Annotation of libwww/Library/src/HTMIME.c, revision 2.67
2.15 frystyk 1: /* HTMIME.c
2: ** MIME MESSAGE PARSE
3: **
2.22 frystyk 4: ** (c) COPYRIGHT MIT 1995.
2.15 frystyk 5: ** Please first read the full copyright statement in the file COPYRIGH.
2.67 ! frystyk 6: ** @(#) $Id: HTMIME.c,v 2.66 1996/07/02 22:54:42 frystyk Exp $
2.1 timbl 7: **
8: ** This is RFC 1341-specific code.
9: ** The input stream pushed into this parser is assumed to be
10: ** stripped on CRs, ie lines end with LF, not CR LF.
11: ** (It is easy to change this except for the body part where
12: ** conversion can be slow.)
13: **
14: ** History:
15: ** Feb 92 Written Tim Berners-Lee, CERN
2.13 duns 16: ** 8 Jul 94 FM Insulate free() from _free structure element.
2.18 frystyk 17: ** 14 Mar 95 HFN Now using anchor for storing data. No more `\n',
18: ** static buffers etc.
2.1 timbl 19: */
2.17 frystyk 20:
21: /* Library include files */
2.57 frystyk 22: #include "sysdep.h"
2.60 frystyk 23: #include "WWWUtil.h"
2.61 frystyk 24: #include "WWWCore.h"
25: #include "HTReqMan.h"
26: #include "HTNetMan.h"
2.36 frystyk 27: #include "HTHeader.h"
2.64 eric 28: #include "HTWWWStr.h"
2.14 frystyk 29: #include "HTMIME.h" /* Implemented here */
2.1 timbl 30:
2.64 eric 31: #define MIME_HASH_SIZE 101
32:
2.1 timbl 33: /* MIME Object
34: ** -----------
35: */
36: struct _HTStream {
2.57 frystyk 37: const HTStreamClass * isa;
2.18 frystyk 38: HTRequest * request;
2.32 frystyk 39: HTNet * net;
40: HTParentAnchor * anchor;
2.18 frystyk 41: HTStream * target;
42: HTFormat target_format;
2.64 eric 43: HTChunk * token;
44: HTChunk * value;
45: int hash;
2.59 frystyk 46: HTEOLState EOLstate;
2.18 frystyk 47: BOOL transparent;
2.48 frystyk 48: BOOL head_only;
2.35 frystyk 49: BOOL nntp;
2.62 frystyk 50: BOOL footer;
2.64 eric 51: BOOL haveToken;
2.1 timbl 52: };
53:
2.18 frystyk 54: /* ------------------------------------------------------------------------- */
2.1 timbl 55:
2.64 eric 56: PRIVATE int pumpData (HTStream * me)
2.18 frystyk 57: {
2.64 eric 58: HTRequest * request = me->request;
59: HTParentAnchor * anchor = me->anchor;
2.48 frystyk 60: me->transparent = YES; /* Pump rest of data right through */
2.27 frystyk 61:
2.48 frystyk 62: /* If this request us a source in PostWeb then pause here */
2.66 frystyk 63: if (HTRequest_isSource(request)) return HT_PAUSE;
2.47 frystyk 64:
2.48 frystyk 65: /* If HEAD method then we just stop here */
2.66 frystyk 66: if (me->head_only || me->footer ||
67: request->method == METHOD_HEAD) return HT_LOADED;
2.43 frystyk 68:
2.60 frystyk 69: /*
70: ** Handle any Content Type
71: ** News server almost never send content type or content length
72: */
2.61 frystyk 73: {
74: HTFormat format = HTAnchor_format(anchor);
75: if (format != WWW_UNKNOWN || me->nntp) {
76: if (STREAM_TRACE) HTTrace("Building.... C-T stack from %s to %s\n",
77: HTAtom_name(format),
78: HTAtom_name(me->target_format));
79: me->target = HTStreamStack(format, me->target_format,
80: me->target, request, YES);
81: }
2.18 frystyk 82: }
2.60 frystyk 83:
84: /* Handle any Content Encoding */
2.61 frystyk 85: {
86: HTList * cc = HTAnchor_encoding(anchor);
87: if (cc) {
88: if (STREAM_TRACE) HTTrace("Building.... C-E stack\n");
89: me->target = HTContentDecodingStack(cc, me->target, request, NULL);
90: }
2.60 frystyk 91: }
92:
93: /* Handle any Transfer encoding */
2.61 frystyk 94: {
95: HTEncoding transfer = HTAnchor_transfer(anchor);
96: if (!HTFormat_isUnityTransfer(transfer)) {
97: if (STREAM_TRACE) HTTrace("Building.... C-T-E stack\n");
98: me->target = HTTransferCodingStack(transfer, me->target,
99: request, NULL, NO);
100: }
101: }
2.27 frystyk 102: return HT_OK;
2.1 timbl 103: }
104:
2.65 eric 105: /* _dispatchParsers - call request's MIME header parser.
