Annotation of libwww/Library/src/HTFormat.c, revision 1.3

1.1       timbl       1: /*             Manage different file formats                   HTFormat.c
                      2: **             =============================
                      3: **
                      4: ** Bugs:
                      5: **     Not reentrant.
                      6: **
                      7: **     Assumes the incoming stream is ASCII, rather than a local file
                      8: **     format, and so ALWAYS converts from ASCII on non-ASCII machines.
                      9: **     Therefore, non-ASCII machines can't read local files.
1.2       timbl      10: **
                     11: */
                     12: 
                     13: /* Implements:
1.1       timbl      14: */
1.2       timbl      15: #include "HTFormat.h"
                     16: 
                     17: PUBLIC float HTMaxSecs = 1e10;         /* No effective limit */
                     18: PUBLIC float HTMaxLength = 1e10;       /* No effective limit */
                     19: 
                     20: #ifdef unix
                     21: #ifdef NeXT
                     22: #define PRESENT_POSTSCRIPT "open %s; /bin/rm -f %s\n"
                     23: #else
                     24: #define PRESENT_POSTSCRIPT "(ghostview %s ; /bin/rm -f %s)&\n" 
                     25:        /* Full pathname would be better! */
                     26: #endif
                     27: #endif
                     28: 
1.1       timbl      29: 
                     30: #include "HTUtils.h"
                     31: #include "tcp.h"
                     32: 
                     33: #include "HTML.h"
1.2       timbl      34: #include "HTMLDTD.h"
1.1       timbl      35: #include "HText.h"
1.2       timbl      36: #include "HTAlert.h"
                     37: #include "HTList.h"
                     38: #include "HTInit.h"
                     39: /*     Streams and structured streams which we use:
                     40: */
                     41: #include "HTFWriter.h"
                     42: #include "HTPlain.h"
                     43: #include "SGML.h"
                     44: #include "HTML.h"
                     45: #include "HTMLGen.h"
                     46: 
                     47: PUBLIC BOOL HTOutputSource = NO;       /* Flag: shortcut parser to stdout */
                     48: extern  BOOL interactive;
                     49: 
                     50: struct _HTStream {
                     51:       CONST HTStreamClass*     isa;
                     52:       /* ... */
                     53: };
                     54: 
                     55: 
                     56: /*     Presentation methods
                     57: **     --------------------
                     58: */
                     59: 
                     60: PUBLIC  HTList * HTPresentations = 0;
                     61: PUBLIC  HTPresentation* default_presentation = 0;
                     62: 
                     63: 
                     64: /*     Define a presentation system command for a content-type
                     65: **     -------------------------------------------------------
                     66: */
                     67: PUBLIC void HTSetPresentation ARGS5(
                     68:        CONST char *, representation,
                     69:        CONST char *, command,
                     70:        float,  quality,
                     71:        float,  secs, 
                     72:        float,  secs_per_byte
                     73: ){
                     74: 
                     75:     HTPresentation * pres = (HTPresentation *)malloc(sizeof(HTPresentation));
                     76:     if (pres == NULL) outofmem(__FILE__, "HTSetPresentation");
                     77:     
                     78:     pres->rep = HTAtom_for(representation);
                     79:     pres->rep_out = WWW_PRESENT;               /* Fixed for now ... :-) */
                     80:     pres->converter = HTSaveAndExecute;                /* Fixed for now ...     */
                     81:     pres->quality = quality;
                     82:     pres->secs = secs;
                     83:     pres->secs_per_byte = secs_per_byte;
                     84:     pres->rep = HTAtom_for(representation);
                     85:     pres->command = 0;
                     86:     StrAllocCopy(pres->command, command);
                     87:     
                     88:     if (!HTPresentations) HTPresentations = HTList_new();
                     89:     
                     90:     if (strcmp(representation, "*")==0) {
                     91:         if (default_presentation) free(default_presentation);
                     92:        default_presentation = pres;
                     93:     } else {
                     94:         HTList_addObject(HTPresentations, pres);
                     95:     }
                     96: }
                     97: 
                     98: 
                     99: /*     Define a built-in function for a content-type
                    100: **     ---------------------------------------------
                    101: */
                    102: PUBLIC void HTSetConversion ARGS6(
                    103:        CONST char *, representation_in,
                    104:        CONST char *, representation_out,
                    105:        HTConverter,    converter,
                    106:        float,  quality,
                    107:        float,  secs, 
                    108:        float,  secs_per_byte
                    109: ){
1.1       timbl     110: 
1.2       timbl     111:     HTPresentation * pres = (HTPresentation *)malloc(sizeof(HTPresentation));
                    112:     if (pres == NULL) outofmem(__FILE__, "HTSetPresentation");
                    113:     
                    114:     pres->rep = HTAtom_for(representation_in);
                    115:     pres->rep_out = HTAtom_for(representation_out);
                    116:     pres->converter = converter;
                    117:     pres->command = NULL;              /* Fixed */
                    118:     pres->quality = quality;
                    119:     pres->secs = secs;
                    120:     pres->secs_per_byte = secs_per_byte;
                    121:     pres->command = 0;
                    122:     
                    123:     if (!HTPresentations) HTPresentations = HTList_new();
                    124:     
                    125:     if (strcmp(representation_in, "*")==0) {
                    126:         if (default_presentation) free(default_presentation);
                    127:        default_presentation = pres;
                    128:     } else {
                    129:         HTList_addObject(HTPresentations, pres);
                    130:     }
                    131: }
1.1       timbl     132: 
                    133: 
                    134: 
                    135: /*     File buffering
                    136: **     --------------
                    137: **
                    138: **     The input file is read using the macro which can read from
                    139: **     a socket or a file.
