Annotation of libwww/Library/src/HTBind.c, revision 2.12

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"
                     37: #include "HTAnchor.h"
                     38: #include "HTAtom.h"
                     39: #include "HTParse.h"
                     40: 
                     41: #ifdef VMS
                     42: #include "HTVMSUtils.h"
                     43: #endif
                     44: 
                     45: #include "HTBind.h"                                     /* Implemented here */
                     46: 
                     47: typedef struct _HTBind {
                     48:     char *     suffix;
                     49:     HTFormat   type;                   /* Content-Type */
                     50:     HTEncoding encoding;               /* Content-Encoding */
                     51:     HTLanguage language;               /* Content-Language */
                     52:     double     quality;
                     53: } HTBind;
                     54: 
                     55: #define HASH_SIZE      101        /* Arbitrary prime. Memory/speed tradeoff */
                     56: 
                     57: /* Suffix registration */
                     58: PRIVATE BOOL HTCaseSen = YES;                /* Are suffixes case sensitive */
                     59: PRIVATE char *HTDelimiters = NULL;                       /* Set of suffixes */
                     60: 
                     61: PRIVATE HTList **HTBindings = NULL;   /* Point to table of lists of bindings */
                     62: 
                     63: PRIVATE HTBind no_suffix = { "*", NULL, NULL, NULL, 1.0 };
                     64: PRIVATE HTBind unknown_suffix = { "*.*", NULL, NULL, NULL, 1.0};
                     65: 
                     66: /* ------------------------------------------------------------------------- */
                     67: 
                     68: /*     
                     69: **     Set up the list of suffix bindings. Done by HTLibInit
                     70: */
2.11      frystyk    71: PUBLIC BOOL HTBind_init (void)
2.1       frystyk    72: {
2.8       frystyk    73:     if (!HTBindings) {
2.1       frystyk    74:        HTBindings = (HTList**) calloc(HASH_SIZE, sizeof(HTList *));
2.8       frystyk    75:        if (!HTBindings) outofmem(__FILE__, "HTBind_init");
                     76:     }
2.1       frystyk    77:     StrAllocCopy(HTDelimiters, DEFAULT_SUFFIXES);
                     78:     return YES;
                     79: }
                     80: 
                     81: 
                     82: /*
                     83: **     Cleans up the memory allocated by file bindings
                     84: **     Done by HTLibTerminate().
                     85: **     Written by Eric Sink, eric@spyglass.com, and Henrik
                     86: */
2.11      frystyk    87: PUBLIC BOOL HTBind_deleteAll (void)
2.1       frystyk    88: {
                     89:     int cnt;
                     90:     HTList *cur;
                     91:     if (!HTBindings)
                     92:        return NO;
                     93:     for (cnt=0; cnt<HASH_SIZE; cnt++) {
                     94:        if ((cur = HTBindings[cnt])) { 
                     95:            HTBind *pres;
                     96:            while ((pres = (HTBind *) HTList_nextObject(cur)) != NULL) {
                     97:                FREE(pres->suffix);
                     98:                free(pres);
                     99:            }
                    100:        }
                    101:        HTList_delete(HTBindings[cnt]);
                    102:        HTBindings[cnt] = NULL;
                    103:     }
                    104:     FREE(HTDelimiters);
                    105:     return YES;
                    106: }
                    107: 
                    108: 
                    109: /*     Make suffix bindings case sensitive
                    110: **     -----------------------------------
                    111: */
2.11      frystyk   112: PUBLIC void HTBind_caseSensitive (BOOL sensitive)
2.1       frystyk   113: {
                    114:     HTCaseSen = sensitive;
                    115: }
                    116: 
                    117: 
                    118: /*     Get set of suffixes
                    119: **     -------------------
                    120: */
2.11      frystyk   121: PUBLIC CONST char *HTBind_delimiters (void)
2.1       frystyk   122: {
                    123:     return HTDelimiters;
                    124: }
                    125: 
                    126: 
                    127: /*     Change set of suffixes
                    128: **     ----------------------
                    129: */
2.11      frystyk   130: PUBLIC void HTBind_setDelimiters (CONST char * new_suffixes)
2.1       frystyk   131: {
                    132:     if (new_suffixes && *new_suffixes)
                    133:        StrAllocCopy(HTDelimiters, new_suffixes);
                    134: }
                    135: 
                    136: 
                    137: /*     Define the representation associated with a file suffix
                    138: **     -------------------------------------------------------
                    139: **
                    140: **     Calling this with suffix set to "*" will set the default
                    141: **     representation.
                    142: **     Calling this with suffix set to "*.*" will set the default
                    143: **     representation for unknown suffix files which contain a "."
