/*
* HP-UX 8/9.01 strcat reads a word past source when doing unaligned copies!
* Work around it here. The bug report has been filed with HP.
*/
char *strcat(char *d, const char *s)
{
dps_strcpy(d + dps_strlen(d), s);
return d;
}
char * DpsURLNormalizePath(char * str){
char * s=str;
char * q;
char * d;
/* Hide query string */
if((q=strchr(s,'?'))){
*q++='\0';
if(!*q)q=NULL;
}
/* Remove all "/../" entries */
while((d=strstr(str,"/../"))){
char * p;
if(d>str){
/* Skip non-slashes */
for(p=d-1;(*p!='/')&&(p>str);p--);
/* Skip slashes */
while((p>(str+1))&&(*(p-1)=='/'))p--;
}else{
/* We are at the top level and have ../ */
/* Remove it too to avoid loops */
p=str;
}
dps_memmove(p,d+3,dps_strlen(d)-2);
}
/* Remove remove trailig "/.." */
d=str+dps_strlen(str);
if((d-str>2)&&(!strcmp(d-3,"/.."))){
d-=4;
while((d>str)&&(*d!='/'))d--;
if(*d=='/')d[1]='\0';
else dps_strcpy(str,"/");
}
/* Remove all "/./" entries */
while((d=strstr(str,"/./"))){
dps_memmove(d,d+2,dps_strlen(d)-1);
}
/* Remove the trailing "/." */
if((d=str+dps_strlen(str))>(str+2))
if(!strcmp(d-2,"/."))
*(d-1)='\0';
/* Remove all "//" entries */
while((d=strstr(str,"//"))){
dps_memmove(d,d+1,dps_strlen(d));
}
/* Replace "%7E" with "~" */
/* Actually it is to be done */
/* for all valid characters */
/* which do not require escaping */
/* However I'm lazy, do it for 7E */
/* as the most often "abused" */
while((d=strstr(str,"%7E"))){
*d='~';
dps_memmove(d+1,d+3,dps_strlen(d+3)+1);
}
/* Restore query string */
if(q){
char * e=str+dps_strlen(str);
*e='?';
dps_memmove(e+1,q,dps_strlen(q)+1);
}
return str;
}
void RelLink(DPS_AGENT *Indexer, DPS_URL *curURL, DPS_URL *newURL, char **str, int ReverseAliasFlag) {
const char *schema = newURL->schema ? newURL->schema : curURL->schema;
const char *hostname = newURL->hostname ? newURL->hostname : curURL->hostname;
const char *auth = newURL->auth ? newURL->auth : curURL->auth;
const char *path = (newURL->path && newURL->path[0]) ? newURL->path : curURL->path;
const char *fname = ((newURL->filename && newURL->filename[0]) || (newURL->path && newURL->path[0])) ?
newURL->filename : curURL->filename;
const char *query_string = newURL->query_string;
char *pathfile = (char*)DpsMalloc(dps_strlen(DPS_NULL2EMPTY(path)) + dps_strlen(DPS_NULL2EMPTY(fname)) +
dps_strlen(DPS_NULL2EMPTY(query_string)) + 5);
int cascade;
DPS_MATCH *Alias;
char *alias = NULL;
size_t aliassize, nparts = 10;
DPS_MATCH_PART Parts[10];
if (newURL->hostinfo == NULL) newURL->charset_id = curURL->charset_id;
if (pathfile == NULL) return;
/* sprintf(pathfile, "/%s%s%s", DPS_NULL2EMPTY(path), DPS_NULL2EMPTY(fname), DPS_NULL2EMPTY(query_string));*/
