (info->page_redirect >= 300 && info->page_redirect <= 399 && !strchr(&(page_name[4]), ':')) {
 505         // 3xx require an absolute URL
 506         debugs(0, DBG_CRITICAL, "FATAL: status " << info->page_redirect << " requires a URL on '" << page_name << "'");
 507         self_destruct();
 508     } else if (info->page_redirect >= 400 /* && <= 599 */ && strchr(&(page_name[4]), ':')) {
 509         // 4xx/5xx require a local template file
 510         debugs(0, DBG_CRITICAL, "FATAL: status " << info->page_redirect << " requires a template on '" << page_name << "'");
 511         self_destruct();
 512     }
 513     // else okay.
 514
 515     return info;
 516 }
 517
 518 /// \ingroup ErrorPageInternal
 519 static void
 520 errorDynamicPageInfoDestroy(ErrorDynamicPageInfo * info)
 521 {
 522     assert(info);
 523     safe_free(info->page_name);
 524     delete info;
 525 }
 526
 527 /// \ingroup ErrorPageInternal
 528 static int
 529 errorPageId(const char *page_name)
 530 {
 531     for (int i = 0; i < ERR_MAX; ++i) {
 532         if (strcmp(err_type_str[i], page_name) == 0)
 533             return i;
 534     }
 535
 536     for (size_t j = 0; j < ErrorDynamicPages.size(); ++j) {
 537         if (strcmp(ErrorDynamicPages.items[j]->page_name, page_name) == 0)
 538             return j + ERR_MAX;
 539     }
 540
 541     return ERR_NONE;
 542 }
 543
 544 err_type
 545 errorReservePageId(const char *page_name)
 546 {
 547     ErrorDynamicPageInfo *info;
 548     int id = errorPageId(page_name);
 549
 550     if (id == ERR_NONE) {
 551         info = errorDynamicPageInfoCreate(ERR_MAX + ErrorDynamicPages.size(), page_name);
 552         ErrorDynamicPages.push_back(info);
 553         id = info->id;
 554     }
 555
 556     return (err_type)id;
 557 }
 558
 559 /// \ingroup ErrorPageInternal
 560 const char *
 561 errorPageName(int pageId)
 562 {
 563     if (pageId >= ERR_NONE && pageId < ERR_MAX)         /* common case */
 564         return err_type_str[pageId];
 565
 566     if (pageId >= ERR_MAX && pageId - ERR_MAX < (ssize_t)ErrorDynamicPages.size())
 567         return ErrorDynamicPages.items[pageId - ERR_MAX]->page_name;
 568
 569     return "ERR_UNKNOWN";       /* should not happen */
 570 }
 571
 572 ErrorState::ErrorState(err_type t, http_status status, HttpRequest * req) :
 573         type(t),
 574         page_id(t),
 575         err_language(NULL),
 576         httpStatus(status),
 577 #if USE_AUTH
 578         auth_user_request (NULL),
 579 #endif
 580         request(NULL),
 581         url(NULL),
 582         xerrno(0),
 583         port(0),
 584         dnsError(),
 585         ttl(0),
 586         src_addr(),
 587         redirect_url(NULL),
 588         callback(NULL),
 589         callback_data(NULL),
 590         request_hdrs(NULL),
 591         err_msg(NULL),
 592 #if USE_SSL
 593         detail(NULL),
 594 #endif
 595         detailCode(ERR_DETAIL_NONE)
 596 {
 597     memset(&flags, 0, sizeof(flags));
 598     memset(&ftp, 0, sizeof(ftp));
 599
 600     if (page_id >= ERR_MAX && ErrorDynamicPages.items[page_id - ERR_MAX]->page_redirect != HTTP_STATUS_NONE)
 601         httpStatus = ErrorDynamicPages.items[page_id - ERR_MAX]->page_redirect;
 602
 603     if (req != NULL) {
 604         request = HTTPMSGLOCK(req);
 605         src_addr = req->client_addr;
 606     }
 607 }
 608
 609 void
 610 errorAppendEntry(StoreEntry * entry, ErrorState * err)
 611 {
 612     assert(entry->mem_obj != NULL);
 613     assert (entry->isEmpty());
 614     debugs(4, 4, "Creating an error page for entry " << entry <<
 615            " with errorstate " << err <<
 616            " page id " << err->page_id);
 617
 618     if (entry->store_status != STORE_PENDING) {
 619         debugs(4, 2, "Skipping error page due to store_status: " << entry->store_status);