                    140: **     The input buffer size, if large will give greater efficiency and
                    141: **     release the server faster, and if small will save space on PCs etc.
                    142: */
                    143: #define INPUT_BUFFER_SIZE 4096         /* Tradeoff */
                    144: PRIVATE char input_buffer[INPUT_BUFFER_SIZE];
                    145: PRIVATE char * input_pointer;
                    146: PRIVATE char * input_limit;
                    147: PRIVATE int input_file_number;
                    148: 
                    149: 
                    150: /*     Set up the buffering
                    151: **
                    152: **     These routines are public because they are in fact needed by
                    153: **     many parsers, and on PCs and Macs we should not duplicate
                    154: **     the static buffer area.
                    155: */
                    156: PUBLIC void HTInitInput ARGS1 (int,file_number)
                    157: {
                    158:     input_file_number = file_number;
                    159:     input_pointer = input_limit = input_buffer;
                    160: }
                    161: 
                    162: 
                    163: PUBLIC char HTGetChararcter NOARGS
                    164: {
                    165:     char ch;
                    166:     do {
                    167:        if (input_pointer >= input_limit) {
                    168:            int status = NETREAD(
                    169:                    input_file_number, input_buffer, INPUT_BUFFER_SIZE);
                    170:            if (status <= 0) {
                    171:                if (status == 0) return (char)EOF;
                    172:                if (TRACE) fprintf(stderr,
                    173:                    "HTFormat: File read error %d\n", status);
                    174:                return (char)EOF; /* -1 is returned by UCX at end of HTTP link */
                    175:            }
                    176:            input_pointer = input_buffer;
                    177:            input_limit = input_buffer + status;
                    178:        }
                    179:        ch = *input_pointer++;
                    180:     } while (ch == (char) 13); /* Ignore ASCII carriage return */
                    181:     
                    182:     return FROMASCII(ch);
                    183: }
                    184: 
                    185: /*     Stream the data to an ouput file as binary
                    186: */
                    187: PUBLIC int HTOutputBinary ARGS2( int,          input,
                    188:                                  FILE *,       output)
                    189: {
                    190:     do {
                    191:            int status = NETREAD(
                    192:                    input, input_buffer, INPUT_BUFFER_SIZE);
                    193:            if (status <= 0) {
                    194:                if (status == 0) return 0;
                    195:                if (TRACE) fprintf(stderr,
                    196:                    "HTFormat: File read error %d\n", status);
                    197:                return 2;                       /* Error */
                    198:            }
                    199:            fwrite(input_buffer, sizeof(char), status, output);
                    200:     } while (YES);
                    201: }
                    202: 
                    203: 
1.2       timbl     204: /*             Create a filter stack
                    205: **             ---------------------
                    206: **
                    207: **     If a widlcard match is made, a temporary HTPresentation
                    208: **     structure is made to hold the destination format while the
                    209: **     new stack is generated. This is just to pass the out format to
                    210: **     MIME so far.  Storing the format of a stream in the stream might
                    211: **     be a lot neater.