                    144: **
                    145: **     If filename suffix is already defined its previous
                    146: **     definition is overridden (or modified)
                    147: */
2.12    ! frystyk   148: PUBLIC BOOL HTBind_addType (CONST char *       suffix,
        !           149:                            CONST char *        representation,
        !           150:                            double              value)
2.1       frystyk   151: {
2.11      frystyk   152:     return HTBind_add(suffix, representation, NULL, NULL, value);
2.1       frystyk   153: }
                    154: 
2.11      frystyk   155: PUBLIC BOOL HTBind_addEncoding (CONST char *   suffix,
                    156:                                CONST char *    encoding,
                    157:                                double          value)
2.1       frystyk   158: {
2.11      frystyk   159:     return HTBind_add(suffix, NULL, encoding, NULL, value);
2.1       frystyk   160: }
                    161: 
2.11      frystyk   162: PUBLIC BOOL HTBind_addLanguage (CONST char *   suffix,
                    163:                                CONST char *    language,
                    164:                                double          value)
2.1       frystyk   165: {
2.11      frystyk   166:     return HTBind_add(suffix, NULL, NULL, language, value);
2.1       frystyk   167: }
                    168: 
2.11      frystyk   169: PUBLIC BOOL HTBind_add (CONST char *   suffix,
                    170:                        CONST char *    representation,
                    171:                        CONST char *    encoding,
                    172:                        CONST char *    language,
                    173:                        double          value)
2.1       frystyk   174: {
                    175:     HTBind * suff;
                    176:     if (!suffix)
                    177:        return NO;
                    178:     if (!strcmp(suffix, "*"))
                    179:        suff = &no_suffix;
                    180:     else if (!strcmp(suffix, "*.*"))
                    181:        suff = &unknown_suffix;
                    182:     else {
                    183:        HTList *suflist;
                    184:        int hash=0;
                    185:        CONST char *ptr=suffix;
                    186: 
                    187:        /* Select list from hash table */
                    188:        for( ; *ptr; ptr++)
                    189:            hash = (int) ((hash * 3 + (*(unsigned char*)ptr)) % HASH_SIZE);
                    190: 
                    191:        if (!HTBindings[hash]) HTBindings[hash] = HTList_new();
                    192:        suflist = HTBindings[hash];
                    193: 
                    194:        /* Look for existing binding */
                    195:        {
                    196:            HTList *cur = suflist;
                    197:            while ((suff = (HTBind *) HTList_nextObject(cur)) != NULL) {
                    198:                if (!strcmp(suff->suffix, suffix))
                    199:                    break;
                    200:            }
                    201:        }
                    202: 
                    203:        /* If not found -- create a new node */
                    204:        if (!suff) {
                    205:            if ((suff = (HTBind *) calloc(1, sizeof(HTBind))) == NULL)
2.11      frystyk   206:                outofmem(__FILE__, "HTBind_add");
2.1       frystyk   207:            HTList_addObject(suflist, (void *) suff);
                    208:            StrAllocCopy(suff->suffix, suffix);
                    209:        }
                    210:     }
                    211: 
                    212:     /* Set the appropriate values */
                    213:     {
                    214:        char *str = NULL;
                    215:        char *ptr;
                    216:        if (representation) {
                    217:            StrAllocCopy(str, representation);
                    218:            for (ptr=str; *ptr; ptr++)
                    219:                *ptr = TOLOWER(*ptr);
                    220:            suff->type = HTAtom_for(str);
                    221:        }
                    222:        if (language) {
                    223:            StrAllocCopy(str, language);
                    224:            for (ptr=str; *ptr; ptr++)
                    225:                *ptr = TOLOWER(*ptr);
                    226:            suff->language = HTAtom_for(str);
                    227:        }
                    228:        if (encoding) {
                    229:            StrAllocCopy(str, encoding);
                    230:            for (ptr=str; *ptr; ptr++)
                    231:                *ptr = TOLOWER(*ptr);
                    232:            suff->encoding = HTAtom_for(str);
                    233:        }
                    234:        FREE(str);
                    235:        suff->quality = value;
                    236:     }
                    237:     return YES;
                    238: }
                    239: 
                    240: 
                    241: /*     Determine a suitable suffix
                    242: **     ---------------------------
                    243: **  Use the set of bindings to find a suitable suffix (or index)
                    244: **  for a certain combination of language, media type and encoding
                    245: **  given in the anchor.
                    246: **
                    247: **  Returns a pointer to a suitable suffix string that must be freed 
                    248: **  by the caller. If more than one suffix is found they are all
                    249: **  concatenated using the first delimiter in HTDelimiters.
                    250: **  If no suffix is found, NULL is returned.
