PRIVATE int SendCommand (HTRequest *request, news_info *news, char *token, char *pars) { HTStream * input = HTRequest_inputStream(request); int len = strlen(token) + (pars ? strlen(pars)+1:0) + 2; HTChunk_setSize(news->cmd, len); if (pars && *pars) sprintf(HTChunk_data(news->cmd), "%s %s%c%c", token, pars, CR, LF); else sprintf(HTChunk_data(news->cmd), "%s%c%c", token, CR, LF); HTTRACE(PROT_TRACE, "News Tx..... %s" _ HTChunk_data(news->cmd)); return (*input->isa->put_block)(input, HTChunk_data(news->cmd), len); }
PRIVATE void HTML_end_element (HTStructured * me, int element_number) { if (!me->started) { HTextImp_build(me->text, HTEXT_BEGIN); me->started = YES; } /* Update our parse stack */ if (me->overflow > 0) { me->overflow--; return; } me->sp++; if (me->sp > me->stack + MAX_NESTING - 1) { HTTRACE(SGML_TRACE, "HTML Parser. Bottom of parse stack reached\n"); me->sp = me->stack + MAX_NESTING - 1; } /* Look at what element was closed */ switch(element_number) { case HTML_TITLE: HTAnchor_setTitle(me->node_anchor, HTChunk_data(me->title)); break; case HTML_PRE: if (me->comment_start) HTextImp_addText(me->text, me->comment_start, strlen(me->comment_start)); break; } /* Call out to the layout engine */ HTextImp_endElement(me->text, element_number); }
PRIVATE int terminate_handler (HTRequest * request, HTResponse * response, void * param, int status) { if (status == HT_LOADED && result && HTChunk_data(result)) { HTPrint("%s", HTChunk_data(result)); HTChunk_delete(result); } /* We are done with this request */ HTRequest_delete(request); /* Terminate libwww */ HTProfile_delete(); exit(0); }
/* ** Check whether the application provides us with a cookie or more. */ PRIVATE int HTCookie_beforeFilter (HTRequest * request, void * param, int mode) { if ((CookieMode & HT_COOKIE_SEND) && FindCookie) { HTAssocList * cookies = (*FindCookie)(request, FindCookieContext); if (cookies) { HTChunk * cookie_header = HTChunk_new(64); HTAssocList * cur = cookies; HTAssoc * pres; BOOL first=YES; while ((pres = (HTAssoc *) HTAssocList_nextObject(cur))) { if (!first) HTChunk_putc(cookie_header, ';'); HTChunk_puts(cookie_header, HTAssoc_name(pres)); HTChunk_putc(cookie_header, '='); HTChunk_puts(cookie_header, HTAssoc_value(pres)); first = NO; } HTRequest_addExtraHeader(request, "Cookie", HTChunk_data(cookie_header)); HTChunk_delete(cookie_header); /* Also delete the association list */ HTAssocList_delete(cookies); } } return HT_OK; }
/* L A B E L P A R S E R S */ PRIVATE StateRet_t callErrorHandler(CSParse_t * pCSParse, const char * errorLocation, char demark, StateRet_t errorCode) { char * token = HTChunk_data(pCSParse->token); pCSParse->pParseContext->pTokenError = (char *) errorLocation; return (*pCSParse->pParseContext->pParseErrorHandler)(pCSParse, token, demark, StateRet_ERROR_BAD_CHAR); }
PRIVATE int HTRule_flush (HTStream * me) { if (me) { char * flush = HTChunk_data(me->buffer); if (flush) HTRule_parseLine(rules, flush); HTChunk_clear(me->buffer); } return HT_OK; }
PRIVATE BOOL CSUserList_addLine(HTStream * me) { UserListStruct_t * newEl; char * pURL; char * end; end = strrchr(HTChunk_data(me->buffer), ' '); pURL = end + 1; while (*(end-1) == ' ') end--; *end = 0; if ((newEl = (UserListStruct_t *)HT_CALLOC(1, sizeof(UserListStruct_t))) == NULL) HT_OUTOFMEM("UserListStruct_t"); StrAllocCopy(newEl->name, HTChunk_data(me->buffer)); StrAllocCopy(newEl->url, pURL); /* HTList_addObject(me->URLs, (void *)pURL); */ HTList_addObject(UserList, (void *)newEl); return YES; }