                    212: */
                    213: PUBLIC HTStream * HTStreamStack ARGS4(
                    214:        HTFormat,               format_in,
                    215:        HTFormat,               rep_out,
                    216:        HTStream*,              sink,
                    217:        HTParentAnchor*,        anchor)
                    218: {
                    219:     HTAtom * wildcard = HTAtom_for("*");
                    220:     HTPresentation temp;
                    221:     if (TRACE) fprintf(stderr,
                    222:        "HTFormat: Constructing stream stack for %s to %s\n",
                    223:        HTAtom_name(format_in), 
                    224:        HTAtom_name(rep_out));
                    225:                
                    226:     if (rep_out == WWW_SOURCE ||
                    227:        rep_out == format_in) return sink;
                    228: 
                    229:     if (!HTPresentations) HTFormatInit();      /* set up the list */
                    230:     
                    231:     {
                    232:        int n = HTList_count(HTPresentations);
                    233:        int i;
                    234:        HTPresentation * pres;
                    235:        for(i=0; i<n; i++) {
                    236:            pres = HTList_objectAt(HTPresentations, i);
                    237:            if (pres->rep == format_in) {
                    238:                if (pres->rep_out == rep_out)
                    239:                    return (*pres->converter)(pres, anchor, sink);
                    240:                if (pres->rep_out == wildcard) {
                    241:                    temp = *pres;/* make temp conversion to needed fmt */
                    242:                    temp.rep_out = rep_out;             /* yuk */
                    243:                    return (*pres->converter)(&temp, anchor, sink);
                    244:                }
                    245:            }
                    246:        }
                    247:     }
                    248:     
1.3     ! timbl     249: #ifdef XMOSAIC_HACK
        !           250:     return sink;
        !           251: #else
1.2       timbl     252:     return NULL;
1.3     ! timbl     253: #endif
1.2       timbl     254: }
                    255:        
                    256: 
                    257: /*             Find the cost of a filter stack
                    258: **             -------------------------------
                    259: **
                    260: **     Must return the cost of the same stack which StreamStack would set up.
                    261: **
                    262: ** On entry,
                    263: **     length  The size of the data to be converted
                    264: */
                    265: PUBLIC float HTStackValue ARGS4(
                    266:        HTFormat,               format_in,
                    267:        HTFormat,               rep_out,
                    268:        float,                  initial_value,
                    269:        long int,               length)
                    270: {
                    271:     HTAtom * wildcard = HTAtom_for("*");
                    272: 
                    273:     if (TRACE) fprintf(stderr,
                    274:        "HTFormat: Evaluating stream stack for %s worth %.3f to %s\n",
                    275:        HTAtom_name(format_in), initial_value,
                    276:        HTAtom_name(rep_out));
                    277:                
                    278:     if (rep_out == WWW_SOURCE ||
                    279:        rep_out == format_in) return 0.0;
                    280: 
                    281:     if (!HTPresentations) HTFormatInit();      /* set up the list */
                    282:     
                    283:     {
                    284:        int n = HTList_count(HTPresentations);
                    285:        int i;
                    286:        HTPresentation * pres;
                    287:        for(i=0; i<n; i++) {
                    288:            pres = HTList_objectAt(HTPresentations, i);
                    289:            if (pres->rep == format_in && (
                    290:                        pres->rep_out == rep_out ||
                    291:                        pres->rep_out == wildcard)) {
                    292:                float value = initial_value * pres->quality;
                    293:                if (HTMaxSecs != 0.0)
                    294:                value = value - (length*pres->secs_per_byte + pres->secs)
                    295:                                         /HTMaxSecs;
                    296:                return value;
                    297:            }
                    298:        }
                    299:     }
                    300:     
                    301:     return -1e30;              /* Really bad */
                    302: 
                    303: }
                    304:        
1.1       timbl     305: 
1.2       timbl     306: /*     Push data from a socket down a stream
                    307: **     -------------------------------------
1.1       timbl     308: **
1.2       timbl     309: **   This routine is responsible for creating and PRESENTING any
1.1       timbl     310: **   graphic (or other) objects described by the file.
1.2       timbl     311: **
                    312: **   The file number given is assumed to be a TELNET stream ie containing
                    313: **   CRLF at the end of lines which need to be stripped to LF for unix
                    314: **   when the format is textual.
                    315: **
1.1       timbl     316: */
1.2       timbl     317: PUBLIC void HTCopy ARGS2(
                    318:        int,                    file_number,
                    319:        HTStream*,              sink)
1.1       timbl     320: {
1.2       timbl     321:     HTStreamClass targetClass;    
                    322:     
                    323: /*     Push the data, ignoring CRLF, down the stream
                    324: **
                    325: */
                    326:     targetClass = *(sink->isa);        /* Copy pointers to procedures */
                    327:     
                    328:     /* Push binary from socket down sink
                    329:     */
                    330:     for(;;) {
                    331:        int status = NETREAD(
                    332:                file_number, input_buffer, INPUT_BUFFER_SIZE);
                    333:        if (status <= 0) {
                    334:            if (status == 0) break;
                    335:            if (TRACE) fprintf(stderr,
                    336:                "HTFormat: Read error, read returns %d\n", status);
                    337:            break;
                    338:        }
                    339:        (*targetClass.write)(sink, input_buffer, status);
                    340:     } /* next bufferload */
                    341:        
                    342: }
                    343: 
1.1       timbl     344: 
1.2       timbl     345: /*     Push data from a socket down a stream STRIPPING CR
                    346: **     --------------------------------------------------
                    347: **
                    348: **   This routine is responsible for creating and PRESENTING any
                    349: **   graphic (or other) objects described by the file.
