#ifndef PROJECT_H_INCLUDED
#define PROJECT_H_INCLUDED
/** Version string. */
-#define PROJECT_H_VERSION "$Id: project.h,v 1.90 2007/02/07 10:36:16 fabiankeil Exp $"
+#define PROJECT_H_VERSION "$Id: project.h,v 1.110 2008/03/29 12:13:46 fabiankeil Exp $"
/*********************************************************************
*
* File : $Source: /cvsroot/ijbswa/current/project.h,v $
* project. Does not define any variables or functions
* (though it does declare some macros).
*
- * Copyright : Written by and Copyright (C) 2001 - 2007 the SourceForge
+ * Copyright : Written by and Copyright (C) 2001-2008 the SourceForge
* Privoxy team. http://www.privoxy.org/
*
* Based on the Internet Junkbuster originally written
*
* Revisions :
* $Log: project.h,v $
+ * Revision 1.110 2008/03/29 12:13:46 fabiankeil
+ * Remove send-wafer and send-vanilla-wafer actions.
+ *
+ * Revision 1.109 2008/03/28 15:13:41 fabiankeil
+ * Remove inspect-jpegs action.
+ *
+ * Revision 1.108 2008/03/27 18:27:36 fabiankeil
+ * Remove kill-popups action.
+ *
+ * Revision 1.107 2008/03/26 18:07:08 fabiankeil
+ * Add hostname directive. Closes PR#1918189.
+ *
+ * Revision 1.106 2008/03/24 11:21:03 fabiankeil
+ * Share the action settings for multiple patterns in the same
+ * section so we waste less memory for gigantic block lists
+ * (and load them slightly faster). Reported by Franz Schwartau.
+ *
+ * Revision 1.105 2008/03/21 11:16:27 fabiankeil
+ * Garbage-collect csp->my_ip_addr_str and csp->my_hostname.
+ *
+ * Revision 1.104 2008/03/04 18:30:40 fabiankeil
+ * Remove the treat-forbidden-connects-like-blocks action. We now
+ * use the "blocked" page for forbidden CONNECT requests by default.
+ *
+ * Revision 1.103 2008/03/01 14:00:45 fabiankeil
+ * Let the block action take the reason for the block
+ * as argument and show it on the "blocked" page.
+ *
+ * Revision 1.102 2008/02/03 13:46:14 fabiankeil
+ * Add SOCKS5 support. Patch #1862863 by Eric M. Hopper with minor changes.
+ *
+ * Revision 1.101 2007/12/07 18:29:23 fabiankeil
+ * Remove now-obsolete csp member x_forwarded.
+ *
+ * Revision 1.100 2007/09/02 13:42:11 fabiankeil
+ * - Allow port lists in url patterns.
+ * - Ditch unused url_spec member pathlen.
+ *
+ * Revision 1.99 2007/07/21 11:51:36 fabiankeil
+ * As Hal noticed, checking dispatch_cgi() as the last cruncher
+ * looks like a bug if CGI requests are blocked unintentionally,
+ * so don't do it unless the user enabled the new config option
+ * "allow-cgi-request-crunching".
+ *
+ * Revision 1.98 2007/07/14 07:31:26 fabiankeil
+ * Add new csp->content_type flag (CT_DECLARED).
+ *
+ * Revision 1.97 2007/05/27 12:38:08 fabiankeil
+ * - Remove some left-overs from the switch to dedicated header filters.
+ * - Adjust "X-Filter: No" to disable dedicated header filters.
+ * - Prepare for forward-override{}
+ *
+ * Revision 1.96 2007/05/14 10:41:15 fabiankeil
+ * Ditch the csp member cookie_list[] which isn't used anymore.
+ *
+ * Revision 1.95 2007/04/30 15:02:19 fabiankeil
+ * Introduce dynamic pcrs jobs that can resolve variables.
+ *
+ * Revision 1.94 2007/04/15 16:39:21 fabiankeil
+ * Introduce tags as alternative way to specify which
+ * actions apply to a request. At the moment tags can be
+ * created based on client and server headers.
+ *
+ * Revision 1.93 2007/03/20 15:16:34 fabiankeil
+ * Use dedicated header filter actions instead of abusing "filter".