pathfile[0] = '/';
dps_strcpy(pathfile + 1, DPS_NULL2EMPTY(path)); dps_strcat(pathfile, DPS_NULL2EMPTY(fname)); dps_strcat(pathfile, DPS_NULL2EMPTY(query_string));
DpsURLNormalizePath(pathfile);
if (!strcasecmp(DPS_NULL2EMPTY(schema), "mailto")
|| !strcasecmp(DPS_NULL2EMPTY(schema), "javascript")
|| !strcasecmp(DPS_NULL2EMPTY(schema), "feed")
) {
*str = (char*)DpsMalloc(dps_strlen(DPS_NULL2EMPTY(schema)) + dps_strlen(DPS_NULL2EMPTY(newURL->specific)) + 4);
if (*str == NULL) return;
/* sprintf(*str, "%s:%s", DPS_NULL2EMPTY(schema), DPS_NULL2EMPTY(newURL->specific));*/
dps_strcpy(*str, DPS_NULL2EMPTY(schema)); dps_strcat(*str, ":"); dps_strcat(*str, DPS_NULL2EMPTY(newURL->specific));
} else if(/*!strcasecmp(DPS_NULL2EMPTY(schema), "file") ||*/ !strcasecmp(DPS_NULL2EMPTY(schema), "htdb")) {
*str = (char*)DpsMalloc(dps_strlen(DPS_NULL2EMPTY(schema)) + dps_strlen(pathfile) + 4);
if (*str == NULL) return;
/* sprintf(*str, "%s:%s", DPS_NULL2EMPTY(schema), pathfile);*/
dps_strcpy(*str, DPS_NULL2EMPTY(schema)); dps_strcat(*str, ":"); dps_strcat(*str, pathfile);
}else{
*str = (char*)DpsMalloc(dps_strlen(DPS_NULL2EMPTY(schema)) + dps_strlen(pathfile) + dps_strlen(DPS_NULL2EMPTY(hostname)) + dps_strlen(DPS_NULL2EMPTY(auth)) + 8);
if (*str == NULL) return;
/* sprintf(*str, "%s://%s%s", DPS_NULL2EMPTY(schema), DPS_NULL2EMPTY(hostinfo), pathfile);*/
dps_strcpy(*str, DPS_NULL2EMPTY(schema)); dps_strcat(*str, "://");
if (auth) {
dps_strcat(*str, auth); dps_strcat(*str,"@");
}
dps_strcat(*str, DPS_NULL2EMPTY(hostname)); dps_strcat(*str, pathfile);
}
if(!strncmp(*str, "ftp://", 6) && (strstr(*str, ";type=")))
*(strstr(*str, ";type")) = '\0';
DPS_FREE(pathfile);
if (ReverseAliasFlag) {
const char *alias_prog = DpsVarListFindStr(&Indexer->Vars, "ReverseAliasProg", NULL);
if (alias_prog) {
int result;
aliassize = 256 + 2 * dps_strlen(*str);
alias = (char*)DpsRealloc(alias, aliassize);
if (alias == NULL) {
DpsLog(Indexer, DPS_LOG_ERROR, "No memory (%d bytes). %s line %d", aliassize, __FILE__, __LINE__);
goto ret;
}
alias[0] = '\0';
result = DpsAliasProg(Indexer, alias_prog, *str, alias, aliassize - 1);
DpsLog(Indexer, DPS_LOG_EXTRA, "ReverseAliasProg result: '%s'", alias);
if(result != DPS_OK) goto ret;
DPS_FREE(*str);
*str = (char*)DpsStrdup(alias);
}
for(cascade = 0; ((Alias=DpsMatchListFind(&Indexer->Conf->ReverseAliases,*str,nparts,Parts))) && (cascade < 1024); cascade++) {
aliassize = dps_strlen(Alias->arg) + dps_strlen(Alias->pattern) + dps_strlen(*str) + 128;
alias = (char*)DpsRealloc(alias, aliassize);
if (alias == NULL) {