/* ** Folding is either of CF LWS, LF LWS, CRLF LWS */ PRIVATE int HTRule_put_block (HTStream * me, const char * b, int l) { while (l > 0) { if (me->EOLstate == EOL_FCR) { if (*b == LF) /* CRLF */ me->EOLstate = EOL_FLF; else if (isspace((int) *b)) /* Folding: CR SP */ me->EOLstate = EOL_DOT; else { /* New line */ HTRule_parseLine(rules, HTChunk_data(me->buffer)); me->EOLstate = EOL_BEGIN; HTChunk_clear(me->buffer); continue; } } else if (me->EOLstate == EOL_FLF) { if (isspace((int) *b)) /* Folding: LF SP or CR LF SP */ me->EOLstate = EOL_DOT; else { /* New line */ HTRule_parseLine(rules, HTChunk_data(me->buffer)); me->EOLstate = EOL_BEGIN; HTChunk_clear(me->buffer); continue; } } else if (me->EOLstate == EOL_DOT) { if (isspace((int) *b)) { me->EOLstate = EOL_BEGIN; HTChunk_putc(me->buffer, ' '); } else { HTRule_parseLine(rules, HTChunk_data(me->buffer)); me->EOLstate = EOL_BEGIN; HTChunk_clear(me->buffer); continue; } } else if (*b == CR) { me->EOLstate = EOL_FCR; } else if (*b == LF) { me->EOLstate = EOL_FLF; /* Line found */ } else HTChunk_putc(me->buffer, *b); l--; b++; } return HT_OK; }
/* _stream2dispatchParsers - extracts the arguments from a * MIME stream before calling the generic _dispatchParser * function. */ PRIVATE int _stream2dispatchParsers (HTStream * me) { char * token = HTChunk_data(me->token); char * value = HTChunk_data(me->value); /* In case we get an empty header consisting of a CRLF, we fall thru */ HTTRACE(STREAM_TRACE, "MIME header. %s: %s\n" _ token ? token : "<null>" _ value ? value : "<null>"); if (!token) return HT_OK; /* Ignore noop token */ /* ** Remember the original header */ HTResponse_addHeader(me->response, token, value); /* call the parsers to set the headers */ return (_dispatchParsers (me->request, token, value)); }
/* Handle entity ** ------------- ** ** On entry, ** s contains the entity name zero terminated */ PRIVATE void handle_entity (HTStream * context) { const char ** entities = context->dtd->entity_names; const char *s = HTChunk_data(context->string); int high, low, i, diff; for(low=0, high = context->dtd->number_of_entities; high > low ; diff < 0 ? (low = i+1) : (high = i)) { i = (low + (high-low)/2); diff = strcmp(entities[i], s); /* Case sensitive! */ if (diff==0) { /* success: found it */ (*context->actions->put_entity)(context->target, i); return; } } /* If entity string not found */ HTTRACE(SGML_TRACE, "Unknown entity %s\n" _ s); (*context->actions->unparsed_entity) (context->target, HTChunk_data(context->string), HTChunk_size(context->string)); }
PRIVATE int NewsPost_put_block (HTStream * me, const char* b, int l) { if (!me->target) { return HT_WOULD_BLOCK; } else if (me->transparent) return b ? PUTBLOCK(b, l) : HT_OK; else { int status; NewsPost_start(me, me->request); if ((status = PUTBLOCK(HTChunk_data(me->buffer), HTChunk_size(me->buffer))) == HT_OK) { me->transparent = YES; return b ? PUTBLOCK(b, l) : HT_OK; } return status; } }
/* Handle attribute value ** ---------------------- */ PRIVATE void handle_attribute_value (HTStream * context) { /* Deal with attributes only if tag is known, ignore silently otherwise */ if (context->current_tag) { if (context->current_attribute_number != INVALID) context->value[context->current_attribute_number] = context->token; else { char * data = HTChunk_data(context->string); HTTRACE(SGML_TRACE, "Attribute value %s ignored\n" _ data ? data+context->token : "<null>"); } } context->current_attribute_number = INVALID; /* can't have two assignments! */ }
/* NewsPost_start ** -------------- ** NNTP needs two extra headers: "From" and "Newsgroups". ** Take the newsgroups from the Postweb model as destinations for this ** anchor. ** Return YES if OK else NO */ PRIVATE BOOL NewsPost_start (HTStream * me, HTRequest * request) { char linebuf[128]; /* @@@ */ HTChunk *header = me->buffer; HTUserProfile * up = HTRequest_userProfile(request); const char * mailaddress = HTUserProfile_email(up); if (mailaddress) { sprintf(linebuf, "From: %s%c%c", mailaddress, CR, LF); HTChunk_puts(header, linebuf); } /* ** Find all the newsgroups we are posting to by looking at all the ** destinations from the source of this request. ** First the main link and then the sub links */ HTChunk_puts(header, "Newsgroups :"); if (HTRequest_isDestination(request)) { HTRequest *src_req = HTRequest_source(request); HTParentAnchor *src_anchor = HTRequest_anchor(src_req); HTLink *link = HTAnchor_mainLink((HTAnchor *) src_anchor); HTAnchor *dest = HTLink_destination(link); HTMethod method = HTLink_method(link); if (link && method == METHOD_POST && HTLink_result(link) == HT_LINK_NONE) { char *desturl = HTAnchor_physical((HTParentAnchor *) dest); char *access = HTParse(desturl, "", PARSE_ACCESS); if (!strcasecomp(access, "news") || !strcasecomp(access, "nntp")) { char *newsgroup = HTParse(desturl, "", PARSE_PATH); HTUnEscape(newsgroup); HTCleanTelnetString(newsgroup); HTChunk_puts(header, newsgroup); HT_FREE(newsgroup); } HT_FREE(access); } /* DO FOR ALL SUB ANCHOR DESTINATION S AS WELL */ } HTTRACE(PROT_TRACE, "News Tx..... %s" _ HTChunk_data(header)); return YES; }
