Annotation of libwww/Library/src/HTBind.c, revision 2.18

2.1       frystyk     1: /*                                                                  HTBind.c
                      2: **     FILE SUFFIX BIND MANAGER
                      3: **
                      4: **     (c) COPYRIGHT MIT 1995
                      5: **     Please first read the full copyright statement in the file COPYRIGH.
                      6: **
                      7: **     This module sets up the binding between a file Bind and a media
                      8: **     type, language, encoding etc. In a client application the Binds
                      9: **     are used in protocols that does not support media types etc., like
                     10: **     FTP, and in server applications they are used to make the bindings
                     11: **     between the server and the local file store that the server can
                     12: **     serve to the rest of the world (well almost). The HTFormat module
                     13: **     holds this information against the accept headers received in a
                     14: **     request and uses if for format negotiation. All the binding management
                     15: **     can all be replace by a database interface. 
                     16: **
                     17: ** History:
                     18: **        Feb 91       Written Tim Berners-Lee CERN/CN
                     19: **        Apr 91       vms-vms access included using DECnet syntax
                     20: **     26 Jun 92 (JFG) When running over DECnet, suppressed FTP.
                     21: **                     Fixed access bug for relative names on VMS.
                     22: **        Sep 93 (MD)  Access to VMS files allows sharing.
                     23: **     15 Nov 93 (MD)  Moved HTVMSname to HTVMSUTILS.C
                     24: **     22 Feb 94 (MD)  Excluded two routines if we are not READING directories
                     25: **     18 May 94 (HF)  Directory stuff removed and stream handling updated,
                     26: **                     error messages introduced etc.
                     27: **     10 Maj 95 HF    Spawned off from HTFile in order to make it easier to
                     28: **                     override by a new module. It's now based on anchors
                     29: **                     and hash tables
                     30: ** Bugs:
                     31: */
                     32: 
                     33: /* Library Includes */
                     34: #include "tcp.h"
                     35: #include "HTUtils.h"
                     36: #include "HTString.h"
2.16      frystyk    37: #include "HTAncMan.h"
2.1       frystyk    38: #include "HTAtom.h"
                     39: #include "HTParse.h"
                     40: #include "HTBind.h"                                     /* Implemented here */
                     41: 
                     42: typedef struct _HTBind {
                     43:     char *     suffix;
                     44:     HTFormat   type;                   /* Content-Type */
                     45:     HTEncoding encoding;               /* Content-Encoding */
                     46:     HTLanguage language;               /* Content-Language */
                     47:     double     quality;
                     48: } HTBind;
                     49: 
                     50: #define HASH_SIZE      101        /* Arbitrary prime. Memory/speed tradeoff */
                     51: 
                     52: /* Suffix registration */
                     53: PRIVATE BOOL HTCaseSen = YES;                /* Are suffixes case sensitive */
                     54: PRIVATE char *HTDelimiters = NULL;                       /* Set of suffixes */
                     55: 
                     56: PRIVATE HTList **HTBindings = NULL;   /* Point to table of lists of bindings */
                     57: 
2.15      frystyk    58: PRIVATE HTBind no_suffix = { "*", NULL, NULL, NULL, 0.5 };
                     59: PRIVATE HTBind unknown_suffix = { "*.*", NULL, NULL, NULL, 0.5 };
2.1       frystyk    60: 
                     61: /* ------------------------------------------------------------------------- */
                     62: 
                     63: /*     
                     64: **     Set up the list of suffix bindings. Done by HTLibInit
                     65: */
2.11      frystyk    66: PUBLIC BOOL HTBind_init (void)
2.1       frystyk    67: {
2.8       frystyk    68:     if (!HTBindings) {
2.17      frystyk    69:        if ((HTBindings = (HTList* *) HT_CALLOC(HASH_SIZE, sizeof(HTList *))) == NULL)
                     70:            HT_OUTOFMEM("HTBind_init");
2.8       frystyk    71:     }
2.1       frystyk    72:     StrAllocCopy(HTDelimiters, DEFAULT_SUFFIXES);
2.15      frystyk    73:     no_suffix.type = WWW_UNKNOWN;
                     74:     no_suffix.encoding = WWW_ENC_BINARY;
                     75:     unknown_suffix.type = WWW_UNKNOWN;
                     76:     unknown_suffix.encoding = WWW_ENC_BINARY;
2.1       frystyk    77:     return YES;
                     78: }
                     79: 
                     80: 
                     81: /*
                     82: **     Cleans up the memory allocated by file bindings
                     83: **     Done by HTLibTerminate().