 620         /*
 621          * If the entry is not STORE_PENDING, then no clients
 622          * care about it, and we don't need to generate an
 623          * error message
 624          */
 625         assert(EBIT_TEST(entry->flags, ENTRY_ABORTED));
 626         assert(entry->mem_obj->nclients == 0);
 627         delete err;
 628         return;
 629     }
 630
 631     if (err->page_id == TCP_RESET) {
 632         if (err->request) {
 633             debugs(4, 2, "RSTing this reply");
 634             err->request->flags.resetTCP_=true;
 635         }
 636     }
 637
 638     entry->lock();
 639     entry->buffer();
 640     entry->replaceHttpReply( err->BuildHttpReply() );
 641     EBIT_CLR(entry->flags, ENTRY_FWD_HDR_WAIT);
 642     entry->flush();
 643     entry->complete();
 644     entry->negativeCache();
 645     entry->releaseRequest();
 646     entry->unlock();
 647     delete err;
 648 }
 649
 650 void
 651 errorSend(const Comm::ConnectionPointer &conn, ErrorState * err)
 652 {
 653     HttpReply *rep;
 654     debugs(4, 3, HERE << conn << ", err=" << err);
 655     assert(Comm::IsConnOpen(conn));
 656
 657     /* moved in front of errorBuildBuf @?@ */
 658     err->flags.flag_cbdata = 1;
 659
 660     rep = err->BuildHttpReply();
 661
 662     MemBuf *mb = rep->pack();
 663     AsyncCall::Pointer call = commCbCall(78, 5, "errorSendComplete",
 664                                          CommIoCbPtrFun(&errorSendComplete, err));
 665     Comm::Write(conn, mb, call);
 666     delete mb;
 667
 668     delete rep;
 669 }
 670
 671 /**
 672  \ingroup ErrorPageAPI
 673  *
 674  * Called by commHandleWrite() after data has been written
 675  * to the client socket.
 676  *
 677  \note If there is a callback, the callback is responsible for
 678  *     closing the FD, otherwise we do it ourselves.
 679  */
 680 static void
 681 errorSendComplete(const Comm::ConnectionPointer &conn, char *bufnotused, size_t size, comm_err_t errflag, int xerrno, void *data)
 682 {
 683     ErrorState *err = static_cast<ErrorState *>(data);
 684     debugs(4, 3, HERE << conn << ", size=" << size);
 685
 686     if (errflag != COMM_ERR_CLOSING) {
 687         if (err->callback) {
 688             debugs(4, 3, "errorSendComplete: callback");
 689             err->callback(conn->fd, err->callback_data, size);
 690         } else {
 691             debugs(4, 3, "errorSendComplete: comm_close");
 692             conn->close();
 693         }
 694     }
 695
 696     delete err;
 697 }
 698
 699 ErrorState::~ErrorState()
 700 {
 701     HTTPMSGUNLOCK(request);
 702     safe_free(redirect_url);
 703     safe_free(url);
 704     safe_free(request_hdrs);
 705     wordlistDestroy(&ftp.server_msg);
 706     safe_free(ftp.request);
 707     safe_free(ftp.reply);
 708 #if USE_AUTH
 709     auth_user_request = NULL;
 710 #endif
 711     safe_free(err_msg);
 712 #if USE_ERR_LOCALES
 713     if (err_language != Config.errorDefaultLanguage)
 714 #endif
 715         safe_free(err_language);
 716 #if USE_SSL
 717     delete detail;
 718 #endif
 719 }
 720
 721 int
 722 ErrorState::Dump(MemBuf * mb)
 723 {
 724     MemBuf str;
 725     char ntoabuf[MAX_IPSTRLEN];
 726
 727     str.reset();
 728     /* email subject line */
 729     str.Printf("CacheErrorInfo - %s", errorPageName(type));
 730     mb->Printf("?subject=%s", rfc1738_escape_part(str.buf));
 731     str.reset();
 732     /* email body */
 733     str.Printf("CacheHost: %s\r\n", getMyHostname());
 734     /* - Err Msgs */
 735     str.Printf("ErrPage: %s\r\n", errorPageName(type));
 736
 737     if (xerrno) {
 738         str.Printf("Err: (%d) %s\r\n", xerrno, strerror(xerrno));
 739     } else {
 740         str.Printf("Err: [none]\r\n");
 741     }
 742 #if USE_AUTH
 743     if (auth_user_request->denyMessage())
