Annotation of libwww/Library/src/HTAnchor.c, revision 1.7

1.1       timbl       1: /*     Hypertext "Anchor" Object                               HTAnchor.c
                      2: **     ==========================
                      3: **
                      4: ** An anchor represents a region of a hypertext document which is linked to
                      5: ** another anchor in the same or a different document.
                      6: **
                      7: ** History
                      8: **
                      9: **         Nov 1990  Written in Objective-C for the NeXT browser (TBL)
                     10: **     24-Oct-1991 (JFG), written in C, browser-independant 
                     11: **     21-Nov-1991 (JFG), first complete version
                     12: **
                     13: **     (c) Copyright CERN 1991 - See Copyright.html
                     14: */
                     15: 
                     16: #define HASH_SIZE 101          /* Arbitrary prime. Memory/speed tradeoff */
                     17: 
                     18: #include <ctype.h>
                     19: #include "tcp.h"
1.7     ! luotonen   20: #include "HTFormat.h"
1.1       timbl      21: #include "HTAnchor.h"
1.6       timbl      22: 
1.1       timbl      23: #include "HTUtils.h"
                     24: #include "HTParse.h"
1.6       timbl      25: #include "HTFWriter.h" /* for cache stuff */
1.1       timbl      26: 
                     27: typedef struct _HyperDoc Hyperdoc;
1.5       duns       28: #ifdef VMS
1.1       timbl      29: struct _HyperDoc {
                     30:        int junk;       /* VMS cannot handle pointers to undefined structs */
                     31: };
                     32: #endif
                     33: 
                     34: PRIVATE HTList **adult_table=0;  /* Point to table of lists of all parents */
                     35: 
                     36: /*                             Creation Methods
                     37: **                             ================
                     38: **
                     39: **     Do not use "new" by itself outside this module. In order to enforce
                     40: **     consistency, we insist that you furnish more information about the
                     41: **     anchor you are creating : use newWithParent or newWithAddress.
                     42: */
                     43: 
                     44: PRIVATE HTParentAnchor * HTParentAnchor_new
                     45:   NOARGS
                     46: {
                     47:   HTParentAnchor *newAnchor = 
                     48:     (HTParentAnchor *) calloc (1, sizeof (HTParentAnchor));  /* zero-filled */
                     49:   newAnchor->parent = newAnchor;
                     50:   return newAnchor;
                     51: }
                     52: 
                     53: PRIVATE HTChildAnchor * HTChildAnchor_new
                     54:   NOARGS
                     55: {
                     56:   return (HTChildAnchor *) calloc (1, sizeof (HTChildAnchor));  /* zero-filled */
                     57: }
                     58: 
                     59: 
                     60: /*     Case insensitive string comparison
                     61: **     ----------------------------------
                     62: ** On entry,
                     63: **     s       Points to one string, null terminated
                     64: **     t       points to the other.
                     65: ** On exit,
                     66: **     returns YES if the strings are equivalent ignoring case
                     67: **             NO if they differ in more than  their case.
                     68: */
                     69: 
                     70: PRIVATE BOOL equivalent
                     71:   ARGS2 (CONST char *,s, CONST char *,t)
                     72: {
                     73:   if (s && t) {  /* Make sure they point to something */
                     74:     for ( ; *s && *t ; s++, t++) {
                     75:         if (TOUPPER(*s) != TOUPPER(*t))
                     76:          return NO;
                     77:     }
                     78:     return TOUPPER(*s) == TOUPPER(*t);
                     79:   } else
                     80:     return s == t;  /* Two NULLs are equivalent, aren't they ? */
                     81: }
                     82: 
                     83: 
                     84: /*     Create new or find old sub-anchor
                     85: **     ---------------------------------
                     86: **
1.3       timbl      87: **     Me one is for a new anchor being edited into an existing
1.1       timbl      88: **     document. The parent anchor must already exist.