/* ** Scan the request line for METHOD, URI and VERSION ** Returns: HT_OK if 1.x request and OK ** HT_LOADED if 0.9 request and OK ** HT_ERROR if invalid request line */ PRIVATE int ParseRequest (HTStream * me) { HTRequest * client = HTList_firstObject(me->http->clients); char * line = HTChunk_data(me->buffer); char * method_str = HTNextField(&line); char * request_uri = HTNextField(&line); char * version_str = HTNextField(&line); HTMethod method; /* Check if method is allowed */ if (!method_str || (method = HTMethod_enum(method_str))==METHOD_INVALID) { HTRequest_addError(client, ERR_FATAL, NO, HTERR_NOT_ALLOWED, NULL, 0, "ParseRequest"); return HT_ERROR; } HTRequest_setMethod(client, method); /* Find an anchor for the request URI */ if (request_uri) { char * uri = HTParse(request_uri, "file:", PARSE_ALL); HTRequest_setAnchor(client, HTAnchor_findAddress(uri)); HT_FREE(uri); } else { HTRequest_addError(client, ERR_FATAL, NO, HTERR_BAD_REQUEST, NULL, 0, "ParseRequest"); return HT_ERROR; } /* Get ready to get the rest of the request */ if (version_str) { me->target = HTStreamStack(WWW_MIME_HEAD, HTRequest_debugFormat(client), HTRequest_debugStream(client), client, NO); return HT_OK; } else { HTRequest_addError(client, ERR_FATAL, NO, HTERR_BAD_VERSION, NULL, 0, "ParseRequest"); return HT_ERROR; } }
/* Start an element ** ---------------- */ PRIVATE void start_element (HTStream * context) { int i; char *value[MAX_ATTRIBUTES]; HTTag *tag = context->current_tag; HTTRACE(SGML_TRACE, "Start <%s>\n" _ tag->name); context->contents = tag->contents; /* ** Build the actual pointers to the value strings stored in the ** chunk buffer. (Must use offsets while collecting the values, ** because the string chunk may get resized during the collection ** and potentially relocated). */ for (i = 0; i < MAX_ATTRIBUTES; ++i) value[i] = context->value[i] < 0 ? NULL : HTChunk_data(context->string) + context->value[i]; (*context->actions->start_element) (context->target, tag - context->dtd->tags, context->present, (const char**)value); /* coerce type for think c */ }
PUBLIC NowIn_t CSParse_targetParser(CSParse_t * pCSParse, char demark, void * pVoid) { /* ParseContext_t * pParseContext = pCSParse->pParseContext; */ TargetObject_t * pTargetObject = pCSParse->pTargetObject; BOOL failedOnPunct = NO; char * token = 0; StateRet_t ret = StateRet_OK; int i; static NowIn_t lastRet = NowIn_END; if (HTChunk_size(pCSParse->token)) { HTChunk_terminate(pCSParse->token); token = HTChunk_data(pCSParse->token); } for (i = 0; i < pTargetObject->stateTokenCount; i++) { StateToken_t * pStateToken = pTargetObject->stateTokens + i; pCSParse->pStateToken = pStateToken; if (!(pCSParse->currentSubState & pStateToken->validSubStates)) continue; if (pStateToken->pCheck) { /* use check function */ StateRet_t checkRes; checkRes = (*pStateToken->pCheck)(pCSParse, pStateToken, token, demark); switch (checkRes) { case StateRet_WARN_BAD_PUNCT: failedOnPunct = YES; case StateRet_WARN_NO_MATCH: continue; case StateRet_ERROR_BAD_CHAR: (*pCSParse->pParseContext->pParseErrorHandler)(pCSParse, token, demark, StateRet_ERROR_BAD_CHAR); /* if (pTargetObject->pDestroy) (*pTargetObject->pDestroy)(pCSParse); */ return NowIn_ERROR; default: break; } } else { /* or match by name[s] */ if (!