+ * Replace "filter-client-headers" and "filter-client-headers"
+ * with "server-header-filter" and "client-header-filter".
+ *
+ * Revision 1.92 2007/03/17 15:20:05 fabiankeil
+ * New config option: enforce-blocks.
+ *
+ * Revision 1.91 2007/03/05 13:28:03 fabiankeil
+ * Add some CSP_FLAGs for the header parsers.
+ *
* Revision 1.90 2007/02/07 10:36:16 fabiankeil
* Add new http_response member to save
* the reason why the response was generated.
};
/**
- * A URL pattern.
+ * A URL or a tag pattern.
*/
struct url_spec
{
Used for debugging or display only. */
char *spec;
+#ifdef FEATURE_PCRE_HOST_PATTERNS
+ regex_t *host_regex;/**< Regex for host matching */
+#else
char *dbuffer; /**< Buffer with '\0'-delimited domain name, or NULL to match all hosts. */
char **dvec; /**< List of pointers to the strings in dbuffer. */
int dcount; /**< How many parts to this domain? (length of dvec) */
int unanchored; /**< Bitmap - flags are ANCHOR_LEFT and ANCHOR_RIGHT. */
+#endif /* defined FEATURE_PCRE_HOST_PATTERNS */
- int port; /**< The port number, or 0 to match all ports. */
+ char *port_list; /**< List of acceptable ports, or NULL to match all ports */
- char *path; /**< The source for the regex. */
- size_t pathlen; /**< ==strlen(path). Needed for prefix matching. FIXME: Now obsolete? */
+ char *path; /**< The source for the regex. */
regex_t *preg; /**< Regex for matching path part */
+ regex_t *tag_regex; /**< Regex for matching tags */
};
/**
* If you declare a static url_spec, this is the value to initialize it to zero.
*/
-#define URL_SPEC_INITIALIZER { NULL, NULL, NULL, 0, 0, 0, NULL, 0, NULL }
+#ifndef FEATURE_PCRE_HOST_PATTERNS
+#define URL_SPEC_INITIALIZER { NULL, NULL, NULL, 0, 0, NULL, NULL, NULL, NULL }
+#else
+#define URL_SPEC_INITIALIZER { NULL, NULL, NULL, NULL, NULL, NULL }
+#endif /* def FEATURE_PCRE_HOST_PATTERNS */
/**
* Constant for host part matching in URLs. If set, indicates that the start of
#define CT_TEXT 0x0001U /**< Suitable for pcrs filtering. */
#define CT_GIF 0x0002U /**< Suitable for GIF filtering. */
#define CT_TABOO 0x0004U /**< DO NOT filter, irrespective of other flags. */
-#define CT_JPEG 0x0008U /**< Suitable for JPEG filtering. */
/* Although these are not, strictly speaking, content types
* (they are content encodings), it is simple to handle them
#define CT_GZIP 0x0010U /**< gzip-compressed data. */
#define CT_DEFLATE 0x0020U /**< zlib-compressed data. */
+/**
+ * Flag to signal that the server declared the content type,
+ * so we can differentiate between unknown and undeclared
+ * content types.
+ */
+#define CT_DECLARED 0x0040U
+
/**
* The mask which includes all actions.