 744         str.Printf("Auth ErrMsg: %s\r\n", auth_user_request->denyMessage());
 745 #endif
 746     if (dnsError.size() > 0)
 747         str.Printf("DNS ErrMsg: %s\r\n", dnsError.termedBuf());
 748
 749     /* - TimeStamp */
 750     str.Printf("TimeStamp: %s\r\n\r\n", mkrfc1123(squid_curtime));
 751
 752     /* - IP stuff */
 753     str.Printf("ClientIP: %s\r\n", src_addr.NtoA(ntoabuf,MAX_IPSTRLEN));
 754
 755     if (request && request->hier.host[0] != '\0') {
 756         str.Printf("ServerIP: %s\r\n", request->hier.host);
 757     }
 758
 759     str.Printf("\r\n");
 760     /* - HTTP stuff */
 761     str.Printf("HTTP Request:\r\n");
 762
 763     if (NULL != request) {
 764         Packer pck;
 765         String urlpath_or_slash;
 766
 767         if (request->urlpath.size() != 0)
 768             urlpath_or_slash = request->urlpath;
 769         else
 770             urlpath_or_slash = "/";
 771
 772         str.Printf("%s " SQUIDSTRINGPH " %s/%d.%d\n",
 773                    RequestMethodStr(request->method),
 774                    SQUIDSTRINGPRINT(urlpath_or_slash),
 775                    AnyP::ProtocolType_str[request->http_ver.protocol],
 776                    request->http_ver.major, request->http_ver.minor);
 777         packerToMemInit(&pck, &str);
 778         request->header.packInto(&pck);
 779         packerClean(&pck);
 780     }
 781
 782     str.Printf("\r\n");
 783     /* - FTP stuff */
 784
 785     if (ftp.request) {
 786         str.Printf("FTP Request: %s\r\n", ftp.request);
 787         str.Printf("FTP Reply: %s\r\n", (ftp.reply? ftp.reply:"[none]"));
 788         str.Printf("FTP Msg: ");
 789         wordlistCat(ftp.server_msg, &str);
 790         str.Printf("\r\n");
 791     }
 792
 793     str.Printf("\r\n");
 794     mb->Printf("&body=%s", rfc1738_escape_part(str.buf));
 795     str.clean();
 796     return 0;
 797 }
 798
 799 /// \ingroup ErrorPageInternal
 800 #define CVT_BUF_SZ 512
 801
 802 const char *
 803 ErrorState::Convert(char token, bool building_deny_info_url, bool allowRecursion)
 804 {
 805     static MemBuf mb;
 806     const char *p = NULL;       /* takes priority over mb if set */
 807     int do_quote = 1;
 808     int no_urlescape = 0;       /* if true then item is NOT to be further URL-encoded */
 809     char ntoabuf[MAX_IPSTRLEN];
 810
 811     mb.reset();
 812
 813     switch (token) {
 814
 815     case 'a':
 816 #if USE_AUTH
 817         if (request && request->auth_user_request != NULL)
 818             p = request->auth_user_request->username();
 819         if (!p)
 820 #endif
 821             p = "-";
 822         break;
 823
 824     case 'b':
 825         mb.Printf("%d", getMyPort());
 826         break;
 827
 828     case 'B':
 829         if (building_deny_info_url) break;
 830         p = request ? ftpUrlWith2f(request) : "[no URL]";
 831         break;
 832
 833     case 'c':
 834         if (building_deny_info_url) break;
 835         p = errorPageName(type);
 836         break;
 837
 838     case 'D':
 839         if (!allowRecursion)
 840             p = "%D";  // if recursion is not allowed, do not convert
 841 #if USE_SSL
 842         // currently only SSL error details implemented
 843         else if (detail) {
 844             detail->useRequest(request);
 845             const String &errDetail = detail->toString();
 846             if (errDetail.defined()) {
 847                 MemBuf *detail_mb  = ConvertText(errDetail.termedBuf(), false);
 848                 mb.append(detail_mb->content(), detail_mb->contentSize());
 849                 delete detail_mb;
 850                 do_quote = 0;
 851             }
 852         }
 853 #endif
 854         if (!mb.contentSize())
 855             mb.Printf("[No Error Detail]");
 856         break;
 857
 858     case 'e':
 859         mb.Printf("%d", xerrno);
 860         break;