(pStateToken->command & Command_MATCHANY)) { if (token && pStateToken->name1) { if (strcasecomp(token, pStateToken->name1) && (!pStateToken->name2 || strcasecomp(token, pStateToken->name2))) continue; } else { if (token != pStateToken->name1) continue; } } if (Punct_badDemark(pStateToken->validPunctuation, demark)) { failedOnPunct = YES; continue; } } /* open or close and do the appropriate callbacks */ if (lastRet != NowIn_CHAIN) ParseTrace("%30s %c ", token ? token : "", demark); ParseTrace("%10s - %s:%10s => ", pCSParse->pTargetObject->note, CSParse_subState2str(pCSParse->currentSubState), pStateToken->note); if (pStateToken->command & Command_NOTOKEN) { HTChunk_clear(pCSParse->token); token = 0; } if (pStateToken->command & Command_OPEN && pTargetObject->pOpen) if ((*pTargetObject->pOpen)(pCSParse, token, demark) == StateRet_ERROR) return NowIn_ERROR; if (pStateToken->command & (Command_OPEN|Command_CLOSE) && pCSParse->pParseContext->pTargetChangeCallback) { ParseTrace("%3d", pStateToken->command & Command_CLOSE ? -pTargetObject->targetChange : pTargetObject->targetChange); if ((*pCSParse->pParseContext->pTargetChangeCallback)(pCSParse, pTargetObject, pTargetObject->targetChange, (BOOL)(pStateToken->command & Command_CLOSE), pVoid) == StateRet_ERROR) return NowIn_ERROR; } else ParseTrace(" "); if (pStateToken->command & Command_CLOSE && pTargetObject->pClose) ret = (*pTargetObject->pClose)(pCSParse, token, demark); if (pStateToken->pPrep && ret != NowIn_ERROR) ret = (*pStateToken->pPrep)(pCSParse, token, demark); if (pStateToken->pNextTargetObject) pCSParse->pTargetObject = pStateToken->pNextTargetObject; if (pStateToken->nextSubState != SubState_X) pCSParse->currentSubState = pStateToken->nextSubState; ParseTrace("%10s - %s", pCSParse->pTargetObject->note, CSParse_subState2str(pCSParse->currentSubState)); if (pStateToken->command & Command_CHAIN) { ParseTrace(" -O-O-"); return lastRet = NowIn_CHAIN; } ParseTrace("\n"); return lastRet = ret == StateRet_ERROR_BAD_CHAR ? NowIn_ERROR : ret == StateRet_DONE ? NowIn_END : NowIn_ENGINE; } (*pCSParse->pParseContext->pParseErrorHandler)(pCSParse, token, demark, failedOnPunct ? StateRet_WARN_BAD_PUNCT : StateRet_WARN_NO_MATCH); if (pTargetObject->pDestroy) (*pTargetObject->pDestroy)(pCSParse); return NowIn_ERROR; }
PUBLIC BOOL HTBind_add (const char * suffix, const char * representation, const char * encoding, const char * transfer, const char * language, double value) { HTBind * suff; if (!suffix) return NO; if (!strcmp(suffix, "*")) suff = &no_suffix; else if (!strcmp(suffix, "*.*")) suff = &unknown_suffix; else { HTList * suflist; int hash; const unsigned char * p; /* Select list from hash table */ for (p=suffix, hash=0; *p; p++) { hash = (hash * 3 + TOLOWER(*p)) % HT_L_HASH_SIZE; } if (!HTBindings) HTBind_init(); if (!HTBindings[hash]) HTBindings[hash] = HTList_new(); suflist = HTBindings[hash]; /* Look for existing binding */ { HTList *cur = suflist; while ((suff = (HTBind *) HTList_nextObject(cur)) != NULL) { if (!strcmp(suff->suffix, suffix)) break; } } /* If not found -- create a new node */ if (!suff) { if ((suff = (HTBind *) HT_CALLOC(1, sizeof(HTBind))) == NULL) HT_OUTOFMEM("HTBind_add"); HTList_addObject(suflist, (void *) suff); StrAllocCopy(suff->suffix, suffix); } } /* Set the appropriate values */ { HTChunk * chunk = HTChunk_new(32); char *ptr; if (representation) { HTChunk_puts(chunk, representation); ptr = HTChunk_data(chunk); for (; *ptr; ptr++) *ptr = TOLOWER(*ptr); suff->type = HTAtom_for(HTChunk_data(chunk)); HTChunk_truncate(chunk,0); } if (encoding) { HTChunk_puts(chunk, encoding); ptr = HTChunk_data(chunk); for (; *ptr; ptr++) *ptr = TOLOWER(*ptr); suff->encoding = HTAtom_for(HTChunk_data(chunk)); HTChunk_truncate(chunk,0); } if (transfer) { HTChunk_puts(chunk, transfer); ptr = HTChunk_data(chunk); for (; *ptr; ptr++) *ptr = TOLOWER(*ptr); suff->transfer = HTAtom_for(HTChunk_data(chunk)); HTChunk_truncate(chunk,0); } if (language) { HTChunk_puts(chunk, language); ptr = HTChunk_data(chunk); for (; *ptr; ptr++) *ptr = TOLOWER(*ptr); suff->language = HTAtom_for(HTChunk_data(chunk)); HTChunk_truncate(chunk,0); } HTChunk_delete(chunk); suff->quality = value; } return YES; }