                    350: **
                    351: **   The file number given is assumed to be a TELNET stream ie containing
                    352: **   CRLF at the end of lines which need to be stripped to LF for unix
                    353: **   when the format is textual.
                    354: **
1.1       timbl     355: */
1.2       timbl     356: PUBLIC void HTCopyNoCR ARGS2(
                    357:        int,                    file_number,
                    358:        HTStream*,              sink)
                    359: {
                    360:     HTStreamClass targetClass;    
1.1       timbl     361:     
1.2       timbl     362: /*     Push the data, ignoring CRLF, down the stream
                    363: **
                    364: */
                    365:     targetClass = *(sink->isa);        /* Copy pointers to procedures */
                    366: 
                    367: /*     Push text from telnet socket down sink
                    368: **
                    369: **     @@@@@ To push strings could be faster? (especially is we
                    370: **     cheat and don't ignore CR! :-}
                    371: */  
1.1       timbl     372:     HTInitInput(file_number);
1.2       timbl     373:     for(;;) {
                    374:        char character;
                    375:        character = HTGetChararcter();
                    376:        if (character == (char)EOF) break;
                    377:        (*targetClass.put_character)(sink, character);           
                    378:     }
                    379: }
1.1       timbl     380: 
1.2       timbl     381: 
                    382: /*     Parse a socket given format and file number
                    383: **
                    384: **   This routine is responsible for creating and PRESENTING any
                    385: **   graphic (or other) objects described by the file.
                    386: **
                    387: **   The file number given is assumed to be a TELNET stream ie containing
                    388: **   CRLF at the end of lines which need to be stripped to LF for unix
                    389: **   when the format is textual.
                    390: **
                    391: */
                    392: PUBLIC int HTParseSocket ARGS5(
                    393:        HTFormat,               format_in,
                    394:        HTFormat,               format_out,
                    395:        HTParentAnchor *,       anchor,
                    396:        int,                    file_number,
                    397:        HTStream*,              sink)
                    398: {
                    399:     HTStream * stream;
                    400:     HTStreamClass targetClass;    
1.1       timbl     401: 
1.2       timbl     402:     stream = HTStreamStack(format_in,
                    403:                        format_out,
                    404:                        sink , anchor);
                    405:     
                    406:     if (!stream) {
                    407:         char buffer[1024];     /* @@@@@@@@ */
                    408:        sprintf(buffer, "Sorry, can't convert from %s to %s.",
                    409:                HTAtom_name(format_in), HTAtom_name(format_out));
1.3     ! timbl     410:        if (TRACE) fprintf(stderr, "HTFormat: %s\n", buffer);
1.2       timbl     411:         return HTLoadError(sink, 501, buffer);
                    412:     }
1.1       timbl     413:     
1.3     ! timbl     414: /*     Push the data, ignoring CRLF if necessary, down the stream
        !           415: **
1.2       timbl     416: **
1.3     ! timbl     417: **   @@  Bug:  This decision ought to be made based on "encoding"
        !           418: **   rather than on format.  @@@  Whne we handle encoding.
        !           419: **   The current method smells anyway.
1.2       timbl     420: */
                    421:     targetClass = *(stream->isa);      /* Copy pointers to procedures */
1.3     ! timbl     422:     if (format_in == WWW_BINARY || HTOutputSource
        !           423:         || strstr(HTAtom_name(format_in), "image/")
        !           424:        || strstr(HTAtom_name(format_in), "video/")) { /* @@@@@@ */
1.2       timbl     425:         HTCopy(file_number, stream);
                    426:     } else {   /* ascii text with CRLFs :-( */
                    427:         HTCopyNoCR(file_number, stream);
                    428:     }
                    429:     (*targetClass.end_document)(stream);
                    430:     (*targetClass.free)(stream);
1.1       timbl     431:     
1.2       timbl     432:     return HT_LOADED;
1.1       timbl     433: }
1.2       timbl     434: 
                    435: 
                    436: 

Webmaster