 861
 862     case 'E':
 863         if (xerrno)
 864             mb.Printf("(%d) %s", xerrno, strerror(xerrno));
 865         else
 866             mb.Printf("[No Error]");
 867         break;
 868
 869     case 'f':
 870         if (building_deny_info_url) break;
 871         /* FTP REQUEST LINE */
 872         if (ftp.request)
 873             p = ftp.request;
 874         else
 875             p = "nothing";
 876         break;
 877
 878     case 'F':
 879         if (building_deny_info_url) break;
 880         /* FTP REPLY LINE */
 881         if (ftp.reply)
 882             p = ftp.reply;
 883         else
 884             p = "nothing";
 885         break;
 886
 887     case 'g':
 888         if (building_deny_info_url) break;
 889         /* FTP SERVER RESPONSE */
 890         if (ftp.listing) {
 891             mb.append(ftp.listing->content(), ftp.listing->contentSize());
 892             do_quote = 0;
 893         } else if (ftp.server_msg) {
 894             wordlistCat(ftp.server_msg, &mb);
 895         }
 896         break;
 897
 898     case 'h':
 899         mb.Printf("%s", getMyHostname());
 900         break;
 901
 902     case 'H':
 903         if (request) {
 904             if (request->hier.host[0] != '\0') // if non-empty string.
 905                 p = request->hier.host;
 906             else
 907                 p = request->GetHost();
 908         } else if (!building_deny_info_url)
 909             p = "[unknown host]";
 910         break;
 911
 912     case 'i':
 913         mb.Printf("%s", src_addr.NtoA(ntoabuf,MAX_IPSTRLEN));
 914         break;
 915
 916     case 'I':
 917         if (request && request->hier.tcpServer != NULL)
 918             p = request->hier.tcpServer->remote.NtoA(ntoabuf,MAX_IPSTRLEN);
 919         else if (!building_deny_info_url)
 920             p = "[unknown]";
 921         break;
 922
 923     case 'l':
 924         if (building_deny_info_url) break;
 925         mb.append(error_stylesheet.content(), error_stylesheet.contentSize());
 926         do_quote = 0;
 927         break;
 928
 929     case 'L':
 930         if (building_deny_info_url) break;
 931         if (Config.errHtmlText) {
 932             mb.Printf("%s", Config.errHtmlText);
 933             do_quote = 0;
 934         } else
 935             p = "[not available]";
 936         break;
 937
 938     case 'm':
 939         if (building_deny_info_url) break;
 940 #if USE_AUTH
 941         p = auth_user_request->denyMessage("[not available]");
 942 #else
 943         p = "-";
 944 #endif
 945         break;
 946
 947     case 'M':
 948         if (request)
 949             p = RequestMethodStr(request->method);
 950         else if (!building_deny_info_url)
 951             p= "[unknown method]";
 952         break;
 953
 954     case 'o':
 955         p = request ? request->extacl_message.termedBuf() : external_acl_message;
 956         if (!p && !building_deny_info_url)
 957             p = "[not available]";
 958         break;
 959
 960     case 'p':
 961         if (request) {
 962             mb.Printf("%d", (int) request->port);
 963         } else if (!building_deny_info_url) {
 964             p = "[unknown port]";
 965         }
 966         break;
 967
 968     case 'P':
 969         if (request) {
 970             p = AnyP::ProtocolType_str[request->protocol];
 971         } else if (!building_deny_info_url) {
 972             p = "[unknown protocol]";
 973         }
 974         break;
 975
 976     case 'R':
 977         if (building_deny_info_url) {
 978             p = (request->urlpath.size() != 0 ? request->urlpath.termedBuf() : "/");
 979             no_urlescape = 1;
 980             break;
 981         }
 982         if (NULL != request) {
 983             Packer pck;
 984             String urlpath_or_slash;
 985
