-const char loaders_rcs[] = "$Id: loaders.c,v 1.29 2001/10/23 21:38:53 jongfoster Exp $";
+const char loaders_rcs[] = "$Id: loaders.c,v 1.89 2012/03/09 16:24:36 fabiankeil Exp $";
/*********************************************************************
*
* File : $Source: /cvsroot/ijbswa/current/loaders.c,v $
* the list of active loaders, and to automatically
* unload files that are no longer in use.
*
- * Copyright : Written by and Copyright (C) 2001 the SourceForge
- * IJBSWA team. http://ijbswa.sourceforge.net
+ * Copyright : Written by and Copyright (C) 2001-2010 the
+ * Privoxy team. http://www.privoxy.org/
*
* Based on the Internet Junkbuster originally written
* by and Copyright (C) 1997 Anonymous Coders and
* or write to the Free Software Foundation, Inc., 59
* Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*
- * Revisions :
- * $Log: loaders.c,v $
- * Revision 1.29 2001/10/23 21:38:53 jongfoster
- * Adding error-checking to create_url_spec()
- *
- * Revision 1.28 2001/10/07 15:40:39 oes
- * Replaced 6 boolean members of csp with one bitmap (csp->flags)
- *
- * Revision 1.27 2001/09/22 16:36:59 jongfoster
- * Removing unused parameter fs from read_config_line()
- *
- * Revision 1.26 2001/09/22 14:05:22 jongfoster
- * Bugfix: Multiple escaped "#" characters in a configuration
- * file are now permitted.
- * Also removing 3 unused headers.
- *
- * Revision 1.25 2001/09/13 22:44:03 jongfoster
- * Adding {} to an if statement
- *
- * Revision 1.24 2001/07/30 22:08:36 jongfoster
- * Tidying up #defines:
- * - All feature #defines are now of the form FEATURE_xxx
- * - Permanently turned off WIN_GUI_EDIT
- * - Permanently turned on WEBDAV and SPLIT_PROXY_ARGS
- *
- * Revision 1.23 2001/07/20 15:51:54 oes
- * Fixed indentation of prepocessor commands
- *
- * Revision 1.22 2001/07/20 15:16:17 haroon
- * - per Guy's suggestion, added a while loop in sweep() to catch not just
- * the last inactive CSP but all other consecutive inactive CSPs after that
- * as well
- *
- * Revision 1.21 2001/07/18 17:26:24 oes
- * Changed to conform to new pcrs interface
- *
- * Revision 1.20 2001/07/17 13:07:01 oes
- * Fixed segv when last line in config files
- * lacked a terminating (\r)\n
- *
- * Revision 1.19 2001/07/13 14:01:54 oes
- * Removed all #ifdef PCRS
- *
- * Revision 1.18 2001/06/29 21:45:41 oes
- * Indentation, CRLF->LF, Tab-> Space
- *
- * Revision 1.17 2001/06/29 13:31:51 oes
- * Various adaptions
- *
- * Revision 1.16 2001/06/09 10:55:28 jongfoster
- * Changing BUFSIZ ==> BUFFER_SIZE
- *
- * Revision 1.15 2001/06/07 23:14:14 jongfoster
- * Removing ACL and forward file loaders - these
- * files have been merged into the config file.
- * Cosmetic: Moving unloader funcs next to their
- * respective loader funcs
- *
- * Revision 1.14 2001/06/01 03:27:04 oes
- * Fixed line continuation problem
- *
- * Revision 1.13 2001/05/31 21:28:49 jongfoster
- * Removed all permissionsfile code - it's now called the actions
- * file, and (almost) all the code is in actions.c
- *
- * Revision 1.12 2001/05/31 17:32:31 oes
- *
- * - Enhanced domain part globbing with infix and prefix asterisk
- * matching and optional unanchored operation
- *
- * Revision 1.11 2001/05/29 23:25:24 oes
- *
- * - load_config_line() and load_permissions_file() now use chomp()
- *
- * Revision 1.10 2001/05/29 09:50:24 jongfoster
- * Unified blocklist/imagelist/permissionslist.
- * File format is still under discussion, but the internal changes
- * are (mostly) done.
- *
- * Also modified interceptor behaviour:
- * - We now intercept all URLs beginning with one of the following
- * prefixes (and *only* these prefixes):
- * * http://i.j.b/
- * * http://ijbswa.sf.net/config/
- * * http://ijbswa.sourceforge.net/config/
- * - New interceptors "home page" - go to http://i.j.b/ to see it.