                     84: **     Written by Eric Sink, eric@spyglass.com, and Henrik
                     85: */
2.11      frystyk    86: PUBLIC BOOL HTBind_deleteAll (void)
2.1       frystyk    87: {
                     88:     int cnt;
                     89:     HTList *cur;
                     90:     if (!HTBindings)
                     91:        return NO;
                     92:     for (cnt=0; cnt<HASH_SIZE; cnt++) {
                     93:        if ((cur = HTBindings[cnt])) { 
                     94:            HTBind *pres;
                     95:            while ((pres = (HTBind *) HTList_nextObject(cur)) != NULL) {
2.17      frystyk    96:                HT_FREE(pres->suffix);
                     97:                HT_FREE(pres);
2.1       frystyk    98:            }
                     99:        }
                    100:        HTList_delete(HTBindings[cnt]);
                    101:        HTBindings[cnt] = NULL;
                    102:     }
2.17      frystyk   103:     HT_FREE(HTDelimiters);
2.1       frystyk   104:     return YES;
                    105: }
                    106: 
                    107: 
                    108: /*     Make suffix bindings case sensitive
                    109: **     -----------------------------------
                    110: */
2.11      frystyk   111: PUBLIC void HTBind_caseSensitive (BOOL sensitive)
2.1       frystyk   112: {
                    113:     HTCaseSen = sensitive;
                    114: }
                    115: 
                    116: 
                    117: /*     Get set of suffixes
                    118: **     -------------------
                    119: */
2.11      frystyk   120: PUBLIC CONST char *HTBind_delimiters (void)
2.1       frystyk   121: {
                    122:     return HTDelimiters;
                    123: }
                    124: 
                    125: 
                    126: /*     Change set of suffixes
                    127: **     ----------------------
                    128: */
2.11      frystyk   129: PUBLIC void HTBind_setDelimiters (CONST char * new_suffixes)
2.1       frystyk   130: {
                    131:     if (new_suffixes && *new_suffixes)
                    132:        StrAllocCopy(HTDelimiters, new_suffixes);
                    133: }
                    134: 
                    135: 
                    136: /*     Define the representation associated with a file suffix
                    137: **     -------------------------------------------------------
                    138: **
                    139: **     Calling this with suffix set to "*" will set the default
                    140: **     representation.
                    141: **     Calling this with suffix set to "*.*" will set the default
                    142: **     representation for unknown suffix files which contain a "."
                    143: **
                    144: **     If filename suffix is already defined its previous
                    145: **     definition is overridden (or modified)
                    146: */
2.12      frystyk   147: PUBLIC BOOL HTBind_addType (CONST char *       suffix,
                    148:                            CONST char *        representation,
                    149:                            double              value)
2.1       frystyk   150: {
2.11      frystyk   151:     return HTBind_add(suffix, representation, NULL, NULL, value);
2.1       frystyk   152: }
                    153: 
2.11      frystyk   154: PUBLIC BOOL HTBind_addEncoding (CONST char *   suffix,
                    155:                                CONST char *    encoding,
                    156:                                double          value)
2.1       frystyk   157: {
2.11      frystyk   158:     return HTBind_add(suffix, NULL, encoding, NULL, value);
2.1       frystyk   159: }
                    160: 
2.11      frystyk   161: PUBLIC BOOL HTBind_addLanguage (CONST char *   suffix,
                    162:                                CONST char *    language,
                    163:                                double          value)
2.1       frystyk   164: {
2.11      frystyk   165:     return HTBind_add(suffix, NULL, NULL, language, value);
2.1       frystyk   166: }
                    167: 
2.11      frystyk   168: PUBLIC BOOL HTBind_add (CONST char *   suffix,
                    169:                        CONST char *    representation,
                    170:                        CONST char *    encoding,
                    171:                        CONST char *    language,
                    172:                        double          value)
2.1       frystyk   173: {
                    174:     HTBind * suff;
                    175:     if (!suffix)
                    176:        return NO;
                    177:     if (!strcmp(suffix, "*"))
                    178:        suff = &no_suffix;
                    179:     else if (!strcmp(suffix, "*.*"))
                    180:        suff = &unknown_suffix;
                    181:     else {
                    182:        HTList *suflist;
                    183:        int hash=0;
                    184:        CONST char *ptr=suffix;
                    185: 
                    186:        /* Select list from hash table */
                    187:        for( ; *ptr; ptr++)
                    188:            hash = (int) ((hash * 3 + (*(unsigned char*)ptr)) % HASH_SIZE);
                    189: 