unsigned int BinURLInputStream::readBytes(XMLByte* const toFill , const unsigned int maxToRead) { unsigned int retval = 0; unsigned int bytesAsked = maxToRead; unsigned int bytesForCopy = 0; // Wipe out the old stuff from the destination buffer to fill. memset((void*)toFill, 0x00, sizeof(XMLByte) * maxToRead); // You can only read till the end of the remote resource file. // So, adjust the count of bytes you want to read now. if (fBytesProcessed + bytesAsked >= fRemoteFileSize) { bytesAsked = fRemoteFileSize - fBytesProcessed; } if (fBufferSize > 0) bytesForCopy = fBufferSize - fBufferIndex; if (bytesAsked <= bytesForCopy) { // ...then you can satisfy this request completely from fBuffer. // Simply copy over the bytes to the destination array. memcpy((void*) toFill, (void*) (fBuffer + fBufferIndex), bytesAsked); fBufferIndex += bytesAsked; if (fBufferIndex >= fBufferSize) { fBufferSize = 0; fBufferIndex = 0; } fBytesProcessed += bytesAsked; retval = bytesAsked; } else { // ...will need to read some more bytes out of the stream. unsigned int bufToFillIndex = 0; HTRequest* request = HTRequest_new(); HTChunk* result = NULL; char ranges[64]; // First copy over what is left in fBuffer, before reading another // chunk out of the stream. if (bytesForCopy != 0) { memcpy((void*) toFill, (void*) (fBuffer + fBufferSize), bytesForCopy); fBufferSize = 0; fBufferIndex = 0; fBytesProcessed += bytesForCopy; bufToFillIndex = bytesForCopy; retval = bytesForCopy; } unsigned int bytesRemainingForCopy = bytesAsked - bytesForCopy; // Now read a new chunk from the stream. HTTP lets you specify the // range of bytes that you would like. sprintf(ranges, "%ld-%ld", fBytesProcessed, fRemoteFileSize<(fBytesProcessed + URLISBUFMAXSIZE)? fRemoteFileSize - 1: fBytesProcessed + URLISBUFMAXSIZE - 1); HTRequest_addRange(request, "bytes", ranges); HTRequest_setOutputFormat(request, WWW_SOURCE); result = HTLoadAnchorToChunk(fAnchor, request); fBufferSize = HTChunk_size(result); if (fBufferSize > 0) { // Store the read chunk in fBuffer. memset((void*) fBuffer, 0x00, URLISBUFMAXSIZE); memcpy((void*) fBuffer, (void*) HTChunk_data(result), fBufferSize); fBufferIndex = 0; } HTRequest_delete(request); HTChunk_delete(result); // Now fill the destination buffer with the new data just read. bytesForCopy = fBufferSize; if (bytesRemainingForCopy > fBufferSize) { bytesRemainingForCopy = fBufferSize; } memcpy((void*) (toFill + bufToFillIndex), (void*) fBuffer, bytesRemainingForCopy); // Update counters. retval += bytesRemainingForCopy; fBufferIndex += bytesRemainingForCopy; fBytesProcessed += bytesRemainingForCopy; } return retval; }
PRIVATE int SGML_write (HTStream * context, const char * b, int l) { const SGML_dtd *dtd = context->dtd; HTChunk *string = context->string; const char *text = b; int count = 0; while (l-- > 0) { char c = *b++; switch(context->state) { got_element_open: /* ** The label is jumped when the '>' of a the element ** start tag has been detected. This DOES NOT FALL TO ** THE CODE S_after_open, only processes the tag and ** sets the state (c should still contain the ** terminating character of the tag ('>')) */ if (context->current_tag && context->current_tag->name) start_element(context); context->state = S_after_open; break; case S_after_open: /* ** State S_after_open is entered only for single ** character after the element opening tag to test ** against newline. Strip one trainling newline only ** after opening nonempty element. - SGML: Ugh! */ text = b; count = 0; if (c == '\n' && (context->contents != SGML_EMPTY)) { context->state = S_text; break; } --text; goto S_text; S_text: context->state = S_text; case S_text: #ifdef ISO_2022_JP if (c == '\033') { context->state = S_esc; ++count; break; } #endif /* ISO_2022_JP */ if (c == '&') { if (count > 0) PUTB(text, count); count = 0; HTChunk_clear(string); context->state = S_ero; } else if (c == '<') { if (count > 0) PUTB(text, count); count = 0; HTChunk_clear(string); /* should scrap LITERAL, and use CDATA and RCDATA -- msa */ context->state = (context->contents == SGML_LITERAL) ? S_literal : S_tag; } else if (c == '\n') /* Newline - ignore if before end tag! */ context->state = S_nl; else ++count; break; case S_nl: if (c == '<') { if (count > 0) PUTB(text, count); count = 0; HTChunk_clear(string); context->state = (context->contents == SGML_LITERAL) ? S_literal : S_nl_tago; } else { ++count; goto S_text; } break; case S_nl_tago: /* Had newline and tag opener */ if (c != '/') PUTC('\n'); /* Only ignore newline before </ */ context->state = S_tag; goto handle_S_tag; #ifdef ISO_2022_JP case S_esc: if (c=='$') context->state = S_dollar; else if (c=='(') context->state = S_paren; else context->state = S_text; ++count; break; case S_dollar: if (c=='@' || c=='B') context->state = S_nonascii_text; else context->state = S_text; ++count; break; case S_paren: if (c=='B' || c=='J') context->state = S_text; else context->state = S_text; ++count; break; case S_nonascii_text: if (c == '\033') context->state = S_esc; ++count; break; #endif /* ISO_2022_JP */ /* In literal mode, waits only for specific end tag! ** Only foir compatibility with old servers. */ case S_literal: HTChunk_putc(string, c); if ( TOUPPER(c) != ((HTChunk_size(string) == 1) ? '/' : context->current_tag->name[HTChunk_size(string)-2])) { /* If complete match, end literal */ if ((c == '>') && (!context->current_tag->name[HTChunk_size(string)-2])) { end_element (context,context->current_tag); /* ...setting SGML_MIXED below is a bit of kludge, but a good guess that currently works, anything other than SGML_LITERAL would work... -- msa */ context->contents = SGML_MIXED; } else { /* If Mismatch: recover string. */ PUTC( '<'); PUTB(HTChunk_data(string), HTChunk_size(string)); } context->state = S_text; text = b; count = 0; } break; /* ** Character reference or Entity */ case S_ero: if (c == '#') { /* &# is Char Ref Open */ context->state = S_cro; break; } context->state = S_entity; /** FALL THROUGH TO S_entity !! ***/ /* ** Handle Entities */ case S_entity: if (isalnum((int) c)) HTChunk_putc(string, c); else { HTChunk_terminate(string); handle_entity(context); text = b; count = 0; if (c != ';') { --text; goto S_text; } context->state = S_text; } break; /* Character reference */ case S_cro: if (isalnum((int)c)) /* accumulate a character NUMBER */ HTChunk_putc(string, c); else { int value; HTChunk_terminate(string); if (sscanf(HTChunk_data(string), "%d", &value)==1) PUTC((char)value); else { PUTB("&#", 2); PUTB(HTChunk_data(string), HTChunk_size(string)-1); } text = b; count = 0; if (c != ';') { --text; goto S_text; } context->state = S_text; } break; case S_tag: /* new tag */ handle_S_tag: if (isalnum((int)c)) HTChunk_putc(string, c); else { /* End of tag name */ int i; if (c == '/') { if (HTChunk_size(string) > 0) HTTRACE(SGML_TRACE, "`<%s/' found!\n" _ HTChunk_data(string)); context->state = S_end; break; } else if (c == '!') { if (HTChunk_size(string) > 0) HTTRACE(SGML_TRACE, " `<%s!' found!\n" _ HTChunk_data(string)); context->state = S_md; break; } HTChunk_terminate(string); context->current_tag = SGMLFindTag(dtd, HTChunk_data(string)); if (context->current_tag == NULL) { HTTRACE(SGML_TRACE, "*** Unknown element %s\n" _ HTChunk_data(string)); (*context->actions->unparsed_begin_element) (context->target, HTChunk_data(string), HTChunk_size(string)); } else { for (i=0; i<context->current_tag->number_of_attributes; i++) { context->present[i] = NO; context->value[i] = -1; } } context->token = 0; HTChunk_clear(string); context->current_attribute_number = INVALID; goto S_tag_gap; } break; S_tag_gap: context->state = S_tag_gap; case S_tag_gap: /* Expecting attribute or > */ if (isspace((int) c)) break; /* Gap between attributes */ if (c == '>') goto got_element_open; else goto S_attr; S_attr: /* ** Start collecting the attribute name and collect ** it in S_attr. */ context->state = S_attr; HTChunk_truncate(string, context->token); case S_attr: if (isspace((int) c) || c == '>' || c == '=') goto got_attribute_name; else HTChunk_putc(string, c); break; got_attribute_name: /* ** This label is entered when attribute name has been ** collected. Process