void Indexer::index_directory(const unicode &dir_path) {
/*
* Recursively indexes the files in a directory
*/
for(auto file: kfs::path::list_dir(dir_path.encode())) {
auto full_path = kfs::path::join(dir_path.encode(), file);
if(kfs::path::is_dir(full_path)) {
index_directory(full_path);
} else {
index_file(full_path);
}
}
}
int get_cachedir_file(request * req, struct stat *statbuf)
{
char pathname_with_index[MAX_PATH_LENGTH];
int data_fd;
time_t real_dir_mtime;
real_dir_mtime = statbuf->st_mtime;
sprintf(pathname_with_index, "%s/dir.%d.%ld",
cachedir, (int) statbuf->st_dev, statbuf->st_ino);
data_fd = open(pathname_with_index, O_RDONLY);
if (data_fd != -1) { /* index cache */
fstat(data_fd, statbuf);
if (statbuf->st_mtime > real_dir_mtime) {
statbuf->st_mtime = real_dir_mtime; /* lie */
strcpy(req->request_uri, directory_index); /* for mimetype */
return data_fd;
}
close(data_fd);
unlink(pathname_with_index); /* cache is stale, delete it */
}
if (index_directory(req, pathname_with_index) == -1)
return -1;
data_fd = open(pathname_with_index, O_RDONLY); /* Last chance */
if (data_fd != -1) {
strcpy(req->request_uri, directory_index); /* for mimetype */
fstat(data_fd, statbuf);
statbuf->st_mtime = real_dir_mtime; /* lie */
return data_fd;
}
boa_perror(req, "re-opening dircache");
return -1; /* Nothing worked. */
}
int get_cachedir_file(request * req, struct stat *statbuf)
{
char pathname_with_index[MAX_PATH_LENGTH];
int data_fd;
time_t real_dir_mtime;
real_dir_mtime = statbuf->st_mtime;
sprintf(pathname_with_index, "%s/dir.%d.%ld",
cachedir, (int) statbuf->st_dev, statbuf->st_ino);
data_fd = open(pathname_with_index, O_RDONLY);
if (data_fd != -1) {
fstat(data_fd, statbuf);
if (statbuf->st_mtime > real_dir_mtime) {
statbuf->st_mtime = real_dir_mtime;
strcpy(req->request_uri, directory_index);
return data_fd;
}
close(data_fd);
unlink(pathname_with_index);
}
if (index_directory(req, pathname_with_index) == -1)
return -1;
data_fd = open(pathname_with_index, O_RDONLY);
if (data_fd != -1) {
strcpy(req->request_uri, directory_index);
fstat(data_fd, statbuf);
statbuf->st_mtime = real_dir_mtime;
return data_fd;
}
boa_perror(req, "re-opening dircache");
return -1;
}
int process_get(request * req)
{
off_t bytes_written;
volatile off_t bytes_to_write;
if (req->method == M_HEAD) {
return complete_response(req);
}
bytes_to_write = (req->ranges->stop - req->ranges->start) + 1;
if (bytes_to_write > system_bufsize)
bytes_to_write = system_bufsize;
if (setjmp(env) == 0) {
handle_sigbus = 1;
bytes_written = write(req->fd, req->data_mem + req->ranges->start,
bytes_to_write);
handle_sigbus = 0;
/* OK, SIGBUS **after** this point is very bad! */
} else {
/* sigbus! */
req->status = DEAD;
log_error_doc(req);
fprintf(stderr, "%sGot SIGBUS in write(2)!\n",
get_commonlog_time());
/* sending an error here is inappropriate
* if we are here, the file is mmapped, and thus,
* a content-length has been sent. If we send fewer bytes
* the client knows there has been a problem.
* We run the risk of accidentally sending the right number
* of bytes (or a few too many) and the client
* won't be the wiser.
*/
return 0;
}
if (bytes_written < 0) {
if (errno == EWOULDBLOCK || errno == EAGAIN)
return -1;
/* request blocked at the pipe level, but keep going */
else {
#ifdef QUIET_DISCONNECT
if (errno != EPIPE) {
#else
if (1) {
#endif
log_error_doc(req);
/* Can generate lots of log entries, */
perror("write");
/* OK to disable if your logs get too big */
}
req->status = DEAD;
return 0;
}
}
req->bytes_written += bytes_written;
req->ranges->start += bytes_written;
if ((req->ranges->stop + 1 - req->ranges->start) == 0) {
return complete_response(req);
}
return 1; /* more to do */
}
/*
* Name: get_dir
* Description: Called from process_get if the request is a directory.