 986             if (request->urlpath.size() != 0)
 987                 urlpath_or_slash = request->urlpath;
 988             else
 989                 urlpath_or_slash = "/";
 990
 991             mb.Printf("%s " SQUIDSTRINGPH " %s/%d.%d\n",
 992                       RequestMethodStr(request->method),
 993                       SQUIDSTRINGPRINT(urlpath_or_slash),
 994                       AnyP::ProtocolType_str[request->http_ver.protocol],
 995                       request->http_ver.major, request->http_ver.minor);
 996             packerToMemInit(&pck, &mb);
 997             request->header.packInto(&pck, true); //hide authorization data
 998             packerClean(&pck);
 999         } else if (request_hdrs) {
1000             p = request_hdrs;
1001         } else {
1002             p = "[no request]";
1003         }
1004         break;
1005
1006     case 's':
1007         /* for backward compat we make %s show the full URL. Drop this in some future release. */
1008         if (building_deny_info_url) {
1009             p = request ? urlCanonical(request) : url;
1010             debugs(0, DBG_CRITICAL, "WARNING: deny_info now accepts coded tags. Use %u to get the full URL instead of %s");
1011         } else
1012             p = visible_appname_string;
1013         break;
1014
1015     case 'S':
1016         if (building_deny_info_url) {
1017             p = visible_appname_string;
1018             break;
1019         }
1020         /* signature may contain %-escapes, recursion */
1021         if (page_id != ERR_SQUID_SIGNATURE) {
1022             const int saved_id = page_id;
1023             page_id = ERR_SQUID_SIGNATURE;
1024             MemBuf *sign_mb = BuildContent();
1025             mb.Printf("%s", sign_mb->content());
1026             sign_mb->clean();
1027             delete sign_mb;
1028             page_id = saved_id;
1029             do_quote = 0;
1030         } else {
1031             /* wow, somebody put %S into ERR_SIGNATURE, stop recursion */
1032             p = "[%S]";
1033         }
1034         break;
1035
1036     case 't':
1037         mb.Printf("%s", Time::FormatHttpd(squid_curtime));
1038         break;
1039
1040     case 'T':
1041         mb.Printf("%s", mkrfc1123(squid_curtime));
1042         break;
1043
1044     case 'U':
1045         /* Using the fake-https version of canonical so error pages see https:// */
1046         /* even when the url-path cannot be shown as more than '*' */
1047         if (request)
1048             p = urlCanonicalFakeHttps(request);
1049         else if (url)
1050             p = url;
1051         else if (!building_deny_info_url)
1052             p = "[no URL]";
1053         break;
1054
1055     case 'u':
1056         if (request)
1057             p = urlCanonical(request);
1058         else if (url)
1059             p = url;
1060         else if (!building_deny_info_url)
1061             p = "[no URL]";
1062         break;
1063
1064     case 'w':
1065         if (Config.adminEmail)
1066             mb.Printf("%s", Config.adminEmail);
1067         else if (!building_deny_info_url)
1068             p = "[unknown]";
1069         break;
1070
1071     case 'W':
1072         if (building_deny_info_url) break;
1073         if (Config.adminEmail && Config.onoff.emailErrData)
1074             Dump(&mb);
1075         no_urlescape = 1;
1076         break;
1077
1078     case 'x':
1079 #if USE_SSL
1080         if (detail)
1081             mb.Printf("%s", detail->errorName());
1082         else
1083 #endif
1084             if (!building_deny_info_url)
1085                 p = "[Unknown Error Code]";
1086         break;
1087
1088     case 'z':
1089         if (building_deny_info_url) break;
1090         if (dnsError.size() > 0)
1091             p = dnsError.termedBuf();
1092         else if (ftp.cwd_msg)
1093             p = ftp.cwd_msg;