it and enter S_attr_gap for ** potential value or start of the next attribute. */ HTChunk_terminate(string) ; handle_attribute_name (context, HTChunk_data(string) + context->token); HTChunk_truncate(string, context->token); context->state = S_attr_gap; case S_attr_gap: /* Expecting attribute or = or > */ if (isspace((int) c)) break; /* Gap after attribute */ if (c == '>') goto got_element_open; else if (c == '=') context->state = S_equals; else goto S_attr; /* Get next attribute */ break; case S_equals: /* After attr = */ if (isspace((int) c)) break; /* Before attribute value */ if (c == '>') { /* End of tag */ HTTRACE(SGML_TRACE, "found = but no value\n"); goto got_element_open; } else if (c == '\'') context->state = S_squoted; else if (c == '"') context->state = S_dquoted; else goto S_value; break; S_value: context->state = S_value; HTChunk_truncate(string, context->token); case S_value: if (isspace((int) c) || c == '>') { HTChunk_terminate(string); handle_attribute_value(context); context->token = HTChunk_size(string); goto S_tag_gap; } else HTChunk_putc(string, c); break; case S_squoted: /* Quoted attribute value */ if (c == '\'') { HTChunk_terminate(string); handle_attribute_value(context); context->token = HTChunk_size(string); context->state = S_tag_gap; } else if (c && c != '\n' && c != '\r') HTChunk_putc(string, c); break; case S_dquoted: /* Quoted attribute value */ if (c == '"') { HTChunk_terminate(string); handle_attribute_value(context); context->token = HTChunk_size(string); context->state = S_tag_gap; } else if (c && c != '\n' && c != '\r') HTChunk_putc(string, c); break; case S_end: /* </ */ if (isalnum((int) c)) HTChunk_putc(string, c); else { /* End of end tag name */ HTTag *t; char * first; HTChunk_terminate(string); if ((first=HTChunk_data(string))!=NULL && *first != '\0') t = SGMLFindTag(dtd, HTChunk_data(string)); else /* Empty end tag */ /* Original code popped here one from the stack. If this feature is required, I have to put the stack back... -- msa */ t = NULL; if (!t) { HTTRACE(SGML_TRACE, "Unknown end tag </%s>\n" _ HTChunk_data(string)); (*context->actions->unparsed_end_element) (context->target, HTChunk_data(string), HTChunk_size(string)); } else { context->current_tag = NULL; end_element(context, t); } HTChunk_clear(string); context->current_attribute_number = INVALID; if (c != '>') { if (!isspace((int) c)) HTTRACE(SGML_TRACE, "`</%s%c' found!\n" _ HTChunk_data(string) _ c); context->state = S_junk_tag; } else { text = b; count = 0; context->state = S_text; } } break; case S_junk_tag: if (c == '>') { text = b; count = 0; context->state = S_text; } break; /* ** Scanning (actually skipping) declarations */ case S_md: if (c == '-') context->state = S_com_1; else if (c == '"') context->state = S_md_dqs; else if (c == '\'') context->state = S_md_sqs; else if (c == '>') { text = b; count = 0; context->state = S_text; } break; case S_md_dqs: /* Skip double quoted string */ if (c == '"') context->state = S_md; else if (c == '>') { text = b; count = 0; context->state = S_text; } break; case S_md_sqs: /* Skip single quoted string */ if (c == '\'') context->state = S_md; else if (c == '>') { text = b; count = 0; context->state = S_text; } break; case S_com_1: /* Starting a comment? */ context->state = (c == '-') ? S_com : S_md; if (c == '>') { text = b; count = 0; context->state = S_text; } break; case S_com: /* ..within comment */ if (c == '-') context->state = S_com_2; break; case S_com_2: /* Ending a comment ? */ context->state = (c == '-') ? S_com_2a : S_com; break; case S_com_2a: if (c == '>') { text = b; count = 0; context->state = S_text; } else if (c == '-') { context->state = S_com_2a; } else context->state = S_com; break; } } if (count > 0) PUTB(text, count); return HT_OK; }