                    190:        if (!HTBindings[hash]) HTBindings[hash] = HTList_new();
                    191:        suflist = HTBindings[hash];
                    192: 
                    193:        /* Look for existing binding */
                    194:        {
                    195:            HTList *cur = suflist;
                    196:            while ((suff = (HTBind *) HTList_nextObject(cur)) != NULL) {
                    197:                if (!strcmp(suff->suffix, suffix))
                    198:                    break;
                    199:            }
                    200:        }
                    201: 
                    202:        /* If not found -- create a new node */
                    203:        if (!suff) {
2.17      frystyk   204:            if ((suff = (HTBind *) HT_CALLOC(1, sizeof(HTBind))) == NULL)
                    205:                HT_OUTOFMEM("HTBind_add");
2.1       frystyk   206:            HTList_addObject(suflist, (void *) suff);
                    207:            StrAllocCopy(suff->suffix, suffix);
                    208:        }
                    209:     }
                    210: 
                    211:     /* Set the appropriate values */
                    212:     {
                    213:        char *str = NULL;
                    214:        char *ptr;
                    215:        if (representation) {
                    216:            StrAllocCopy(str, representation);
                    217:            for (ptr=str; *ptr; ptr++)
                    218:                *ptr = TOLOWER(*ptr);
                    219:            suff->type = HTAtom_for(str);
                    220:        }
                    221:        if (language) {
                    222:            StrAllocCopy(str, language);
                    223:            for (ptr=str; *ptr; ptr++)
                    224:                *ptr = TOLOWER(*ptr);
                    225:            suff->language = HTAtom_for(str);
                    226:        }
                    227:        if (encoding) {
                    228:            StrAllocCopy(str, encoding);
                    229:            for (ptr=str; *ptr; ptr++)
                    230:                *ptr = TOLOWER(*ptr);
                    231:            suff->encoding = HTAtom_for(str);
                    232:        }
2.17      frystyk   233:        HT_FREE(str);
2.1       frystyk   234:        suff->quality = value;
                    235:     }
                    236:     return YES;
                    237: }
                    238: 
                    239: 
                    240: /*     Determine a suitable suffix
                    241: **     ---------------------------
                    242: **  Use the set of bindings to find a suitable suffix (or index)
                    243: **  for a certain combination of language, media type and encoding
                    244: **  given in the anchor.
                    245: **
                    246: **  Returns a pointer to a suitable suffix string that must be freed 
                    247: **  by the caller. If more than one suffix is found they are all
                    248: **  concatenated using the first delimiter in HTDelimiters.
                    249: **  If no suffix is found, NULL is returned.
                    250: */
2.11      frystyk   251: PUBLIC char * HTBind_getSuffix (HTParentAnchor * anchor)
2.1       frystyk   252: {
                    253:     int cnt;
                    254:     HTList *cur;
                    255:     char *suffix = NULL;
2.5       frystyk   256:     char delimiter[2];
2.4       frystyk   257:     *delimiter = *HTDelimiters;
2.5       frystyk   258:     *(delimiter+1) = '\0';
2.1       frystyk   259:     if (anchor) {
                    260:        for (cnt=0; cnt<HASH_SIZE; cnt++) {
                    261:            if ((cur = HTBindings[cnt])) { 
                    262:                HTBind *pres;
                    263:                while ((pres = (HTBind *) HTList_nextObject(cur)) != NULL) {
                    264:                    if ((pres->type && pres->type==anchor->content_type) ||
                    265:                        (pres->encoding &&
                    266:                         pres->encoding!=WWW_ENC_7BIT &&
                    267:                         pres->encoding!=WWW_ENC_8BIT &&
                    268:                         pres->encoding!=WWW_ENC_BINARY &&
                    269:                         pres->encoding==anchor->content_encoding) ||
                    270:                        (pres->language &&
                    271:                         pres->language == anchor->content_language)) {
2.3       frystyk   272:                        StrAllocCat(suffix, delimiter);
2.1       frystyk   273:                        StrAllocCat(suffix, pres->suffix);
                    274:                    }
                    275:                }
                    276:            }
                    277:        }
                    278:     }
                    279:     return suffix;
                    280: }
                    281: 
                    282: /*     Determine the description of a file
                    283: **     -----------------------------------
                    284: **  Use the set of bindings to find the combination of language,
                    285: **  media type and encoding of a given file name.