1094         else
1095             p = "[unknown]";
1096         break;
1097
1098     case 'Z':
1099         if (building_deny_info_url) break;
1100         if (err_msg)
1101             p = err_msg;
1102         else
1103             p = "[unknown]";
1104         break;
1105
1106     case '%':
1107         p = "%";
1108         break;
1109
1110     default:
1111         mb.Printf("%%%c", token);
1112         do_quote = 0;
1113         break;
1114     }
1115
1116     if (!p)
1117         p = mb.buf;             /* do not use mb after this assignment! */
1118
1119     assert(p);
1120
1121     debugs(4, 3, "errorConvert: %%" << token << " --> '" << p << "'" );
1122
1123     if (do_quote)
1124         p = html_quote(p);
1125
1126     if (building_deny_info_url && !no_urlescape)
1127         p = rfc1738_escape_part(p);
1128
1129     return p;
1130 }
1131
1132 void
1133 ErrorState::DenyInfoLocation(const char *name, HttpRequest *aRequest, MemBuf &result)
1134 {
1135     char const *m = name;
1136     char const *p = m;
1137     char const *t;
1138
1139     if (m[0] == '3')
1140         m += 4; // skip "3xx:"
1141
1142     while ((p = strchr(m, '%'))) {
1143         result.append(m, p - m);       /* copy */
1144         t = Convert(*++p, true, true);       /* convert */
1145         result.Printf("%s", t);        /* copy */
1146         m = p + 1;                     /* advance */
1147     }
1148
1149     if (*m)
1150         result.Printf("%s", m);        /* copy tail */
1151
1152     assert((size_t)result.contentSize() == strlen(result.content()));
1153 }
1154
1155 HttpReply *
1156 ErrorState::BuildHttpReply()
1157 {
1158     HttpReply *rep = new HttpReply;
1159     const char *name = errorPageName(page_id);
1160     /* no LMT for error pages; error pages expire immediately */
1161
1162     if (name[0] == '3' || (name[0] != '2' && name[0] != '4' && name[0] != '5' && strchr(name, ':'))) {
1163         /* Redirection */
1164         http_status status = HTTP_MOVED_TEMPORARILY;
1165         // Use configured 3xx reply status if set.
1166         if (name[0] == '3')
1167             status = httpStatus;
1168         else {
1169             // Use 307 for HTTP/1.1 non-GET/HEAD requests.
1170             if (request->method != METHOD_GET && request->method != METHOD_HEAD && request->http_ver >= HttpVersion(1,1))
1171                 status = HTTP_TEMPORARY_REDIRECT;
1172         }
1173
1174         rep->setHeaders(status, NULL, "text/html", 0, 0, -1);
1175
1176         if (request) {
1177             MemBuf redirect_location;
1178             redirect_location.init();
1179             DenyInfoLocation(name, request, redirect_location);
1180             httpHeaderPutStrf(&rep->header, HDR_LOCATION, "%s", redirect_location.content() );
1181         }
1182
1183         httpHeaderPutStrf(&rep->header, HDR_X_SQUID_ERROR, "%d %s", httpStatus, "Access Denied");
1184     } else {
1185         MemBuf *content = BuildContent();
1186         rep->setHeaders(httpStatus, NULL, "text/html", content->contentSize(), 0, -1);
1187         /*
1188          * include some information for downstream caches. Implicit
1189          * replaceable content. This isn't quite sufficient. xerrno is not
1190          * necessarily meaningful to another system, so we really should
1191          * expand it. Additionally, we should identify ourselves. Someone
1192          * might want to know. Someone _will_ want to know OTOH, the first
1193          * X-CACHE-MISS entry should tell us who.
1194          */
1195         httpHeaderPutStrf(&rep->header, HDR_X_SQUID_ERROR, "%s %d", name, xerrno);
1196
1197 #if USE_ERR_LOCALES
1198         /*
1199          * If error page auto-negotiate is enabled in any way, send the Vary.
1200          * RFC 2616 section 13.6 and 14.44 says MAY and SHOULD do this.