*/
#define ACTION_NO_COOKIE_READ 0x00001000UL
/** Action bitmap: Block setting cookies. */
#define ACTION_NO_COOKIE_SET 0x00002000UL
-/** Action bitmap: Filter out popups. */
-#define ACTION_NO_POPUPS 0x00004000UL
-/** Action bitmap: Send a vanilla wafer. */
-#define ACTION_VANILLA_WAFER 0x00008000UL
+/** Action bitmap: Override the forward settings in the config file */
+#define ACTION_FORWARD_OVERRIDE 0x00004000UL
+/** Action bitmap: Block as empty document */
+#define ACTION_HANDLE_AS_EMPTY_DOCUMENT 0x00008000UL
/** Action bitmap: Limit CONNECT requests to safe ports. */
#define ACTION_LIMIT_CONNECT 0x00010000UL
-/** Action bitmap: Inspect if it's a JPEG. */
-#define ACTION_JPEG_INSPECT 0x00020000UL
+/** Action bitmap: Redirect request. */
+#define ACTION_REDIRECT 0x00020000UL
/** Action bitmap: Crunch or modify "if-modified-since" header. */
#define ACTION_HIDE_IF_MODIFIED_SINCE 0x00040000UL
/** Action bitmap: Overwrite Content-Type header. */
#define ACTION_OVERWRITE_LAST_MODIFIED 0x02000000UL
/** Action bitmap: Replace or block Accept-Language header */
#define ACTION_HIDE_ACCEPT_LANGUAGE 0x04000000UL
-/** Action bitmap: Block as empty document */
-#define ACTION_HANDLE_AS_EMPTY_DOCUMENT 0x08000000UL
-/** Action bitmap: Redirect request. */
-#define ACTION_REDIRECT 0x10000000UL
-/** Action bitmap: Answer blocked Connects verbosely */
-#define ACTION_TREAT_FORBIDDEN_CONNECTS_LIKE_BLOCKS 0x20000000UL
-/** Action bitmap: Filter server headers with pcre */
-#define ACTION_FILTER_SERVER_HEADERS 0x40000000UL
-/** Action bitmap: Filter client headers with pcre */
-#define ACTION_FILTER_CLIENT_HEADERS 0x80000000UL
/** Action string index: How to deanimate GIFs */
#define ACTION_STRING_REDIRECT 13
/** Action string index: Decode before redirect? */
#define ACTION_STRING_FAST_REDIRECTS 14
+/** Action string index: Overriding forward rule. */
+#define ACTION_STRING_FORWARD_OVERRIDE 15
+/** Action string index: Reason for the block. */
+#define ACTION_STRING_BLOCK 16
/** Number of string actions. */
-#define ACTION_STRING_COUNT 15
+#define ACTION_STRING_COUNT 17
-/*To make the ugly hack in sed easier to understand*/
+/* To make the ugly hack in sed easier to understand */
#define CHECK_EVERY_HEADER_REMAINING 0
/** Index into current_action_spec::multi[] for headers to add. */
-#define ACTION_MULTI_ADD_HEADER 0
-/** Index into current_action_spec::multi[] for headers to add. */
-#define ACTION_MULTI_WAFER 1
-/** Index into current_action_spec::multi[] for filters to apply. */
-#define ACTION_MULTI_FILTER 2
+#define ACTION_MULTI_ADD_HEADER 0
+/** Index into current_action_spec::multi[] for content filters to apply. */
+#define ACTION_MULTI_FILTER 1
+/** Index into current_action_spec::multi[] for server-header filters to apply. */
+#define ACTION_MULTI_SERVER_HEADER_FILTER 2
+/** Index into current_action_spec::multi[] for client-header filters to apply. */
+#define ACTION_MULTI_CLIENT_HEADER_FILTER 3
+/** Index into current_action_spec::multi[] for client-header tags to apply. */
+#define ACTION_MULTI_CLIENT_HEADER_TAGGER 4
+/** Index into current_action_spec::multi[] for server-header tags to apply. */
+#define ACTION_MULTI_SERVER_HEADER_TAGGER 5
/** Number of multi-string actions. */
-#define ACTION_MULTI_COUNT 3
+#define ACTION_MULTI_COUNT 6
/**
/**
- * This structure is used to store the actions list.
+ * This structure is used to store action files.
*
- * It contains a URL pattern, and the chages to the actions.
- * It is a linked list.
+ * It contains an URL or tag pattern, and the changes to
+ * the actions. It's a linked list and should only be
+ * free'd through unload_actions_file() unless there's
+ * only a single entry.
*/
struct url_actions
{
- struct url_spec url[1]; /**< URL pattern. */
+ struct url_spec url[1]; /**< The URL or tag pattern. */
- struct action_spec action[1]; /**< Actions. */
+ struct action_spec *action; /**< Action settings that might be shared with
+ the list entry before or after the current
+ one and can't be free'd willy nilly. */
- struct url_actions * next; /**< Next action in file, or NULL. */
+ struct url_actions *next; /**< Next action section in file, or NULL. */
};
*/
#define CSP_FLAG_HOST_HEADER_IS_SET 0x00000200UL
+/**
+ * Flag for csp->flags: Set if filtering is disabled by X-Filter: No
+ * XXX: As we now have tags we might as well ditch this.