                     89: */
                     90: 
1.2       timbl      91: PUBLIC HTChildAnchor * HTAnchor_findChild
1.1       timbl      92:   ARGS2 (HTParentAnchor *,parent, CONST char *,tag)
                     93: {
                     94:   HTChildAnchor *child;
                     95:   HTList *kids;
                     96: 
                     97:   if (! parent) {
                     98:     if (TRACE) printf ("HTAnchor_findChild called with NULL parent.\n");
                     99:     return NULL;
                    100:   }
1.7     ! luotonen  101:   if ((kids = parent->children)) {  /* parent has children : search them */
1.1       timbl     102:     if (tag && *tag) {         /* TBL */
1.7     ! luotonen  103:        while ((child = HTList_nextObject (kids))) {
1.1       timbl     104:            if (equivalent(child->tag, tag)) { /* Case sensitive 920226 */
1.4       timbl     105:                if (TRACE) fprintf (stderr,
1.1       timbl     106:               "Child anchor %p of parent %p with name `%s' already exists.\n",
1.4       timbl     107:                    (void*)child, (void*)parent, tag);
1.1       timbl     108:                return child;
                    109:            }
                    110:        }
                    111:      }  /*  end if tag is void */
                    112:   } else  /* parent doesn't have any children yet : create family */
                    113:     parent->children = HTList_new ();
                    114: 
                    115:   child = HTChildAnchor_new ();
                    116:   if (TRACE) fprintf(stderr, "new Anchor %p named `%s' is child of %p\n",
1.4       timbl     117:        (void*)child, (int)tag ? tag : (CONST char *)"" , (void*)parent); /* int for apollo */
1.1       timbl     118:   HTList_addObject (parent->children, child);
                    119:   child->parent = parent;
                    120:   StrAllocCopy(child->tag, tag);
                    121:   return child;
                    122: }
                    123: 
                    124: 
                    125: /*     Create or find a child anchor with a possible link
                    126: **     --------------------------------------------------
                    127: **
                    128: **     Create new anchor with a given parent and possibly
                    129: **     a name, and possibly a link to a _relatively_ named anchor.
                    130: **     (Code originally in ParseHTML.h)
                    131: */
                    132: PUBLIC HTChildAnchor * HTAnchor_findChildAndLink
                    133:   ARGS4(
                    134:        HTParentAnchor *,parent,        /* May not be 0 */
                    135:        CONST char *,tag,       /* May be "" or 0 */
                    136:        CONST char *,href,      /* May be "" or 0 */
                    137:        HTLinkType *,ltype      /* May be 0 */
                    138:        )
                    139: {
                    140:   HTChildAnchor * child = HTAnchor_findChild(parent, tag);
                    141:   if (href && *href) {
1.2       timbl     142:     char * relative_to = HTAnchor_address((HTAnchor *) parent);
                    143:     char * parsed_address = HTParse(href, relative_to, PARSE_ALL);
                    144:     HTAnchor * dest = HTAnchor_findAddress(parsed_address);
1.1       timbl     145:     HTAnchor_link((HTAnchor *) child, dest, ltype);
                    146:     free(parsed_address);
1.2       timbl     147:     free(relative_to);
1.1       timbl     148:   }
                    149:   return child;
                    150: }
                    151: 
                    152: 
                    153: /*     Create new or find old named anchor
                    154: **     -----------------------------------
                    155: **
1.3       timbl     156: **     Me one is for a reference which is found in a document, and might
1.1       timbl     157: **     not be already loaded.
                    158: **     Note: You are not guaranteed a new anchor -- you might get an old one,
                    159: **     like with fonts.