                    251: */
2.11      frystyk   252: PUBLIC char * HTBind_getSuffix (HTParentAnchor * anchor)
2.1       frystyk   253: {
                    254:     int cnt;
                    255:     HTList *cur;
                    256:     char *suffix = NULL;
2.5       frystyk   257:     char delimiter[2];
2.4       frystyk   258:     *delimiter = *HTDelimiters;
2.5       frystyk   259:     *(delimiter+1) = '\0';
2.1       frystyk   260:     if (anchor) {
                    261:        for (cnt=0; cnt<HASH_SIZE; cnt++) {
                    262:            if ((cur = HTBindings[cnt])) { 
                    263:                HTBind *pres;
                    264:                while ((pres = (HTBind *) HTList_nextObject(cur)) != NULL) {
                    265:                    if ((pres->type && pres->type==anchor->content_type) ||
                    266:                        (pres->encoding &&
                    267:                         pres->encoding!=WWW_ENC_7BIT &&
                    268:                         pres->encoding!=WWW_ENC_8BIT &&
                    269:                         pres->encoding!=WWW_ENC_BINARY &&
                    270:                         pres->encoding==anchor->content_encoding) ||
                    271:                        (pres->language &&
                    272:                         pres->language == anchor->content_language)) {
2.3       frystyk   273:                        StrAllocCat(suffix, delimiter);
2.1       frystyk   274:                        StrAllocCat(suffix, pres->suffix);
                    275:                    }
                    276:                }
                    277:            }
                    278:        }
                    279:     }
                    280:     return suffix;
                    281: }
                    282: 
                    283: /*     Determine the description of a file
                    284: **     -----------------------------------
                    285: **  Use the set of bindings to find the combination of language,
                    286: **  media type and encoding of a given file name.
                    287: **
                    288: **  If more than one suffix is found they are all searched. The last suffix
                    289: **  has highest priority, the first one lowest. See also HTBind_getFormat()
                    290: **
                    291: **  Returns a contentdescription object with the representations found. This
                    292: **  must be free by the caller
                    293: */
2.11      frystyk   294: PUBLIC HTContentDescription * HTBind_getDescription (char * file)
2.1       frystyk   295: {
                    296:     HTContentDescription * cd;
                    297:     cd = (HTContentDescription *) calloc(1, sizeof(HTContentDescription));
                    298:     if (!cd) outofmem(__FILE__, "HTContentDescription");
                    299:     if (HTBind_getFormat(file, &cd->content_type, &cd->content_encoding,
                    300:                         &cd->content_language, &cd->quality))
                    301:        return cd;
                    302:     else {
                    303:        free(cd);
                    304:        return NULL;
                    305:     }
                    306: }
                    307: 
                    308: /*     Determine the content of an Anchor
                    309: **     ----------------------------------
                    310: **  Use the set of bindings to find the combination of language,
                    311: **  media type and encoding of a given anchor.
                    312: **
                    313: **  If more than one suffix is found they are all searched. The last suffix
                    314: **  has highest priority, the first one lowest. See also HTBind_getFormat()
                    315: **
                    316: **  Returns the anchor object with the representations found
                    317: */
2.11      frystyk   318: PUBLIC BOOL HTBind_getBindings (HTParentAnchor * anchor)
2.1       frystyk   319: {
                    320:     BOOL status = NO;
                    321:     double quality=1.0;                  /* @@@ Should we add this into the anchor? */
                    322:     if (anchor) {
                    323:        char *addr = HTAnchor_physical(anchor);
                    324:        char *path = HTParse(addr, "", PARSE_PATH+PARSE_PUNCTUATION);
2.9       frystyk   325:        char *file;
                    326:        char *end;
                    327:        if ((end = strchr(path, ';')) || (end = strchr(path, '?')) ||
                    328:            (end = strchr(path, '#')))
                    329:            *end = '\0';
                    330:        if ((file = strrchr(path, '/'))) {
                    331:            if (BIND_TRACE)
                    332:                fprintf(TDEST,"Get Binding. for file: `%s\'\n", path);
2.1       frystyk   333:            status = HTBind_getFormat(file, &anchor->content_type,
                    334:                                      &anchor->content_encoding,
                    335:                                      &anchor->content_language,
                    336:                                      &quality);
                    337:        }
                    338:        FREE(path);
                    339:     }
                    340:     return status;
                    341: }
                    342: 
                    343: 
                    344: /*     Determine the content of an file name
                    345: **     -------------------------------------
                    346: **  Use the set of bindings to find the combination of language,
                    347: **  media type and encoding of a given anchor.