106: ** Use global parser if no appropriate one is found for request.
107: */
108: PRIVATE int _dispatchParsers (HTStream * me)
109: {
110: int status;
111: char * token = HTChunk_data(me->token);
112: char * value = HTChunk_data(me->value);
113: BOOL found, local;
114: HTMIMEParseSet * parseSet;
115:
116: /* In case we get an empty header consisting of a CRLF, we fall thru */
2.66 frystyk 117: if (STREAM_TRACE) HTTrace("MIME header. %s: %s\n",
118: token ? token : "<null>",
119: value ? value : "<null>");
120: if (!token) return HT_OK; /* Ignore noop token */
2.65 eric 121:
122: if ((parseSet = HTRequest_MIMEParseSet(me->request, &local)) != NULL) {
123: status = HTMIMEParseSet_dispatch(parseSet, me->request,
124: token, value, &found);
125: if (found)
126: return status;
127: if (local)
128: return HT_OK; /* not found, but that's OK */
129: }
130:
131: if ((parseSet = HTHeader_MIMEParseSet()) == NULL)
132: return HT_OK;
133: status = HTMIMEParseSet_dispatch(parseSet, me->request,
134: token, value, &found);
135: if (found)
136: return status;
137: if (STREAM_TRACE) HTTrace("Ignoring MIME header: %s: %s.\n", token, value);
138:
139: return HT_OK;
140: }
141:
2.18 frystyk 142: /*
143: ** Header is terminated by CRCR, LFLF, CRLFLF, CRLFCRLF
144: ** Folding is either of CF LWS, LF LWS, CRLF LWS
145: */
2.57 frystyk 146: PRIVATE int HTMIME_put_block (HTStream * me, const char * b, int l)
2.18 frystyk 147: {
2.57 frystyk 148: const char * start = b;
149: const char * end = start;
2.64 eric 150: const char * value = me->value->size ? b : NULL;
151: long cl;
152: int status;
153: /* enum {Line_CHAR, Line_END, Line_FOLD, Line_LINE} line = Line_CHAR; */
154:
155: while (!me->transparent) {
2.18 frystyk 156: if (me->EOLstate == EOL_FCR) {
2.64 eric 157: if (*b == CR) /* End of header */
158: me->EOLstate = EOL_END;
159: else if (*b == LF) /* CRLF */
2.18 frystyk 160: me->EOLstate = EOL_FLF;
2.64 eric 161: else if (WHITE(*b)) /* Folding: CR SP */
162: me->EOLstate = EOL_FOLD;
163: else /* New line */
164: me->EOLstate = EOL_LINE;
2.18 frystyk 165: } else if (me->EOLstate == EOL_FLF) {
166: if (*b == CR) /* LF CR or CR LF CR */
167: me->EOLstate = EOL_SCR;
2.64 eric 168: else if (*b == LF) /* End of header */
169: me->EOLstate = EOL_END;
170: else if (WHITE(*b)) /* Folding: LF SP or CR LF SP */
171: me->EOLstate = EOL_FOLD;
172: else /* New line */
173: me->EOLstate = EOL_LINE;
174: } else if (me->EOLstate == EOL_SCR) {
175: if (*b==CR || *b==LF) /* End of header */
176: me->EOLstate = EOL_END;
177: else if (WHITE(*b)) /* Folding: LF CR SP or CR LF CR SP */
178: me->EOLstate = EOL_FOLD;
179: else /* New line */
180: me->EOLstate = EOL_LINE;
181: } else if (*b == CR)
182: me->EOLstate = EOL_FCR;
183: else if (*b == LF)
184: me->EOLstate = EOL_FLF; /* Line found */
185: else {
186: if (!me->haveToken) {
187: if (*b == ':' || isspace(*b)) {
188: HTChunk_putb(me->token, start, end-start);
189: HTChunk_putc(me->token, '\0');
190: me->haveToken = YES;
191: } else {
192: unsigned char ch = *(unsigned char *) b;
193: tolower(ch);
194: /* if (ch >= 'A' && ch <= 'Z')
195: ch += ('a' - 'A'); */
196: me->hash = (me->hash * 3 + ch) % MIME_HASH_SIZE;
197: }
198: } else if (value == NULL && *b != ':' && !isspace(*b))
199: value = b;
200: end++;
201: }
202: switch (me->EOLstate) {