                    286: **
                    287: **  If more than one suffix is found they are all searched. The last suffix
                    288: **  has highest priority, the first one lowest. See also HTBind_getFormat()
                    289: **
                    290: **  Returns a contentdescription object with the representations found. This
                    291: **  must be free by the caller
                    292: */
2.11      frystyk   293: PUBLIC HTContentDescription * HTBind_getDescription (char * file)
2.1       frystyk   294: {
                    295:     HTContentDescription * cd;
2.17      frystyk   296:     if ((cd = (HTContentDescription  *) HT_CALLOC(1, sizeof(HTContentDescription))) == NULL)
                    297:         HT_OUTOFMEM("HTContentDescription");
2.1       frystyk   298:     if (HTBind_getFormat(file, &cd->content_type, &cd->content_encoding,
                    299:                         &cd->content_language, &cd->quality))
                    300:        return cd;
                    301:     else {
2.17      frystyk   302:        HT_FREE(cd);
2.1       frystyk   303:        return NULL;
                    304:     }
                    305: }
                    306: 
                    307: /*     Determine the content of an Anchor
                    308: **     ----------------------------------
                    309: **  Use the set of bindings to find the combination of language,
                    310: **  media type and encoding of a given anchor.
                    311: **
                    312: **  If more than one suffix is found they are all searched. The last suffix
                    313: **  has highest priority, the first one lowest. See also HTBind_getFormat()
                    314: **
                    315: **  Returns the anchor object with the representations found
                    316: */
2.11      frystyk   317: PUBLIC BOOL HTBind_getBindings (HTParentAnchor * anchor)
2.1       frystyk   318: {
                    319:     BOOL status = NO;
                    320:     double quality=1.0;                  /* @@@ Should we add this into the anchor? */
                    321:     if (anchor) {
                    322:        char *addr = HTAnchor_physical(anchor);
                    323:        char *path = HTParse(addr, "", PARSE_PATH+PARSE_PUNCTUATION);
2.9       frystyk   324:        char *file;
                    325:        char *end;
                    326:        if ((end = strchr(path, ';')) || (end = strchr(path, '?')) ||
                    327:            (end = strchr(path, '#')))
                    328:            *end = '\0';
                    329:        if ((file = strrchr(path, '/'))) {
                    330:            if (BIND_TRACE)
2.18    ! eric      331:                HTTrace("Get Binding. for file: `%s\'\n", path);
2.1       frystyk   332:            status = HTBind_getFormat(file, &anchor->content_type,
                    333:                                      &anchor->content_encoding,
                    334:                                      &anchor->content_language,
                    335:                                      &quality);
                    336:        }
2.17      frystyk   337:        HT_FREE(path);
2.1       frystyk   338:     }
                    339:     return status;
                    340: }
                    341: 
                    342: 
                    343: /*     Determine the content of an file name
                    344: **     -------------------------------------
                    345: **  Use the set of bindings to find the combination of language,
                    346: **  media type and encoding of a given anchor.