* statbuf must describe directory on input, since we may need its
* device, inode, and mtime.
* statbuf is updated, since we may need to check mtimes of a cache.
* returns:
* -1 error
* 0 cgi (either gunzip or auto-generated)
* >0 file descriptor of file
*/
int get_dir(request * req, struct stat *statbuf)
{
char pathname_with_index[MAX_PATH_LENGTH];
int data_fd;
if (directory_index) { /* look for index.html first?? */
unsigned int l1, l2;
l1 = strlen(req->pathname);
l2 = strlen(directory_index);
#ifdef GUNZIP
if (l1 + l2 + 3 + 1 > sizeof(pathname_with_index)) { /* for .gz */
#else
if (l1 + l2 + 1 > sizeof(pathname_with_index)) {
#endif
errno = ENOMEM;
boa_perror(req, "pathname_with_index not large enough for pathname + index");
return -1;
}
memcpy(pathname_with_index, req->pathname, l1); /* doesn't copy NUL */
memcpy(pathname_with_index + l1, directory_index, l2 + 1); /* does */
data_fd = open(pathname_with_index, O_RDONLY|O_LARGEFILE);
if (data_fd != -1) { /* user's index file */
/* We have to assume that directory_index will fit, because
* if it doesn't, well, that's a huge configuration problem.
* this is only the 'index.html' pathname for mime type
*/
memcpy(req->request_uri, directory_index, l2 + 1); /* for mimetype */
fstat(data_fd, statbuf);
return data_fd;
}
if (errno == EACCES) {
send_r_forbidden(req);
return -1;
} else if (errno != ENOENT) {
/* if there is an error *other* than EACCES or ENOENT */
send_r_not_found(req);
return -1;
}
#ifdef GUNZIP
/* if we are here, trying index.html didn't work
* try index.html.gz
*/
strcat(pathname_with_index, ".gz");
data_fd = open(pathname_with_index, O_RDONLY|O_LARGEFILE);
if (data_fd != -1) { /* user's index file */
close(data_fd);
req->response_status = R_REQUEST_OK;
SQUASH_KA(req);
if (req->pathname)
free(req->pathname);
req->pathname = strdup(pathname_with_index);
if (!req->pathname) {
boa_perror(req, "strdup of pathname_with_index for .gz files " __FILE__ ":" STR(__LINE__));
return 0;
}
if (req->http_version != HTTP09) {
req_write(req, http_ver_string(req->http_version));
req_write(req, " 200 OK-GUNZIP" CRLF);
print_http_headers(req);
print_last_modified(req);
req_write(req, "Content-Type: ");
req_write(req, get_mime_type(directory_index));
req_write(req, CRLF CRLF);
req_flush(req);
}
if (req->method == M_HEAD)
return 0;
return init_cgi(req);
}
#endif
}
/* only here if index.html, index.html.gz don't exist */
if (dirmaker != NULL) { /* don't look for index.html... maybe automake? */
req->response_status = R_REQUEST_OK;
SQUASH_KA(req);
/* the indexer should take care of all headers */
if (req->http_version != HTTP09) {
req_write(req, http_ver_string(req->http_version));
req_write(req, " 200 OK" CRLF);
print_http_headers(req);
print_last_modified(req);
req_write(req, "Content-Type: text/html" CRLF CRLF);
req_flush(req);
}
if (req->method == M_HEAD)
return 0;
return init_cgi(req);