DpsLog(Indexer, DPS_LOG_ERROR, "No memory (%d bytes). %s line %d", aliassize, __FILE__, __LINE__);
goto ret;
}
DpsMatchApply(alias,aliassize,*str,Alias->arg,Alias,nparts,Parts);
if(alias[0]){
DpsLog(Indexer,DPS_LOG_DEBUG,"ReverseAlias%d: pattern:%s, arg:%s -> '%s'", cascade, Alias->pattern, Alias->arg, alias);
DPS_FREE(*str);
*str = (char*)DpsStrdup(alias);
} else break;
if (Alias->last) break;
}
}
ret:
DPS_FREE(alias);
}
int __DPSCALL DpsSearchdCatAction(DPS_AGENT *A, DPS_CATEGORY *C, int cmd, void *db) {
DPS_DB *searchd = db;
DPS_SEARCHD_PACKET_HEADER hdr;
char *buf;
ssize_t nsent, nrecv;
int done = 0;
int rc=DPS_OK;
char *msg = NULL;
char *dinfo = NULL;
TRACE_IN(A, "DpsSearchdCatAction");
hdr.cmd = DPS_SEARCHD_CMD_CATINFO;
hdr.len = sizeof(int) + dps_strlen(C->addr) + 1;
if ((buf = (char*)DpsMalloc(hdr.len + 1)) == NULL) {
DpsLog(A, DPS_LOG_ERROR, "Out of memory");
TRACE_OUT(A);
return DPS_ERROR;
}
*((int*)buf) = cmd;
dps_strcpy(buf + sizeof(int), C->addr);
nsent = DpsSearchdSendPacket(searchd->searchd, &hdr, buf);
DPS_FREE(buf);
while(!done) {
char * tok, * lt;
nrecv = DpsRecvall(searchd->searchd, &hdr, sizeof(hdr), 360);
if(nrecv != sizeof(hdr)){
DpsLog(A, DPS_LOG_ERROR, "Received incomplete header from searchd (%d bytes)", (int)nrecv);
TRACE_OUT(A);
return(DPS_ERROR);
}else{
#ifdef DEBUG_SDP
DpsLog(A, DPS_LOG_ERROR, "Received header cmd=%d len=%d\n", hdr.cmd, hdr.len);
#endif
}
switch(hdr.cmd){
case DPS_SEARCHD_CMD_ERROR:
msg = (char*)DpsMalloc(hdr.len + 1);
if (msg == NULL) {
done=1;
break;
}
nrecv = DpsRecvall(searchd->searchd, msg, hdr.len, 360);
msg[(nrecv >= 0) ? nrecv : 0 ] = '\0';
sprintf(A->Conf->errstr, "Searchd error: '%s'", msg);
rc=DPS_ERROR;
DPS_FREE(msg);
done=1;
break;
case DPS_SEARCHD_CMD_MESSAGE:
msg=(char*)DpsMalloc(hdr.len+1);
if (msg == NULL) {
done=1;
break;
}
nrecv = DpsRecvall(searchd->searchd, msg, hdr.len, 360);
msg[(nrecv >= 0) ? nrecv : 0] = '\0';
#ifdef DEBUG_SDP
DpsLog(A, DPS_LOG_ERROR, "Message from searchd: '%s'\n",msg);
#endif
DPS_FREE(msg);
break;
case DPS_SEARCHD_CMD_CATINFO:
dinfo=(char*)DpsMalloc(hdr.len+1);
if (dinfo == NULL) {
done=1;
break;
}
nrecv = DpsRecvall(searchd->searchd, dinfo, hdr.len, 360);
dinfo[(nrecv >= 0) ? nrecv : 0] = '\0';
#ifdef DEBUG_SDP
DpsLog(A, DPS_LOG_ERROR, "Received CATINFO size=%d buf=%s\n",hdr.len,dinfo);
#endif
C->ncategories = 0;
tok = dps_strtok_r(dinfo, "\r\n", <, NULL);
while(tok){
DpsCatFromTextBuf(C, tok);
tok = dps_strtok_r(NULL, "\r\n", <, NULL);
}
DPS_FREE(dinfo);
done=1;
break;
default:
sprintf(A->Conf->errstr, "Unknown searchd response: cmd=%d len=%d", hdr.cmd, hdr.len);
rc=DPS_ERROR;
done = 1;
break;
}
}
TRACE_OUT(A);
return rc;
}