- * - Internal changes so that intercepted and fast redirect pages
- * are not replaced with an image.
- * - Interceptors now have the option to send a binary page direct
- * to the client. (i.e. ijb-send-banner uses this)
- * - Implemented show-url-info interceptor. (Which is why I needed
- * the above interceptors changes - a typical URL is
- * "http://i.j.b/show-url-info?url=www.somesite.com/banner.gif".
- * The previous mechanism would not have intercepted that, and
- * if it had been intercepted then it then it would have replaced
- * it with an image.)
- *
- * Revision 1.9 2001/05/26 17:12:07 jongfoster
- * Fatal errors loading configuration files now give better error messages.
- *
- * Revision 1.8 2001/05/26 00:55:20 jongfoster
- * Removing duplicated code. load_forwardfile() now uses create_url_spec()
- *
- * Revision 1.7 2001/05/26 00:28:36 jongfoster
- * Automatic reloading of config file.
- * Removed obsolete SIGHUP support (Unix) and Reload menu option (Win32).
- * Most of the global variables have been moved to a new
- * struct configuration_spec, accessed through csp->config->globalname
- * Most of the globals remaining are used by the Win32 GUI.
- *
- * Revision 1.6 2001/05/23 12:27:33 oes
- *
- * Fixed ugly indentation of my last changes
- *
- * Revision 1.5 2001/05/23 10:39:05 oes
- * - Added support for escaping the comment character
- * in config files by a backslash
- * - Added support for line continuation in config
- * files
- * - Fixed a buffer overflow bug with long config lines
- *
- * Revision 1.4 2001/05/22 18:56:28 oes
- * CRLF -> LF
- *
- * Revision 1.3 2001/05/20 01:21:20 jongfoster
- * Version 2.9.4 checkin.
- * - Merged popupfile and cookiefile, and added control over PCRS
- * filtering, in new "permissionsfile".
- * - Implemented LOG_LEVEL_FATAL, so that if there is a configuration
- * file error you now get a message box (in the Win32 GUI) rather
- * than the program exiting with no explanation.
- * - Made killpopup use the PCRS MIME-type checking and HTTP-header
- * skipping.
- * - Removed tabs from "config"
- * - Moved duplicated url parsing code in "loaders.c" to a new funcition.
- * - Bumped up version number.
- *
- * Revision 1.2 2001/05/17 23:01:01 oes
- * - Cleaned CRLF's from the sources and related files
- *
- * Revision 1.1.1.1 2001/05/15 13:58:59 oes
- * Initial import of version 2.9.3 source tree
- *
- *
*********************************************************************/
-\f
+
#include "config.h"
#include <stdlib.h>
#include <sys/types.h>
#include <string.h>
-#include <malloc.h>
#include <errno.h>
#include <sys/stat.h>
#include <ctype.h>
#include "miscutil.h"
#include "errlog.h"
#include "actions.h"
+#include "urlmatch.h"
+#include "encode.h"
const char loaders_h_rcs[] = LOADERS_H_VERSION;
-/* Fix a problem with Solaris. There should be no effect on other
- * platforms.
- * Solaris's isspace() is a macro which uses it's argument directly
- * as an array index. Therefore we need to make sure that high-bit
- * characters generate +ve values, and ideally we also want to make
- * the argument match the declared parameter type of "int".
- */
-#define ijb_isspace(__X) isspace((int)(unsigned char)(__X))
-
-
/*
* Currently active files.
* These are also entered in the main linked list of files.
static struct file_list *current_trustfile = NULL;
#endif /* def FEATURE_TRUST */
-static struct file_list *current_re_filterfile = NULL;
+static int load_one_re_filterfile(struct client_state *csp, int fileid);
+
+static struct file_list *current_re_filterfile[MAX_AF_FILES] = {
+ NULL, NULL, NULL, NULL, NULL,
+ NULL, NULL, NULL, NULL, NULL
+};
+/*
+ * Pseudo filter type for load_one_re_filterfile
+ */
+#define NO_NEW_FILTER -1
/*********************************************************************
*
* Parameters : None
*
- * Returns : N/A
+ * Returns : The number of threads that are still active.