                    347: **
                    348: **  If more than one suffix is found they are all searched. The last suffix
                    349: **  has highest priority, the first one lowest. See also HTBind_getBindings()
2.10      frystyk   350: **  Either of format, encoding, or language can be NULL
2.1       frystyk   351: **  Returns the format, encoding, and language found
                    352: */
2.10      frystyk   353: PUBLIC BOOL HTBind_getFormat (CONST char * filename, HTFormat * format,
                    354:                              HTEncoding * enc, HTLanguage * lang,
                    355:                              double * quality)
2.1       frystyk   356: {
                    357:     int sufcnt=0;
                    358:     char *file=NULL;
2.6       frystyk   359: #ifdef HT_REENTRANT
                    360:     char *lasts;                                            /* For strtok_r */
                    361: #endif
2.1       frystyk   362:     if (*quality < HT_EPSILON)
                    363:        *quality = 1.0;                            /* Set to a neutral value */
                    364:     StrAllocCopy(file, filename);
                    365:     HTUnEscape(file);                             /* Unescape the file name */
2.6       frystyk   366: #ifdef HT_REENTRANT
                    367:     if (strtok_r(file, HTDelimiters, &lasts)) {         /* Do we have any suffixes? */
                    368: #else
2.1       frystyk   369:     if (strtok(file, HTDelimiters)) {           /* Do we have any suffixes? */
2.6       frystyk   370: #endif /* HT_REENTRANT */
2.1       frystyk   371:        char *suffix;
2.6       frystyk   372: #ifdef HT_REENTRANT
                    373:        while ((suffix=(char*)strtok_r(NULL, HTDelimiters, &lasts)) != NULL) {
                    374: #else
                    375:        while ((suffix=strtok(NULL, HTDelimiters)) != NULL) {
                    376: #endif /* HT_REENTRANT */
2.1       frystyk   377:            HTBind *suff=NULL;
                    378:            int hash=0;
                    379:            char *ptr=suffix;
                    380:            if (BIND_TRACE)
2.18    ! eric      381:                HTTrace("Get Binding. Look for '%s\' ", suffix);
2.1       frystyk   382:            sufcnt++;
                    383: 
                    384:            /* Select list from hash table */
                    385:            for( ; *ptr; ptr++)
                    386:                hash = (int)((hash*3+(*(unsigned char*)ptr)) % HASH_SIZE);
                    387: 
                    388:            /* Now search list for entries (case or non case sensitive) */
                    389:            if (HTBindings[hash]) {
                    390:                HTList *cur = HTBindings[hash];
                    391:                while ((suff = (HTBind *) HTList_nextObject(cur))) {
                    392:                    if ((HTCaseSen && !strcmp(suff->suffix, suffix)) ||
                    393:                        !strcasecomp(suff->suffix, suffix)) {
2.18    ! eric      394:                        if (BIND_TRACE) HTTrace("Found!\n");
2.10      frystyk   395:                        if (suff->type && format) *format = suff->type;
                    396:                        if (suff->encoding && enc) *enc = suff->encoding;
                    397:                        if (suff->language && lang) *lang = suff->language;
2.1       frystyk   398:                        if (suff->quality > HT_EPSILON)
                    399:                            *quality *= suff->quality;
                    400:                        break;
                    401:                    }
                    402:                }
                    403:            }
                    404:            if (!suff) {        /* We don't have this suffix - use default */
                    405:                if (BIND_TRACE)
2.18    ! eric      406:                    HTTrace("Not found - use default for \'*.*\'\n");
2.10      frystyk   407:                if (format) *format = unknown_suffix.type;
                    408:                if (enc) *enc = unknown_suffix.encoding;
                    409:                if (lang) *lang = unknown_suffix.language;
2.1       frystyk   410:                *quality = unknown_suffix.quality;
                    411:            }
                    412:        } /* while we still have suffixes */
                    413:     }
                    414:     if (!sufcnt) {             /* No suffix so use default value */
                    415:        if (BIND_TRACE)
2.18    ! eric      416:            HTTrace("Get Binding. No suffix found - using default '%s\'\n", filename);
2.10      frystyk   417:        if (format) *format = no_suffix.type;
                    418:        if (enc) *enc = no_suffix.encoding;
                    419:        if (lang) *lang = no_suffix.language;
2.1       frystyk   420:        *quality = no_suffix.quality;
                    421:     }
                    422:     if (BIND_TRACE)
2.18    ! eric      423:        HTTrace("Get Binding. Result for '%s\' is: type='%s\', encoding='%s\', language='%s\' with quality %.2f\n",
2.1       frystyk   424:                filename,
2.10      frystyk   425:                (format && *format) ? HTAtom_name(*format) : "unknown",
                    426:                (enc && *enc) ? HTAtom_name(*enc) : "unknown",
                    427:                (lang && *lang) ? HTAtom_name(*lang) : "unknown",
2.1       frystyk   428:                *quality);
2.17      frystyk   429:     HT_FREE(file);
2.1       frystyk   430:     return YES;
                    431: }
                    432: 

Webmaster