X-Git-Url: http://www.privoxy.org/gitweb/?p=privoxy.git;a=blobdiff_plain;f=parsers.c;h=8e0ea45706de79a331bd58c60fd3e1968b513525;hp=27ada5f76c44b1263a97f0d42c7b47c0ce50caa2;hb=1d99cd7a1e89f84d3f750afe14aebe4aa9647a09;hpb=320155e7eb1cdd404e8cd8c741d3cca045aa774e diff --git a/parsers.c b/parsers.c index 27ada5f7..8e0ea457 100644 --- a/parsers.c +++ b/parsers.c @@ -1,11 +1,11 @@ -const char parsers_rcs[] = "$Id: parsers.c,v 1.296 2014/11/03 14:41:12 fabiankeil Exp $"; +const char parsers_rcs[] = "$Id: parsers.c,v 1.308 2016/02/26 12:31:12 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 @@ -96,6 +96,7 @@ static char *get_header_line(struct iob *iob); 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); @@ -386,7 +387,7 @@ jb_err add_to_iob(struct iob *iob, const size_t buffer_limit, char *src, long n) void clear_iob(struct iob *iob) { free(iob->buf); - memset(iob, '\0', sizeof(*iob));; + memset(iob, '\0', sizeof(*iob)); } @@ -435,7 +436,9 @@ jb_err decompress_iob(struct client_state *csp) * 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; } @@ -2369,8 +2372,7 @@ static jb_err server_content_encoding(struct client_state *csp, char **header) /* * 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. " @@ -2668,13 +2670,12 @@ static jb_err server_last_modified(struct client_state *csp, char **header) } 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 @@ -3285,6 +3286,13 @@ static jb_err client_host(struct client_state *csp, char **header) { 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') { @@ -3371,11 +3379,10 @@ static jb_err client_if_modified_since(struct client_state *csp, char **header) } 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 @@ -3599,9 +3606,8 @@ static jb_err client_host_adder(struct client_state *csp) 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; } /* @@ -4307,7 +4313,13 @@ static jb_err parse_header_time(const char *header_time, time_t *result) time_t result2; tm = gmtime(result); - strftime(recreated_date, sizeof(recreated_date), time_formats[i], tm); + if (!strftime(recreated_date, sizeof(recreated_date), + time_formats[i], tm)) + { + log_error(LOG_LEVEL_ERROR, "Failed to recreate date '%s' with '%s'.", + header_time, time_formats[i]); + continue; + } memset(&gmt, 0, sizeof(gmt)); if (NULL == strptime(recreated_date, time_formats[i], &gmt)) { @@ -4335,6 +4347,44 @@ static jb_err parse_header_time(const char *header_time, time_t *result) } +/********************************************************************* + * + * 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); + +} + /********************************************************************* * @@ -4363,6 +4413,8 @@ jb_err get_destination_from_headers(const struct list *headers, struct http_requ char *p; char *host; + assert(!http->ssl); + host = get_header_value(headers, "Host:"); if (NULL == host) @@ -4388,12 +4440,12 @@ jb_err get_destination_from_headers(const struct list *headers, struct http_requ } 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) @@ -4401,9 +4453,26 @@ jb_err get_destination_from_headers(const struct list *headers, struct http_requ 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; }