int main (int argc, char ** argv) { HTSQL * sql = NULL; char * sqlserver = DEFAULT_SQL_SERVER; char * sqldb = DEFAULT_SQL_DB; char * sqluser = DEFAULT_SQL_USER; char * sqlpw = DEFAULT_SQL_PW; char * cvsuser = DEFAULT_CVS_USER; time_t cvsdate = -1; FILE * fin = stdin; char * input_buffer[BUFSIZE]; HTChunk * loginfo = HTChunk_new(BUFSIZE); int arg = 0; BOOL create_db = YES; /* Initiate libwww */ HTLibInit(APP_NAME, APP_VERSION); /* Scan command line for parameters */ for (arg=1; arg<argc; arg++) { if (*argv[arg] == '-') { if (!strcmp(argv[arg], "-h") || !strcmp(argv[arg], "-?")) { VersionInfo(); Cleanup(0, sql, loginfo); } else if (!strcmp(argv[arg], "-v")) { HTSetTraceMessageMask("q"); } else if (!strcmp(argv[arg], "-nocreate")) { create_db = NO; } else if (!strncmp(argv[arg], "-sqldb", 5)) { sqldb = (arg+1 < argc && *argv[arg+1] != '-') ? argv[++arg] : DEFAULT_SQL_DB; } else if (!strncmp(argv[arg], "-sqlpassword", 5)) { sqlpw = (arg+1 < argc && *argv[arg+1] != '-') ? argv[++arg] : DEFAULT_SQL_PW; } else if (!strncmp(argv[arg], "-sqlserver", 5)) { sqlserver = (arg+1 < argc && *argv[arg+1] != '-') ? argv[++arg] : DEFAULT_SQL_SERVER; } else if (!strncmp(argv[arg], "-sqluser", 5)) { sqluser = (arg+1 < argc && *argv[arg+1] != '-') ? argv[++arg] : DEFAULT_SQL_USER; } else if (!strncmp(argv[arg], "-cvsuser", 5)) { cvsuser = (arg+1 < argc && *argv[arg+1] != '-') ? argv[++arg] : DEFAULT_CVS_USER; } else if (!strncmp(argv[arg], "-cvsdate", 5)) { cvsdate = (arg+1 < argc && *argv[arg+1] != '-') ? HTParseTime(argv[++arg], NULL, NO) : -1; } else { fprintf(stderr, "Bad Argument (%s)\n", argv[arg]); } } else { fprintf(stderr, "Bad Argument (%s)\n", argv[arg]); } } /* Get an SQL object */ if ((sql = HTSQL_new(sqlserver, sqluser, sqlpw, 0)) == NULL) Cleanup(-1, sql, loginfo); /* Connect to the SQL server */ if (HTSQL_connect(sql) != YES) Cleanup(-1, sql, loginfo); /* Select our database */ if (HTSQL_selectDB(sql, sqldb) != YES) Cleanup(-1, sql, loginfo); /* Create our tables */ if (create_db) createTables(sql, 0); /* Read the arguments from stdin */ for (;;) { int status = fread(input_buffer, 1, BUFSIZE, fin); if (status < 0) Cleanup(-1, sql, loginfo); if (status == 0) break; HTChunk_putb(loginfo, (const char *) input_buffer, status); } /* Parse the input chunk */ { char * ptr = HTChunk_data(loginfo); char * noop1 = HTNextField(&ptr); char * noop2 = HTNextField(&ptr); char * root = HTNextField(&ptr); char * operation = NULL; char * files = NULL; char * comment = NULL; int comment_id = -1; int user_id = -1; char * p, * q; #ifdef HT_REENTRANT char *lasts; /* For strtok_r */ #endif /* Find shared log message and get id */ if ((q = HTStrCaseStr(ptr, "\nLog Message:")) != NULL) { comment = q+14; *q = '\0'; } if ((comment_id = add_comment(sql, comment)) < 0) Cleanup(-1, sql, loginfo); /* Add/find user and get id */ if ((user_id = add_user(sql, cvsuser)) < 0) Cleanup(-1, sql, loginfo); /* For each operation, find the files involved */ while ((q = HTStrCaseStr(ptr, " Files:")) != NULL) { /* Find the operation */ files = q+9; for (p=q; p>HTChunk_data(loginfo) && *p && *p!='\n'; p--); p++; operation = HTNextField(&p); /* Find the next line */ if ((q = strchr(files, '\n')) != NULL) { *q++ = '\0'; ptr = q; } /* Create the query */ if (operation && files && comment) { char * file; int location_id = -1; #ifdef HT_REENTRANT if ((file = strtok_r(files, DELIMITERS, &lasts)) != NULL) { #else if ((file = strtok(files, DELIMITERS)) != NULL) { #endif /* HT_REENTRANT */ do { char * path = NULL; StrAllocMCopy(&path, root, "/", file, NULL); /* Add/find location and get id */ if ((location_id = add_location(sql, path)) < 0) { Cleanup(-1, sql, loginfo); break; } #if 0 fprintf(stderr, "location: `%s\', user: `%s\', operation: `%s\', comment: `%s\'\n", path, cvsuser, operation, comment); #endif /* Add log entry */ { char buf[16384]; char * query = HTSQL_printf(buf, 16384, "insert into %s values (%u,%u,%T,%S,%u)", DEFAULT_SQL_LOG_TABLE, location_id, user_id, cvsdate, operation, comment_id); if (HTSQL_query(sql, query) != YES) { Cleanup(-1, sql, loginfo); break; } } HT_FREE(path); #ifdef HT_REENTRANT } while ((file = (char *) strtok_r(NULL, DELIMITERS, &lasts)) != NULL); #else } while ((file = strtok(NULL, DELIMITERS)) != NULL); #endif /* HT_REENTRANT */ } } } } return 0; }