                    160: */
                    161: 
                    162: HTAnchor * HTAnchor_findAddress
                    163:   ARGS1 (CONST char *,address)
                    164: {
                    165:   char *tag = HTParse (address, "", PARSE_ANCHOR);  /* Anchor tag specified ? */
                    166: 
                    167:   /* If the address represents a sub-anchor, we recursively load its parent,
                    168:      then we create a child anchor within that document. */
                    169:   if (*tag) {
                    170:     char *docAddress = HTParse(address, "", PARSE_ACCESS | PARSE_HOST |
                    171:                                            PARSE_PATH | PARSE_PUNCTUATION);
                    172:     HTParentAnchor * foundParent =
                    173:       (HTParentAnchor *) HTAnchor_findAddress (docAddress);
                    174:     HTChildAnchor * foundAnchor = HTAnchor_findChild (foundParent, tag);
                    175:     free (docAddress);
                    176:     free (tag);
                    177:     return (HTAnchor *) foundAnchor;
                    178:   }
                    179:   
1.3       timbl     180:   else { /* If the address has no anchor tag, 
                    181:            check whether we have this node */
1.1       timbl     182:     int hash;
                    183:     CONST char *p;
                    184:     HTList * adults;
                    185:     HTList *grownups;
                    186:     HTParentAnchor * foundAnchor;
                    187: 
                    188:     free (tag);
                    189:     
                    190:     /* Select list from hash table */
1.2       timbl     191:     for(p=address, hash=0; *p; p++)
                    192:        hash = (hash * 3 + (*(unsigned char*)p))
                    193:         % HASH_SIZE;
1.1       timbl     194:     if (!adult_table)
                    195:         adult_table = (HTList**) calloc(HASH_SIZE, sizeof(HTList*));
                    196:     if (!adult_table[hash]) adult_table[hash] = HTList_new();
                    197:     adults = adult_table[hash];
                    198: 
                    199:     /* Search list for anchor */
                    200:     grownups = adults;
1.7     ! luotonen  201:     while ((foundAnchor = HTList_nextObject (grownups))) {
1.2       timbl     202:        if (equivalent(foundAnchor->address, address)) {
1.1       timbl     203:        if (TRACE) fprintf(stderr, "Anchor %p with address `%s' already exists.\n",
1.4       timbl     204:                          (void*) foundAnchor, address);
1.1       timbl     205:        return (HTAnchor *) foundAnchor;
                    206:       }
                    207:     }
                    208:     
                    209:     /* Node not found : create new anchor */
                    210:     foundAnchor = HTParentAnchor_new ();
                    211:     if (TRACE) fprintf(stderr, "New anchor %p has hash %d and address `%s'\n",
1.4       timbl     212:        (void*)foundAnchor, hash, address);
1.1       timbl     213:     StrAllocCopy(foundAnchor->address, address);
                    214:     HTList_addObject (adults, foundAnchor);
                    215:     return (HTAnchor *) foundAnchor;
                    216:   }
                    217: }
                    218: 
                    219: 
                    220: /*     Delete an anchor and possibly related things (auto garbage collection)
                    221: **     --------------------------------------------
                    222: **
                    223: **     The anchor is only deleted if the corresponding document is not loaded.
                    224: **     All outgoing links from parent and children are deleted, and this anchor
                    225: **     is removed from the sources list of all its targets.
                    226: **     We also try to delete the targets whose documents are not loaded.
                    227: **     If this anchor's source list is empty, we delete it and its children.
                    228: */
                    229: 
                    230: PRIVATE void deleteLinks
1.3       timbl     231:   ARGS1 (HTAnchor *,me)
1.1       timbl     232: {
1.3       timbl     233:   if (! me)
1.1       timbl     234:     return;
                    235: 
                    236:   /* Recursively try to delete target anchors */
1.3       timbl     237:   if (me->mainLink.dest) {
                    238:     HTParentAnchor *parent = me->mainLink.dest->parent;
                    239:     HTList_removeObject (parent->sources, me);
1.1       timbl     240:     if (! parent->document)  /* Test here to avoid calling overhead */
