-const char parsers_rcs[] = "$Id: parsers.c,v 1.293 2014/10/18 11:24:34 fabiankeil Exp $";
+const char parsers_rcs[] = "$Id: parsers.c,v 1.307 2016/01/17 14:31:47 fabiankeil Exp $";
/*********************************************************************
*
* File : $Source: /cvsroot/ijbswa/current/parsers.c,v $
*
* Purpose : Declares functions to parse/crunch headers and pages.
*
- * Copyright : Written by and Copyright (C) 2001-2014 the
+ * Copyright : Written by and Copyright (C) 2001-2016 the
* Privoxy team. http://www.privoxy.org/
*
* Based on the Internet Junkbuster originally written
static jb_err scan_headers(struct client_state *csp);
static jb_err header_tagger(struct client_state *csp, char *header);
static jb_err parse_header_time(const char *header_time, time_t *result);
+static jb_err parse_time_header(const char *header, time_t *result);
static jb_err crumble (struct client_state *csp, char **header);
static jb_err filter_header (struct client_state *csp, char **header);
void clear_iob(struct iob *iob)
{
free(iob->buf);
- memset(iob, '\0', sizeof(*iob));;
+ memset(iob, '\0', sizeof(*iob));
}
* This is to protect the parsing of gzipped data,
* but it should(?) be valid for deflated data also.
*/
- log_error(LOG_LEVEL_ERROR, "Buffer too small decompressing iob");
+ log_error(LOG_LEVEL_ERROR,
+ "Insufficient data to start decompression. Bytes in buffer: %d",
+ csp->iob->eod - csp->iob->cur);
return JB_ERR_COMPRESS;
}
csp->flags |= CSP_FLAG_SERVER_KEEP_ALIVE_TIMEOUT_SET;
}
+ freez(*header);
+
return JB_ERR_OK;
}
*/
if ((csp->content_type & CT_TEXT) || (csp->action->flags & ACTION_FORCE_TEXT_MODE))
{
+ jb_err err;
freez(*header);
*header = strdup_or_die("Content-Type: ");
- string_append(header, csp->action->string[ACTION_STRING_CONTENT_TYPE]);
- if (header == NULL)
+ err = string_append(header, csp->action->string[ACTION_STRING_CONTENT_TYPE]);
+ if (JB_ERR_OK != err)
{
log_error(LOG_LEVEL_HEADER, "Insufficient memory to replace Content-Type!");
return JB_ERR_MEMORY;
/*
* Log a warning if the user expects the content to be filtered.
*/
- if ((csp->rlist != NULL) &&
- (!list_is_empty(csp->action->multi[ACTION_MULTI_FILTER])))
+ if (content_filters_enabled(csp->action))
{
log_error(LOG_LEVEL_INFO,
"Compressed content detected, content filtering disabled. "
}
else if (0 == strcmpic(newval, "randomize"))
{
- const char *header_time = *header + sizeof("Last-Modified:");
-
log_error(LOG_LEVEL_HEADER, "Randomizing: %s", *header);
- if (JB_ERR_OK != parse_header_time(header_time, &last_modified))
+ if (JB_ERR_OK != parse_time_header(*header, &last_modified))
{
- log_error(LOG_LEVEL_HEADER, "Couldn't parse: %s in %s (crunching!)", header_time, *header);
+ log_error(LOG_LEVEL_HEADER,
+ "Couldn't parse time in %s (crunching!)", *header);
freez(*header);
}
else
{
char *p, *q;
+ if (strlen(*header) < 7)
+ {
+ log_error(LOG_LEVEL_HEADER, "Removing empty Host header");
+ freez(*header);
+ return JB_ERR_OK;
+ }
+
if (!csp->http->hostport || (*csp->http->hostport == '*') ||
*csp->http->hostport == ' ' || *csp->http->hostport == '\0')
{
}
else /* add random value */
{
- const char *header_time = *header + sizeof("If-Modified-Since:");
-