*
*********************************************************************/
-void sweep(void)
+unsigned int sweep(void)
{
struct file_list *fl, *nfl;
- struct client_state *csp, *ncsp;
+ struct client_state *csp;
+ struct client_states *last_active, *client_list;
+ int i;
+ unsigned int active_threads = 0;
/* clear all of the file's active flags */
- for ( fl = files->next; NULL != fl; fl = fl->next )
+ for (fl = files->next; NULL != fl; fl = fl->next)
{
fl->active = 0;
}
- for (csp = clients; csp && (ncsp = csp->next) ; csp = csp->next)
+ last_active = clients;
+ client_list = clients->next;
+
+ while (NULL != client_list)
{
- if (ncsp->flags & CSP_FLAG_ACTIVE)
+ csp = &client_list->csp;
+ if (csp->flags & CSP_FLAG_ACTIVE)
{
- /* mark this client's files as active */
+ /* Mark this client's files as active */
/*
* Always have a configuration file.
* (Also note the slightly non-standard extra
* indirection here.)
*/
- ncsp->config->config_file_list->active = 1;
+ csp->config->config_file_list->active = 1;
- if (ncsp->actions_list) /* actions files */
+ /*
+ * Actions files
+ */
+ for (i = 0; i < MAX_AF_FILES; i++)
{
- ncsp->actions_list->active = 1;
+ if (csp->actions_list[i])
+ {
+ csp->actions_list[i]->active = 1;
+ }
}
- if (ncsp->rlist) /* pcrsjob files */
+ /*
+ * Filter files
+ */
+ for (i = 0; i < MAX_AF_FILES; i++)
{
- ncsp->rlist->active = 1;
+ if (csp->rlist[i])
+ {
+ csp->rlist[i]->active = 1;
+ }
}
+ /*
+ * Trust file
+ */
#ifdef FEATURE_TRUST
- if (ncsp->tlist) /* trust files */
+ if (csp->tlist)
{
- ncsp->tlist->active = 1;
+ csp->tlist->active = 1;
}
#endif /* def FEATURE_TRUST */
+ active_threads++;
+
+ last_active = client_list;
+ client_list = client_list->next;
}
else
/*
- * this client is not active, release its resources
- * and the ones of all inactive clients that might
- * follow it
+ * This client is not active. Free its resources.
*/
{
- while (!(ncsp->flags & CSP_FLAG_ACTIVE))
- {
- csp->next = ncsp->next;
+ last_active->next = client_list->next;
- freez(ncsp->ip_addr_str);
- freez(ncsp->my_ip_addr_str);
- freez(ncsp->my_hostname);
-
-#ifdef FEATURE_TRUST
- freez(ncsp->referrer);
-#endif /* def FEATURE_TRUST */
- freez(ncsp->x_forwarded);
- freez(ncsp->iob->buf);
+ freez(csp->ip_addr_str);
+ freez(csp->iob->buf);
+ freez(csp->error_message);
- free_http_request(ncsp->http);
+ if (csp->action->flags & ACTION_FORWARD_OVERRIDE &&
+ NULL != csp->fwd)
+ {
+ unload_forward_spec(csp->fwd);
+ }
+ free_http_request(csp->http);
- destroy_list(ncsp->headers);
- destroy_list(ncsp->cookie_list);
+ destroy_list(csp->headers);
+ destroy_list(csp->tags);
- free_current_action(ncsp->action);
+ free_current_action(csp->action);
#ifdef FEATURE_STATISTICS
- urls_read++;
- if (ncsp->flags & CSP_FLAG_REJECTED)
- {
- urls_rejected++;
- }
-#endif /* def FEATURE_STATISTICS */
-
- freez(ncsp);
-
- /* are there any more in sequence after it? */
- if( !(ncsp = csp->next) )
- break;
+ urls_read++;
+ if (csp->flags & CSP_FLAG_REJECTED)
+ {
+ urls_rejected++;
}
- }
- }
-
- for (fl = files; fl && (nfl = fl->next) ; fl = fl->next)
- {
- if ( ( 0 == nfl->active ) && ( NULL != nfl->unloader ) )
- {
- fl->next = nfl->next;
-
- (nfl->unloader)(nfl->f);
+#endif /* def FEATURE_STATISTICS */
- freez(nfl->filename);
+ freez(client_list);
- freez(nfl);
+ client_list = last_active->next;
}
}
-}
-
-
-/*********************************************************************
- *
- * Function : create_url_spec
- *
- * Description : Creates a "url_spec" structure from a string.
- * When finished, free with unload_url().
- *
- * Parameters :
- * 1 : url = Target url_spec to be filled in. Must be
- * zeroed out before the call (e.g. using zalloc).
- * 2 : buf = Source pattern, null terminated. NOTE: The
- * contents of this buffer are destroyed by this
- * function. If this function succeeds, the
- * buffer is copied to url->spec. If this
- * function fails, the contents of the buffer
- * are lost forever.