                    241:       HTAnchor_delete (parent);
                    242:   }
1.3       timbl     243:   if (me->links) {  /* Extra destinations */
1.1       timbl     244:     HTLink *target;
1.7     ! luotonen  245:     while ((target = HTList_removeLastObject (me->links))) {
1.1       timbl     246:       HTParentAnchor *parent = target->dest->parent;
1.3       timbl     247:       HTList_removeObject (parent->sources, me);
1.1       timbl     248:       if (! parent->document)  /* Test here to avoid calling overhead */
                    249:        HTAnchor_delete (parent);
                    250:     }
                    251:   }
                    252: }
                    253: 
                    254: PUBLIC BOOL HTAnchor_delete
1.3       timbl     255:   ARGS1 (HTParentAnchor *,me)
1.1       timbl     256: {
                    257:   HTChildAnchor *child;
                    258: 
                    259:   /* Don't delete if document is loaded */
1.3       timbl     260:   if (me->document)
1.1       timbl     261:     return NO;
                    262: 
                    263:   /* Recursively try to delete target anchors */
1.3       timbl     264:   deleteLinks ((HTAnchor *) me);
1.1       timbl     265: 
1.3       timbl     266:   if (! HTList_isEmpty (me->sources)) {  /* There are still incoming links */
1.1       timbl     267:     /* Delete all outgoing links from children, if any */
1.3       timbl     268:     HTList *kids = me->children;
1.7     ! luotonen  269:     while ((child = HTList_nextObject (kids)))
1.1       timbl     270:       deleteLinks ((HTAnchor *) child);
                    271:     return NO;  /* Parent not deleted */
                    272:   }
                    273: 
                    274:   /* No more incoming links : kill everything */
                    275:   /* First, recursively delete children */
1.7     ! luotonen  276:   while ((child = HTList_removeLastObject (me->children))) {
1.1       timbl     277:     deleteLinks ((HTAnchor *) child);
                    278:     free (child->tag);
                    279:     free (child);
                    280:   }
                    281: 
                    282:   /* Now kill myself */
1.3       timbl     283:   HTList_delete (me->children);
                    284:   HTList_delete (me->sources);
                    285:   free (me->address);
1.1       timbl     286:   /* Devise a way to clean out the HTFormat if no longer needed (ref count?) */
1.3       timbl     287:   free (me);
1.6       timbl     288:   if (me->cacheItems) {
                    289:       HTCacheClear(me->cacheItems);
                    290:   }
1.1       timbl     291:   return YES;  /* Parent deleted */
                    292: }
                    293: 
                    294: 
                    295: /*             Move an anchor to the head of the list of its siblings
                    296: **             ------------------------------------------------------
                    297: **
                    298: **     This is to ensure that an anchor which might have already existed
                    299: **     is put in the correct order as we load the document.
                    300: */
                    301: 
                    302: void HTAnchor_makeLastChild
1.3       timbl     303:   ARGS1(HTChildAnchor *,me)
1.1       timbl     304: {
1.3       timbl     305:   if (me->parent != (HTParentAnchor *) me) {  /* Make sure it's a child */
                    306:     HTList * siblings = me->parent->children;
                    307:     HTList_removeObject (siblings, me);
                    308:     HTList_addObject (siblings, me);
1.1       timbl     309:   }
                    310: }
                    311: 
                    312: /*     Data access functions
                    313: **     ---------------------
                    314: */
                    315: 
                    316: PUBLIC HTParentAnchor * HTAnchor_parent
1.3       timbl     317:   ARGS1 (HTAnchor *,me)
1.1       timbl     318: {
1.3       timbl     319:   return me ? me->parent : NULL;
1.1       timbl     320: }
                    321: 
                    322: void HTAnchor_setDocument
1.3       timbl     323:   ARGS2 (HTParentAnchor *,me, HyperDoc *,doc)
1.1       timbl     324: {
1.3       timbl     325:   if (me)
                    326:     me->document = doc;
1.1       timbl     327: }
                    328: 
                    329: HyperDoc * HTAnchor_document
1.3       timbl     330:   ARGS1 (HTParentAnchor *,me)
1.1       timbl     331: {
1.3       timbl     332:   return me ? me->document : NULL;
1.1       timbl     333: }
                    334: 
                    335: 
                    336: /* We don't want code to change an address after anchor creation... yet ?