+ */
+#define CSP_FLAG_NO_FILTERING 0x00000400UL
+
/*
* Flags for use in return codes of child processes
As a number. */
long ip_addr_long;
- /** Our IP address. I.e. the IP address that the client used to reach us,
- as a string. */
- char *my_ip_addr_str;
-
- /** Our hostname. I.e. the reverse DNS of the IP address that the client
- used to reach us, as a string. */
- char *my_hostname;
-
/** The URL that was requested */
struct http_request http[1];
+ /*
+ * The final forwarding settings.
+ * XXX: Currently this is only used for forward-override,
+ * so we can free the space in sweep.
+ */
+ struct forward_spec * fwd;
+
/** An I/O buffer used for buffering data read from the network */
struct iob iob[1];
/** List of all headers for this request */
struct list headers[1];
- /** List of all cookies for this request */
- struct list cookie_list[1];
+ /** List of all tags that apply to this request */
+ struct list tags[1];
/** MIME-Type key, see CT_* above */
unsigned int content_type;
- /** The "X-Forwarded-For:" header sent by the client */
- char *x_forwarded;
-
/** Actions files associated with this client */
struct file_list *actions_list[MAX_AF_FILES];
#define SOCKS_NONE 0 /**< Don't use a SOCKS server */
#define SOCKS_4 40 /**< original SOCKS 4 protocol */
#define SOCKS_4A 41 /**< as modified for hosts w/o external DNS */
+#define SOCKS_5 50 /**< as modified for hosts w/o external DNS */
/**
/** URL pattern that this forward_spec is for. */
struct url_spec url[1];
- /** Connection type. Must be SOCKS_NONE, SOCKS_4, or SOCKS_4A. */
+ /** Connection type. Must be SOCKS_NONE, SOCKS_4, SOCKS_4A or SOCKS_5. */
int type;
/** SOCKS server hostname. Only valid if "type" is SOCKS_4 or SOCKS_4A. */
*/
#define FORWARD_SPEC_INITIALIZER { { URL_SPEC_INITIALIZER }, 0, NULL, 0, NULL, 0, NULL }
+/* Supported filter types */
+#define FT_CONTENT_FILTER 0
+#define FT_CLIENT_HEADER_FILTER 1
+#define FT_SERVER_HEADER_FILTER 2
+#define FT_CLIENT_HEADER_TAGGER 3
+#define FT_SERVER_HEADER_TAGGER 4
+
+#define MAX_FILTER_TYPES 5
/**
* This struct represents one filter (one block) from
char *description; /**< Description from FILTER: statement in re_filterfile. */
struct list patterns[1]; /**< The patterns from the re_filterfile. */
pcrs_job *joblist; /**< The resulting compiled pcrs_jobs. */
+ int type; /**< Filter type (content, client-header, server-header). */
+ int dynamic; /**< Set to one if the pattern might contain variables
+ and has to be recompiled for every request. */
struct re_filterfile_spec *next; /**< The pointer for chaining. */
};
/** configuration_spec::feature_flags: Check the host header for requests with host-less request lines. */
#define RUNTIME_FEATURE_ACCEPT_INTERCEPTED_REQUESTS 16
+/** configuration_spec::feature_flags: Don't allow to circumvent blocks with the force prefix. */
+#define RUNTIME_FEATURE_ENFORCE_BLOCKS 32
+
+/** configuration_spec::feature_flags: Allow to block or redirect CGI requests. */
+#define RUNTIME_FEATURE_CGI_CRUNCHING 64
+
/**
* Data loaded from the configuration file.
/** The short names of the pcre filter files. */
const char *re_filterfile_short[MAX_AF_FILES];
+ /** The hostname to show on CGI pages, or NULL to use the real one. */
+ const char *hostname;
+
#ifdef FEATURE_COOKIE_JAR
/** The file name of the cookie jar file */
projects / privoxy.git / blobdiff