1201          * We have even better reasons though:
1202          * see http://wiki.squid-cache.org/KnowledgeBase/VaryNotCaching
1203          */
1204         if (!Config.errorDirectory) {
1205             /* We 'negotiated' this ONLY from the Accept-Language. */
1206             rep->header.delById(HDR_VARY);
1207             rep->header.putStr(HDR_VARY, "Accept-Language");
1208         }
1209
1210         /* add the Content-Language header according to RFC section 14.12 */
1211         if (err_language) {
1212             rep->header.putStr(HDR_CONTENT_LANGUAGE, err_language);
1213         } else
1214 #endif /* USE_ERROR_LOCALES */
1215         {
1216             /* default templates are in English */
1217             /* language is known unless error_directory override used */
1218             if (!Config.errorDirectory)
1219                 rep->header.putStr(HDR_CONTENT_LANGUAGE, "en");
1220         }
1221
1222         rep->body.setMb(content);
1223         /* do not memBufClean() or delete the content, it was absorbed by httpBody */
1224     }
1225
1226     // Make sure error codes get back to the client side for logging and
1227     // error tracking.
1228     if (request) {
1229         int edc = ERR_DETAIL_NONE; // error detail code
1230 #if USE_SSL
1231         if (detail)
1232             edc = detail->errorNo();
1233         else
1234 #endif
1235             if (detailCode)
1236                 edc = detailCode;
1237             else
1238                 edc = xerrno;
1239         request->detailError(type, edc);
1240     }
1241
1242     return rep;
1243 }
1244
1245 MemBuf *
1246 ErrorState::BuildContent()
1247 {
1248     const char *m = NULL;
1249
1250     assert(page_id > ERR_NONE && page_id < error_page_count);
1251
1252 #if USE_ERR_LOCALES
1253     ErrorPageFile *localeTmpl = NULL;
1254
1255     /** error_directory option in squid.conf overrides translations.
1256      * Custom errors are always found either in error_directory or the templates directory.
1257      * Otherwise locate the Accept-Language header
1258      */
1259     if (!Config.errorDirectory && page_id < ERR_MAX) {
1260         if (err_language && err_language != Config.errorDefaultLanguage)
1261             safe_free(err_language);
1262
1263         localeTmpl = new ErrorPageFile(err_type_str[page_id], static_cast<err_type>(page_id));
1264         if (localeTmpl->loadFor(request)) {
1265             m = localeTmpl->text();
1266             assert(localeTmpl->language());
1267             err_language = xstrdup(localeTmpl->language());
1268         }
1269     }
1270 #endif /* USE_ERR_LOCALES */
1271
1272     /** \par
1273      * If client-specific error templates are not enabled or available.
1274      * fall back to the old style squid.conf settings.
1275      */
1276     if (!m) {
1277         m = error_text[page_id];
1278 #if USE_ERR_LOCALES
1279         if (!Config.errorDirectory)
1280             err_language = Config.errorDefaultLanguage;
1281 #endif
1282         debugs(4, 2, HERE << "No existing error page language negotiated for " << errorPageName(page_id) << ". Using default error file.");
1283     }
1284
1285     MemBuf *result = ConvertText(m, true);
1286 #if USE_ERR_LOCALES
1287     if (localeTmpl)
1288         delete localeTmpl;
1289 #endif
1290     return result;
1291 }
1292
1293 MemBuf *ErrorState::ConvertText(const char *text, bool allowRecursion)
1294 {
1295     MemBuf *content = new MemBuf;
1296     const char *p;
1297     const char *m = text;
1298     assert(m);
1299     content->init();
1300
1301     while ((p = strchr(m, '%'))) {
1302         content->append(m, p - m);      /* copy */
1303         const char *t = Convert(*++p, false, allowRecursion);   /* convert */
1304         content->Printf("%s", t);       /* copy */
1305         m = p + 1;                      /* advance */
1306     }
1307
1308     if (*m)
1309         content->Printf("%s", m);       /* copy tail */
1310
1311     assert((size_t)content->contentSize() == strlen(content->content()));
1312
1313     return content;
1314 }