- if (JB_ERR_OK != parse_header_time(header_time, &tm))
+ if (JB_ERR_OK != parse_time_header(*header, &tm))
{
- log_error(LOG_LEVEL_HEADER, "Couldn't parse: %s in %s (crunching!)", header_time, *header);
+ log_error(LOG_LEVEL_HEADER,
+ "Couldn't parse time in %s (crunching!)", *header);
freez(*header);
}
else
if (!csp->http->hostport || !*(csp->http->hostport))
{
- /* XXX: When does this happen and why is it OK? */
- log_error(LOG_LEVEL_INFO, "Weirdness in client_host_adder detected and ignored.");
- return JB_ERR_OK;
+ log_error(LOG_LEVEL_ERROR, "Destination host unknown.");
+ return JB_ERR_PARSE;
}
/*
reason_phrase="";
}
- if (3 != sscanf(*header, "HTTP/%u.%u %u", &major_version,
+ if (3 != sscanf(*header, "HTTP/%u.%u %d", &major_version,
&minor_version, &(csp->http->status)))
{
log_error(LOG_LEVEL_ERROR,
length = sizeof("HTTP/1.1 200 ") + strlen(reason_phrase) + 1;
new_response_line = malloc_or_die(length);
- snprintf(new_response_line, length, "HTTP/%u.%u %u %s",
+ snprintf(new_response_line, length, "HTTP/%u.%u %d %s",
major_version, minor_version, csp->http->status, reason_phrase);
if (0 != strcmp(*header, new_response_line))
}
+/*********************************************************************
+ *
+ * Function : parse_time_header
+ *
+ * Description : Parses the time in an HTTP time header to get
+ * the numerical respresentation.
+ *
+ * Parameters :
+ * 1 : header = HTTP header with a time value
+ * 2 : result = storage for header_time in seconds
+ *
+ * Returns : JB_ERR_OK if the time format was recognized, or
+ * JB_ERR_PARSE otherwise.
+ *
+ *********************************************************************/
+static jb_err parse_time_header(const char *header, time_t *result)
+{
+ const char *header_time;
+
+ header_time = strchr(header, ':');
+
+ /*
+ * Currently this can't happen as all callers are called
+ * through sed() which requires a header name followed by
+ * a colon.
+ */
+ assert(header_time != NULL);
+
+ header_time++;
+ if (*header_time == ' ')
+ {
+ header_time++;
+ }
+
+ return parse_header_time(header_time, result);
+
+}
+
/*********************************************************************
*
char *p;
char *host;
+ assert(!http->ssl);
+
host = get_header_value(headers, "Host:");
if (NULL == host)
}
else
{
- http->port = http->ssl ? 443 : 80;
+ http->port = 80;
}
/* Rebuild request URL */
freez(http->url);
- http->url = strdup(http->ssl ? "https://" : "http://");
+ http->url = strdup("http://");
string_append(&http->url, http->hostport);
string_append(&http->url, http->path);
if (http->url == NULL)
return JB_ERR_MEMORY;
}
- log_error(LOG_LEVEL_HEADER, "Destination extracted from \"Host:\" header. New request URL: %s",
+ log_error(LOG_LEVEL_HEADER,
+ "Destination extracted from \"Host\" header. New request URL: %s",
http->url);
+ /*
+ * Regenerate request line in "proxy format"
+ * to make rewrites more convenient.
+ */
+ assert(http->cmd != NULL);
+ freez(http->cmd);
+ http->cmd = strdup_or_die(http->gpc);
+ string_append(&http->cmd, " ");
+ string_append(&http->cmd, http->url);
+ string_append(&http->cmd, " ");
+ string_append(&http->cmd, http->ver);
+ if (http->cmd == NULL)
+ {
+ return JB_ERR_MEMORY;
+ }
+
return JB_ERR_OK;
}