                    337: void HTAnchor_setAddress
1.3       timbl     338:   ARGS2 (HTAnchor *,me, char *,addr)
1.1       timbl     339: {
1.3       timbl     340:   if (me)
                    341:     StrAllocCopy (me->parent->address, addr);
1.1       timbl     342: }
                    343: */
                    344: 
                    345: char * HTAnchor_address
1.3       timbl     346:   ARGS1 (HTAnchor *,me)
1.1       timbl     347: {
                    348:   char *addr = NULL;
1.3       timbl     349:   if (me) {
                    350:     if (((HTParentAnchor *) me == me->parent) ||
                    351:        !((HTChildAnchor *) me)->tag) {  /* it's an adult or no tag */
                    352:       StrAllocCopy (addr, me->parent->address);
1.1       timbl     353:     }
                    354:     else {  /* it's a named child */
1.3       timbl     355:       addr = malloc (2 + strlen (me->parent->address)
                    356:                     + strlen (((HTChildAnchor *) me)->tag));
1.1       timbl     357:       if (addr == NULL) outofmem(__FILE__, "HTAnchor_address");
1.3       timbl     358:       sprintf (addr, "%s#%s", me->parent->address,
                    359:               ((HTChildAnchor *) me)->tag);
1.1       timbl     360:     }
                    361:   }
                    362:   return addr;
                    363: }
                    364: 
                    365: 
                    366: 
                    367: void HTAnchor_setFormat
1.3       timbl     368:   ARGS2 (HTParentAnchor *,me, HTFormat ,form)
1.1       timbl     369: {
1.3       timbl     370:   if (me)
                    371:     me->format = form;
1.1       timbl     372: }
                    373: 
1.2       timbl     374: HTFormat HTAnchor_format
1.3       timbl     375:   ARGS1 (HTParentAnchor *,me)
1.1       timbl     376: {
1.3       timbl     377:   return me ? me->format : NULL;
1.1       timbl     378: }
                    379: 
                    380: 
                    381: 
                    382: void HTAnchor_setIndex
1.3       timbl     383:   ARGS1 (HTParentAnchor *,me)
1.1       timbl     384: {
1.3       timbl     385:   if (me)
                    386:     me->isIndex = YES;
1.1       timbl     387: }
                    388: 
                    389: BOOL HTAnchor_isIndex
1.3       timbl     390:   ARGS1 (HTParentAnchor *,me)
1.1       timbl     391: {
1.3       timbl     392:   return me ? me->isIndex : NO;
1.1       timbl     393: }
                    394: 
                    395: 
                    396: 
                    397: BOOL HTAnchor_hasChildren
1.3       timbl     398:   ARGS1 (HTParentAnchor *,me)
1.1       timbl     399: {
1.3       timbl     400:   return me ? ! HTList_isEmpty(me->children) : NO;
1.1       timbl     401: }
                    402: 
                    403: /*     Title handling
                    404: */
                    405: CONST char * HTAnchor_title
1.3       timbl     406:   ARGS1 (HTParentAnchor *,me)
1.1       timbl     407: {
1.3       timbl     408:   return me ? me->title : 0;
1.1       timbl     409: }
                    410: 
                    411: void HTAnchor_setTitle
1.3       timbl     412:   ARGS2(HTParentAnchor *,me, CONST char *,title)
1.1       timbl     413: {
1.3       timbl     414:   StrAllocCopy(me->title, title);
1.1       timbl     415: }
                    416: 
                    417: void HTAnchor_appendTitle
1.3       timbl     418:   ARGS2(HTParentAnchor *,me, CONST char *,title)
1.1       timbl     419: {
1.3       timbl     420:   StrAllocCat(me->title, title);
1.1       timbl     421: }
                    422: 
1.3       timbl     423: /*     Link me Anchor to another given one
1.1       timbl     424: **     -------------------------------------
                    425: */
                    426: 
                    427: BOOL HTAnchor_link
                    428:   ARGS3(HTAnchor *,source, HTAnchor *,destination, HTLinkType *,type)
                    429: {
                    430:   if (! (source && destination))
                    431:     return NO;  /* Can't link to/from non-existing anchor */
                    432:   if (TRACE) printf ("Linking anchor %p to anchor %p\n", source, destination);
                    433:   if (! source->mainLink.dest) {
                    434:     source->mainLink.dest = destination;
                    435:     source->mainLink.type = type;
                    436:   } else {
                    437:     HTLink * newLink = (HTLink *) malloc (sizeof (HTLink));
                    438:     if (newLink == NULL) outofmem(__FILE__, "HTAnchor_link");
                    439:     newLink->dest = destination;
                    440:     newLink->type = type;