/* in this case, 0 means success */
} else if (cachedir) {
return get_cachedir_file(req, statbuf);
} else { /* neither index.html nor autogenerate are allowed */
send_r_forbidden(req);
return -1; /* nothing worked */
}
}
static int get_cachedir_file(request * req, struct stat *statbuf)
{
char pathname_with_index[MAX_PATH_LENGTH];
int data_fd;
time_t real_dir_mtime;
real_dir_mtime = statbuf->st_mtime;
/* the sizeof() doesn't need a -1 because snprintf will
* include the NUL when calculating if the size is enough
*/
snprintf(pathname_with_index, sizeof(pathname_with_index),
"%s/dir.%d." PRINTF_OFF_T_ARG, cachedir,
(int) statbuf->st_dev, statbuf->st_ino);
data_fd = open(pathname_with_index, O_RDONLY|O_LARGEFILE);
if (data_fd != -1) { /* index cache */
fstat(data_fd, statbuf);
if (statbuf->st_mtime > real_dir_mtime) {
statbuf->st_mtime = real_dir_mtime; /* lie */
strcpy(req->request_uri, directory_index); /* for mimetype */
return data_fd;
}
close(data_fd);
unlink(pathname_with_index); /* cache is stale, delete it */
}
if (index_directory(req, pathname_with_index) == -1)
return -1;
data_fd = open(pathname_with_index, O_RDONLY|O_LARGEFILE); /* Last chance */
if (data_fd != -1) {
strcpy(req->request_uri, directory_index); /* for mimetype */
fstat(data_fd, statbuf);
statbuf->st_mtime = real_dir_mtime; /* lie */
return data_fd;
}
boa_perror(req, "re-opening dircache");
return -1; /* Nothing worked. */
}
void send_file(per_request *reqInfo, struct stat *fi, char allow_options)
{
FILE *f;
#ifdef BLACKOUT_CODE
int isblack = FALSE;
#endif /* BLACKOUT_CODE */
if ((reqInfo->method != M_GET) && (reqInfo->method != M_HEAD)) {
sprintf(error_msg,"%s to non-script",methods[reqInfo->method]);
die(reqInfo,SC_NOT_IMPLEMENTED,error_msg);
}
set_content_type(reqInfo,reqInfo->filename);
if((allow_options & OPT_INCLUDES) && (!reqInfo->outh_content_encoding[0])) {
#ifdef XBITHACK
if((fi->st_mode & S_IXUSR) ||
(!strcmp(reqInfo->outh_content_type,INCLUDES_MAGIC_TYPE))) {
#else
if(!strcmp(reqInfo->outh_content_type,INCLUDES_MAGIC_TYPE)) {
#endif /* XBITHACK */
reqInfo->bytes_sent = 0;
send_parsed_file(reqInfo, allow_options & OPT_INCNOEXEC);
log_transaction(reqInfo);
return;
}
}
if (reqInfo->path_info[0]) {
strcat(reqInfo->filename,reqInfo->path_info);
strcat(reqInfo->url,reqInfo->path_info);
sprintf(error_msg,"No file matching URL: %s",reqInfo->url);
log_reason(reqInfo, error_msg, reqInfo->filename);
die(reqInfo,SC_NOT_FOUND,reqInfo->url);
}
if(!(f=FOpen(reqInfo->filename,"r"))) {
if (errno == EACCES) {
log_reason(reqInfo,"(1) file permissions deny server access",
reqInfo->filename);
/* we've already established that it exists */
die(reqInfo,SC_FORBIDDEN,reqInfo->url);
} else {
/* We know an error occured, of an unexpected variety.
* This could be due to no more file descriptors. We have this
* child exit after this stage so that errors of state are
* swept under the carpet.