- *
- * Returns : JB_ERR_OK - Success
- * JB_ERR_MEMORY - Out of memory
- * JB_ERR_PARSE - Cannot parse regex (Detailed message
- * written to system log)
- *
- *********************************************************************/
-jb_err create_url_spec(struct url_spec * url, char * buf)
-{
- char *p;
- struct url_spec tmp_url[1];
-
- assert(url);
- assert(buf);
-
- /* save a copy of the orignal specification */
- if ((url->spec = strdup(buf)) == NULL)
- {
- return JB_ERR_MEMORY;
- }
+ nfl = files;
+ fl = files->next;
- if ((p = strchr(buf, '/')))
+ while (fl != NULL)
{
- if (NULL == (url->path = strdup(p)))
+ if ((0 == fl->active) && (NULL != fl->unloader))
{
- freez(url->spec);
- return JB_ERR_MEMORY;
- }
- url->pathlen = strlen(url->path);
- *p = '\0';
- }
- else
- {
- url->path = NULL;
- url->pathlen = 0;
- }
-#ifdef REGEX
- if (url->path)
- {
- int errcode;
- char rebuf[BUFFER_SIZE];
+ nfl->next = fl->next;
- if (NULL == (url->preg = zalloc(sizeof(*url->preg))))
- {
- freez(url->spec);
- freez(url->path);
- return JB_ERR_MEMORY;
- }
+ (fl->unloader)(fl->f);
- sprintf(rebuf, "^(%s)", url->path);
+ freez(fl->filename);
+ freez(fl);
- errcode = regcomp(url->preg, rebuf,
- (REG_EXTENDED|REG_NOSUB|REG_ICASE));
- if (errcode)
+ fl = nfl->next;
+ }
+ else
{
- size_t errlen = regerror(errcode,
- url->preg, rebuf, sizeof(rebuf));
-
- if (errlen > (sizeof(rebuf) - (size_t)1))
- {
- errlen = sizeof(rebuf) - (size_t)1;
- }
- rebuf[errlen] = '\0';
-
- log_error(LOG_LEVEL_ERROR, "error compiling %s: %s",
- url->spec, rebuf);
-
- freez(url->spec);
- freez(url->path);
- freez(url->preg);
-
- return JB_ERR_PARSE;
+ nfl = fl;
+ fl = fl->next;
}
}
-#endif
- if ((p = strchr(buf, ':')) == NULL)
- {
- url->port = 0;
- }
- else
- {
- *p++ = '\0';
- url->port = atoi(p);
- }
-
- if ((url->domain = strdup(buf)) == NULL)
- {
- freez(url->spec);
- freez(url->path);
-#ifdef REGEX
- freez(url->preg);
-#endif /* def REGEX */
- return JB_ERR_MEMORY;
- }
-
- /* split domain into components */
-
- *tmp_url = dsplit(url->domain);
- if (tmp_url->dbuf == NULL)
- {
- freez(url->spec);
- freez(url->path);
- freez(url->domain);
-#ifdef REGEX
- freez(url->preg);
-#endif /* def REGEX */
- return JB_ERR_MEMORY;
- }
-
- url->dbuf = tmp_url->dbuf;
- url->dcnt = tmp_url->dcnt;
- url->dvec = tmp_url->dvec;
- url->unanchored = tmp_url->unanchored;
-
- return JB_ERR_OK;
-
-}
-
-/*********************************************************************
- *
- * Function : free_url
- *
- * Description : Called from the "unloaders". Freez the url
- * structure elements.
- *
- * Parameters :
- * 1 : url = pointer to a url_spec structure.
- *
- * Returns : N/A
- *
- *********************************************************************/
-void free_url(struct url_spec *url)
-{
- if (url == NULL) return;
-
- freez(url->spec);
- freez(url->domain);
- freez(url->dbuf);
- freez(url->dvec);
- freez(url->path);
-#ifdef REGEX
- if (url->preg)
- {
- regfree(url->preg);
- freez(url->preg);
- }
-#endif
+ return active_threads;
}
}
fs = (struct file_list *)zalloc(sizeof(struct file_list));
-
if (fs == NULL)
{
/* Out of memory error */
return 1;
}
+
fs->filename = strdup(filename);
fs->lastmodified = statbuf->st_mtime;
freez (fs);
return 1;
}
-
-
*newfl = fs;
return 1;
+}
+
+
+/*********************************************************************
+ *
+ * Function : simple_read_line
+ *
+ * Description : Read a single line from a file and return it.