203: case EOL_LINE:
204: case EOL_END: {
205: int status;
206: HTChunk_putb(me->value, value, end-value);
207: HTChunk_putc(me->value, '\0');
208: start=b, end=b;
2.65 eric 209: status = _dispatchParsers(me);
2.64 eric 210: if (me->EOLstate == EOL_END) { /* EOL_END */
2.67 ! frystyk 211: if (status == HT_OK) {
! 212: b++, l--;
2.64 eric 213: status = pumpData(me);
2.67 ! frystyk 214: }
2.64 eric 215: HTNet_setBytesRead(me->net, l);
216: } else { /* EOL_LINE */
217: HTChunk_clear(me->token);
218: HTChunk_clear(me->value);
219: me->haveToken = NO;
220: me->hash = 0;
221: value = NULL;
222: }
2.18 frystyk 223: me->EOLstate = EOL_BEGIN;
2.27 frystyk 224: if (status != HT_OK)
225: return status;
2.64 eric 226: break;
227: }
228: case EOL_FOLD:
2.18 frystyk 229: me->EOLstate = EOL_BEGIN;
2.64 eric 230: if (!me->haveToken) {
231: HTChunk_putb(me->token, start, end-start);
232: HTChunk_putc(me->token, '\0');
233: me->haveToken = YES;
234: } else if (value) {
235: HTChunk_putb(me->value, value, end-value);
236: HTChunk_putc(me->value, ' ');
237: }
238: start=b, end=b;
239: break;
240: default:
241: b++;
242: l--;
243: if (!l) {
244: if (!me->haveToken)
245: HTChunk_putb(me->token, start, end-start);
246: else if (value)
247: HTChunk_putb(me->value, value, end-value);
248: return HT_OK;
249: }
250: }
2.18 frystyk 251: }
2.32 frystyk 252:
253: /*
254: ** Put the rest down the stream without touching the data but make sure
255: ** that we get the correct content length of data
256: */
2.66 frystyk 257: if (me->target) {
258: if ((status = (*me->target->isa->put_block)(me->target, b, l)) != HT_OK)
259: return status;
260: /* Check if CL at all - thanks to jwei@hal.com (John Wei) */
261: cl = HTAnchor_length(me->anchor);
262: return (cl>=0 && HTNet_bytesRead(me->net)>=cl) ? HT_LOADED : HT_OK;
263: }
264: return HT_LOADED;
2.18 frystyk 265: }
266:
267:
268: /* Character handling
269: ** ------------------
270: */
2.36 frystyk 271: PRIVATE int HTMIME_put_character (HTStream * me, char c)
2.18 frystyk 272: {
273: return HTMIME_put_block(me, &c, 1);
274: }
275:
2.1 timbl 276:
277: /* String handling
278: ** ---------------
279: */
2.57 frystyk 280: PRIVATE int HTMIME_put_string (HTStream * me, const char * s)
2.1 timbl 281: {
2.18 frystyk 282: return HTMIME_put_block(me, s, (int) strlen(s));
2.1 timbl 283: }
284:
285:
2.18 frystyk 286: /* Flush an stream object
287: ** ---------------------
2.1 timbl 288: */
2.36 frystyk 289: PRIVATE int HTMIME_flush (HTStream * me)
2.1 timbl 290: {
2.47 frystyk 291: return me->target ? (*me->target->isa->flush)(me->target) : HT_OK;
2.1 timbl 292: }
293:
2.18 frystyk 294: /* Free a stream object
295: ** --------------------
2.1 timbl 296: */
2.36 frystyk 297: PRIVATE int HTMIME_free (HTStream * me)
2.1 timbl 298: {
2.18 frystyk 299: int status = HT_OK;
2.64 eric 300: if (!me->transparent)
2.65 eric 301: if (_dispatchParsers(me) == HT_OK)
2.64 eric 302: pumpData(me);
2.25 frystyk 303: if (me->target) {
304: if ((status = (*me->target->isa->_free)(me->target))==HT_WOULD_BLOCK)
305: return HT_WOULD_BLOCK;
306: }
2.26 frystyk 307: if (PROT_TRACE)
2.55 eric 308: HTTrace("MIME........ FREEING....\n");
2.64 eric 309: HTChunk_delete(me->token);
310: HTChunk_delete(me->value);
2.52 frystyk 311: HT_FREE(me);
2.18 frystyk 312: return status;
2.1 timbl 313: }