                    348: **
                    349: **  If more than one suffix is found they are all searched. The last suffix
                    350: **  has highest priority, the first one lowest. See also HTBind_getBindings()
2.10      frystyk   351: **  Either of format, encoding, or language can be NULL
2.1       frystyk   352: **  Returns the format, encoding, and language found
                    353: */
2.10      frystyk   354: PUBLIC BOOL HTBind_getFormat (CONST char * filename, HTFormat * format,
                    355:                              HTEncoding * enc, HTLanguage * lang,
                    356:                              double * quality)
2.1       frystyk   357: {
                    358:     int sufcnt=0;
                    359:     char *file=NULL;
2.6       frystyk   360: #ifdef HT_REENTRANT
                    361:     char *lasts;                                            /* For strtok_r */
                    362: #endif
2.1       frystyk   363:     if (*quality < HT_EPSILON)
                    364:        *quality = 1.0;                            /* Set to a neutral value */
                    365:     StrAllocCopy(file, filename);
                    366:     HTUnEscape(file);                             /* Unescape the file name */
2.6       frystyk   367: #ifdef HT_REENTRANT
                    368:     if (strtok_r(file, HTDelimiters, &lasts)) {         /* Do we have any suffixes? */
                    369: #else
2.1       frystyk   370:     if (strtok(file, HTDelimiters)) {           /* Do we have any suffixes? */
2.6       frystyk   371: #endif /* HT_REENTRANT */
2.1       frystyk   372:        char *suffix;
2.6       frystyk   373: #ifdef HT_REENTRANT
                    374:        while ((suffix=(char*)strtok_r(NULL, HTDelimiters, &lasts)) != NULL) {
                    375: #else
                    376:        while ((suffix=strtok(NULL, HTDelimiters)) != NULL) {
                    377: #endif /* HT_REENTRANT */
2.1       frystyk   378:            HTBind *suff=NULL;
                    379:            int hash=0;
                    380:            char *ptr=suffix;
                    381:            if (BIND_TRACE)
                    382:                fprintf(TDEST, "Get Binding. Look for '%s\' ", suffix);
                    383:            sufcnt++;
                    384: 
                    385:            /* Select list from hash table */
                    386:            for( ; *ptr; ptr++)
                    387:                hash = (int)((hash*3+(*(unsigned char*)ptr)) % HASH_SIZE);
                    388: 
                    389:            /* Now search list for entries (case or non case sensitive) */
                    390:            if (HTBindings[hash]) {
                    391:                HTList *cur = HTBindings[hash];
                    392:                while ((suff = (HTBind *) HTList_nextObject(cur))) {
                    393:                    if ((HTCaseSen && !strcmp(suff->suffix, suffix)) ||
                    394:                        !strcasecomp(suff->suffix, suffix)) {
                    395:                        if (BIND_TRACE) fprintf(TDEST, "Found!\n");
2.10      frystyk   396:                        if (suff->type && format) *format = suff->type;
                    397:                        if (suff->encoding && enc) *enc = suff->encoding;
                    398:                        if (suff->language && lang) *lang = suff->language;
2.1       frystyk   399:                        if (suff->quality > HT_EPSILON)
                    400:                            *quality *= suff->quality;
                    401:                        break;
                    402:                    }
                    403:                }
                    404:            }
                    405:            if (!suff) {        /* We don't have this suffix - use default */
                    406:                if (BIND_TRACE)
                    407:                    fprintf(TDEST,"Not found - use default for \'*.*\'\n");
2.10      frystyk   408:                if (format) *format = unknown_suffix.type;
                    409:                if (enc) *enc = unknown_suffix.encoding;
                    410:                if (lang) *lang = unknown_suffix.language;
2.1       frystyk   411:                *quality = unknown_suffix.quality;
                    412:            }
                    413:        } /* while we still have suffixes */
                    414:     }
                    415:     if (!sufcnt) {             /* No suffix so use default value */
                    416:        if (BIND_TRACE)
                    417:            fprintf(TDEST, "Get Binding. No suffix found - using default '%s\'\n", filename);
2.10      frystyk   418:        if (format) *format = no_suffix.type;
                    419:        if (enc) *enc = no_suffix.encoding;
                    420:        if (lang) *lang = no_suffix.language;
2.1       frystyk   421:        *quality = no_suffix.quality;
                    422:     }
                    423:     if (BIND_TRACE)
                    424:        fprintf(TDEST, "Get Binding. Result for '%s\' is: type='%s\', encoding='%s\', language='%s\' with quality %.2f\n",
                    425:                filename,
2.10      frystyk   426:                (format && *format) ? HTAtom_name(*format) : "unknown",
                    427:                (enc && *enc) ? HTAtom_name(*enc) : "unknown",
                    428:                (lang && *lang) ? HTAtom_name(*lang) : "unknown",
2.1       frystyk   429:                *quality);
                    430:     free(file);
                    431:     return YES;
                    432: }
                    433: 

Webmaster