*/
standalone = 0;
sprintf(error_msg,"File Open error, errno=%d",errno);
log_reason(reqInfo,error_msg,reqInfo->filename);
die(reqInfo,SC_SERVER_ERROR,error_msg);
}
}
reqInfo->bytes_sent = 0;
#ifdef BLACKOUT_CODE
if (!strcmp(reqInfo->outh_content_type,BLACKOUT_MAGIC_TYPE)) {
isblack = TRUE;
strcpy(reqInfo->outh_content_type,"text/html");
}
#endif /* BLACKOUT_CODE */
if(reqInfo->http_version != P_HTTP_0_9) {
/* No length dependent headers since black is parsed */
#ifdef BLACKOUT_CODE
if (isblack == FALSE) {
#endif /* BLACKOUT_CODE */
#ifdef CONTENT_MD5
reqInfo->outh_content_md5 = (unsigned char *)md5digest(f);
#endif /* CONTENT_MD5 */
set_content_length(reqInfo,fi->st_size);
if (set_last_modified(reqInfo,fi->st_mtime)) {
FClose(f);
return;
}
}
if (reqInfo->http_version != P_HTTP_0_9) {
send_http_header(reqInfo);
}
#ifdef BLACKOUT_CODE
}
#endif /* BLACKOUT_CODE */
if(reqInfo->method != M_HEAD) {
#ifdef BLACKOUT_CODE
if (isblack == TRUE)
send_fp_black(reqInfo,f,NULL);
else
#endif /* BLACKOUT_CODE */
send_fp(reqInfo,f,NULL);
}
log_transaction(reqInfo);
FClose(f);
}
void send_dir(per_request *reqInfo,struct stat *finfo, char allow_options) {
char *name_ptr, *end_ptr;
char *ifile, *temp_name;
ifile = newString(HUGE_STRING_LEN,STR_TMP);
temp_name = newString(HUGE_STRING_LEN,STR_TMP);
/* Path Alias (pa) array should now have the trailing slash */
/* if (pa[0] != '/') { */
if ((reqInfo->filename[strlen(reqInfo->filename) - 1] != '/') &&
(reqInfo->path_info[0] != '/')) {
strcpy_dir(ifile,reqInfo->url);
construct_url(temp_name,reqInfo->hostInfo,ifile);
escape_url(temp_name);
die(reqInfo,SC_REDIRECT_PERM,temp_name);
}
/* Don't allow PATH_INFO to directory indexes as a compromise for
error messages for files which don't exist */
if ((reqInfo->path_info[0] != '\0') || (strlen(reqInfo->path_info) > 1)) {
strcat(reqInfo->filename,reqInfo->path_info);
strcat(reqInfo->url,reqInfo->path_info);
sprintf(error_msg,"No file matching URL: %s",reqInfo->url);
log_reason(reqInfo, error_msg, reqInfo->filename);
freeString(temp_name);
freeString(ifile);
die(reqInfo,SC_NOT_FOUND,reqInfo->url);
}
strncpy(temp_name, reqInfo->hostInfo->index_names, HUGE_STRING_LEN-1);
end_ptr = name_ptr = temp_name;
while (*name_ptr) {
while (*name_ptr && isspace (*name_ptr)) ++name_ptr;
end_ptr = name_ptr;
if (strchr(end_ptr, ' ') ) {
end_ptr = strchr(name_ptr, ' ');
*end_ptr = '\0';
end_ptr++;
} else
end_ptr += strlen(end_ptr);
make_full_path(reqInfo->filename,name_ptr,ifile);
if(stat(ifile,finfo) == -1) {
if(! *end_ptr && (allow_options & OPT_INDEXES)) {
if (reqInfo->path_info[0]) {
strcat(reqInfo->filename,reqInfo->path_info);
strcat(reqInfo->url,reqInfo->path_info);
log_reason(reqInfo,"file does not exist",reqInfo->filename);
freeString(ifile);
freeString(temp_name);
die(reqInfo,SC_NOT_FOUND,reqInfo->url);
}
if ((reqInfo->method != M_GET) && (reqInfo->method != M_HEAD)) {
sprintf(error_msg,"%s to non-script",methods[reqInfo->method]);
freeString(ifile);
freeString(temp_name);
die(reqInfo,SC_NOT_IMPLEMENTED,error_msg);
}
index_directory(reqInfo);
freeString(ifile);
freeString(temp_name);
return;
} else if (! *end_ptr) {
log_reason(reqInfo,"(2) file permissions deny server access",
reqInfo->filename);
freeString(ifile);
freeString(temp_name);
die(reqInfo,SC_FORBIDDEN,reqInfo->url);
}
} else {
strcpy(reqInfo->filename,ifile);
probe_content_type(reqInfo,reqInfo->filename);
if(!strcmp(reqInfo->outh_content_type,CGI_MAGIC_TYPE))
send_cgi(reqInfo,finfo,allow_options);
else
send_file(reqInfo,finfo,allow_options);
freeString(ifile);
freeString(temp_name);
return;
}
name_ptr = end_ptr;
}
}