314:
315: /* End writing
316: */
2.38 frystyk 317: PRIVATE int HTMIME_abort (HTStream * me, HTList * e)
2.1 timbl 318: {
2.18 frystyk 319: int status = HT_ERROR;
2.41 frystyk 320: if (me->target) status = (*me->target->isa->abort)(me->target, e);
2.26 frystyk 321: if (PROT_TRACE)
2.55 eric 322: HTTrace("MIME........ ABORTING...\n");
2.64 eric 323: HTChunk_delete(me->token);
324: HTChunk_delete(me->value);
2.52 frystyk 325: HT_FREE(me);
2.18 frystyk 326: return status;
2.1 timbl 327: }
328:
329:
330:
331: /* Structured Object Class
332: ** -----------------------
333: */
2.57 frystyk 334: PRIVATE const HTStreamClass HTMIME =
2.1 timbl 335: {
336: "MIMEParser",
2.18 frystyk 337: HTMIME_flush,
2.1 timbl 338: HTMIME_free,
2.6 timbl 339: HTMIME_abort,
340: HTMIME_put_character,
341: HTMIME_put_string,
2.18 frystyk 342: HTMIME_put_block
2.1 timbl 343: };
344:
345:
2.48 frystyk 346: /* MIME header parser stream.
2.1 timbl 347: ** -------------------------
2.48 frystyk 348: ** This stream parses a complete MIME header and if a content type header
349: ** is found then the stream stack is called. Any left over data is pumped
350: ** right through the stream
2.1 timbl 351: */
2.36 frystyk 352: PUBLIC HTStream* HTMIMEConvert (HTRequest * request,
353: void * param,
354: HTFormat input_format,
355: HTFormat output_format,
356: HTStream * output_stream)
2.1 timbl 357: {
2.62 frystyk 358: HTStream * me;
2.52 frystyk 359: if ((me = (HTStream *) HT_CALLOC(1, sizeof(* me))) == NULL)
360: HT_OUTOFMEM("HTMIMEConvert");
2.1 timbl 361: me->isa = &HTMIME;
2.18 frystyk 362: me->request = request;
2.32 frystyk 363: me->anchor = request->anchor;
364: me->net = request->net;
2.49 frystyk 365: me->target = output_stream;
2.18 frystyk 366: me->target_format = output_format;
2.64 eric 367: me->token = HTChunk_new(256);
368: me->value = HTChunk_new(256);
369: me->hash = 0;
2.18 frystyk 370: me->EOLstate = EOL_BEGIN;
2.64 eric 371: me->haveToken = NO;
2.1 timbl 372: return me;
373: }
2.32 frystyk 374:
2.48 frystyk 375: /* MIME header ONLY parser stream
376: ** ------------------------------
377: ** This stream parses a complete MIME header and then returnes HT_PAUSE.
378: ** It does not set up any streams and resting data stays in the buffer.
379: ** This can be used if you only want to parse the headers before you
380: ** decide what to do next. This is for example the case in a server app.
381: */
382: PUBLIC HTStream * HTMIMEHeader (HTRequest * request,
383: void * param,
384: HTFormat input_format,
385: HTFormat output_format,
386: HTStream * output_stream)
387: {
2.62 frystyk 388: HTStream * me = HTMIMEConvert(request, param, input_format,
389: output_format, output_stream);
390: me->head_only = YES;
2.48 frystyk 391: return me;
392: }
2.62 frystyk 393:
394: /* MIME footer ONLY parser stream
395: ** ------------------------------
396: ** Parse only a footer, for example after a chunked encoding.
397: */
398: PUBLIC HTStream * HTMIMEFooter (HTRequest * request,
399: void * param,
400: HTFormat input_format,
401: HTFormat output_format,
402: HTStream * output_stream)
403: {
404: HTStream * me = HTMIMEConvert(request, param, input_format,
405: output_format, output_stream);
406: me->footer = YES;
2.67 ! frystyk 407: me->EOLstate = EOL_FLF;
2.62 frystyk 408: return me;
409: }
Webmaster