                    441:     if (! source->links)
                    442:       source->links = HTList_new ();
                    443:     HTList_addObject (source->links, newLink);
                    444:   }
                    445:   if (!destination->parent->sources)
                    446:     destination->parent->sources = HTList_new ();
                    447:   HTList_addObject (destination->parent->sources, source);
                    448:   return YES;  /* Success */
                    449: }
                    450: 
                    451: 
                    452: /*     Manipulation of links
                    453: **     ---------------------
                    454: */
                    455: 
                    456: HTAnchor * HTAnchor_followMainLink
1.3       timbl     457:   ARGS1 (HTAnchor *,me)
1.1       timbl     458: {
1.3       timbl     459:   return me->mainLink.dest;
1.1       timbl     460: }
                    461: 
                    462: HTAnchor * HTAnchor_followTypedLink
1.3       timbl     463:   ARGS2 (HTAnchor *,me, HTLinkType *,type)
1.1       timbl     464: {
1.3       timbl     465:   if (me->mainLink.type == type)
                    466:     return me->mainLink.dest;
                    467:   if (me->links) {
                    468:     HTList *links = me->links;
1.1       timbl     469:     HTLink *link;
1.7     ! luotonen  470:     while ((link = HTList_nextObject (links)))
1.1       timbl     471:       if (link->type == type)
                    472:        return link->dest;
                    473:   }
1.3       timbl     474:   return NULL;  /* No link of me type */
1.1       timbl     475: }
                    476: 
1.2       timbl     477: 
                    478: /*     Make main link
                    479: */
1.1       timbl     480: BOOL HTAnchor_makeMainLink
1.3       timbl     481:   ARGS2 (HTAnchor *,me, HTLink *,movingLink)
1.1       timbl     482: {
                    483:   /* Check that everything's OK */
1.3       timbl     484:   if (! (me && HTList_removeObject (me->links, movingLink)))
1.1       timbl     485:     return NO;  /* link not found or NULL anchor */
                    486:   else {
                    487:     /* First push current main link onto top of links list */
                    488:     HTLink *newLink = (HTLink*) malloc (sizeof (HTLink));
                    489:     if (newLink == NULL) outofmem(__FILE__, "HTAnchor_makeMainLink");
1.3       timbl     490:     memcpy (newLink, & me->mainLink, sizeof (HTLink));
                    491:     HTList_addObject (me->links, newLink);
1.1       timbl     492: 
                    493:     /* Now make movingLink the new main link, and free it */
1.3       timbl     494:     memcpy (& me->mainLink, movingLink, sizeof (HTLink));
1.1       timbl     495:     free (movingLink);
                    496:     return YES;
                    497:   }
1.2       timbl     498: }
                    499: 
                    500: 
                    501: /*     Methods List
                    502: **     ------------
                    503: */
                    504: 
1.3       timbl     505: PUBLIC HTList * HTAnchor_methods ARGS1(HTParentAnchor *, me)
1.2       timbl     506: {
1.3       timbl     507:     if (!me->methods) {
                    508:         me->methods = HTList_new();
1.2       timbl     509:     }
1.3       timbl     510:     return me->methods;
1.2       timbl     511: }
                    512: 
                    513: /*     Protocol
                    514: **     --------
                    515: */
                    516: 
1.3       timbl     517: PUBLIC void * HTAnchor_protocol ARGS1(HTParentAnchor *, me)
1.2       timbl     518: {
1.3       timbl     519:     return me->protocol;
1.2       timbl     520: }
                    521: 
1.3       timbl     522: PUBLIC void HTAnchor_setProtocol ARGS2(HTParentAnchor *, me,
1.2       timbl     523:        void*,  protocol)
                    524: {
1.3       timbl     525:     me->protocol = protocol;
1.2       timbl     526: }
                    527: 
                    528: /*     Physical Address
                    529: **     ----------------
                    530: */
                    531: 
1.3       timbl     532: PUBLIC char * HTAnchor_physical ARGS1(HTParentAnchor *, me)
1.2       timbl     533: {
1.3       timbl     534:     return me->physical;
1.2       timbl     535: }
                    536: 
1.3       timbl     537: PUBLIC void HTAnchor_setPhysical ARGS2(HTParentAnchor *, me,
1.2       timbl     538:        char *, physical)
                    539: {
1.3       timbl     540:     StrAllocCopy(me->physical, physical);
1.1       timbl     541: }

Webmaster