+ * This is basically a version of fgets() that malloc()s
+ * it's own line buffer. Note that the buffer will
+ * always be a multiple of BUFFER_SIZE bytes long.
+ * Therefore if you are going to keep the string for
+ * an extended period of time, you should probably
+ * strdup() it and free() the original, to save memory.
+ *
+ *
+ * Parameters :
+ * 1 : dest = destination for newly malloc'd pointer to
+ * line data. Will be set to NULL on error.
+ * 2 : fp = File to read from
+ * 3 : newline = Standard for newlines in the file.
+ * Will be unchanged if it's value on input is not
+ * NEWLINE_UNKNOWN.
+ * On output, may be changed from NEWLINE_UNKNOWN to
+ * actual convention in file.
+ *
+ * Returns : JB_ERR_OK on success
+ * JB_ERR_MEMORY on out-of-memory
+ * JB_ERR_FILE on EOF.
+ *
+ *********************************************************************/
+jb_err simple_read_line(FILE *fp, char **dest, int *newline)
+{
+ size_t len = 0;
+ size_t buflen = BUFFER_SIZE;
+ char * buf;
+ char * p;
+ int ch;
+ int realnewline = NEWLINE_UNKNOWN;
+
+ if (NULL == (buf = malloc(buflen)))
+ {
+ return JB_ERR_MEMORY;
+ }
+
+ p = buf;
+
+/*
+ * Character codes. If you have a weird compiler and the following are
+ * incorrect, you also need to fix NEWLINE() in loaders.h
+ */
+#define CHAR_CR '\r' /* ASCII 13 */
+#define CHAR_LF '\n' /* ASCII 10 */
+
+ for (;;)
+ {
+ ch = getc(fp);
+ if (ch == EOF)
+ {
+ if (len > 0)
+ {
+ *p = '\0';
+ *dest = buf;
+ return JB_ERR_OK;
+ }
+ else
+ {
+ free(buf);
+ *dest = NULL;
+ return JB_ERR_FILE;
+ }
+ }
+ else if (ch == CHAR_CR)
+ {
+ ch = getc(fp);
+ if (ch == CHAR_LF)
+ {
+ if (*newline == NEWLINE_UNKNOWN)
+ {
+ *newline = NEWLINE_DOS;
+ }
+ }
+ else
+ {
+ if (ch != EOF)
+ {
+ ungetc(ch, fp);
+ }
+ if (*newline == NEWLINE_UNKNOWN)
+ {
+ *newline = NEWLINE_MAC;
+ }
+ }
+ *p = '\0';
+ *dest = buf;
+ if (*newline == NEWLINE_UNKNOWN)
+ {
+ *newline = realnewline;
+ }
+ return JB_ERR_OK;
+ }
+ else if (ch == CHAR_LF)
+ {
+ *p = '\0';
+ *dest = buf;
+ if (*newline == NEWLINE_UNKNOWN)
+ {
+ *newline = NEWLINE_UNIX;
+ }
+ return JB_ERR_OK;
+ }
+ else if (ch == 0)
+ {
+ *p = '\0';
+ *dest = buf;
+ return JB_ERR_OK;
+ }
+ *p++ = (char)ch;
+
+ if (++len >= buflen)
+ {
+ buflen += BUFFER_SIZE;
+ if (NULL == (p = realloc(buf, buflen)))
+ {
+ free(buf);
+ return JB_ERR_MEMORY;
+ }
+ buf = p;
+ p = buf + len;
+ }
+ }
}
/*********************************************************************
*
- * Function : read_config_line
+ * Function : edit_read_line
*
* Description : Read a single non-empty line from a file and return
* it. Trims comments, leading and trailing whitespace
* and respects escaping of newline and comment char.
+ * Provides the line in 2 alternative forms: raw and
+ * preprocessed.
+ * - raw is the raw data read from the file. If the
+ * line is not modified, then this should be written
+ * to the new file.
+ * - prefix is any comments and blank lines that were
+ * read from the file. If the line is modified, then
+ * this should be written out to the file followed
+ * by the modified data. (If this string is non-empty
+ * then it will have a newline at the end).
+ * - data is the actual data that will be parsed
+ * further by appropriate routines.
+ * On EOF, the 3 strings will all be set to NULL and
+ * 0 will be returned.
*
* Parameters :
- * 1 : buf = Buffer to use.
- * 2 : buflen = Size of buffer in bytes.
- * 3 : fp = File to read from
- *
- * Returns : NULL on EOF or error
- * Otherwise, returns buf.
+ * 1 : fp = File to read from
+ * 2 : raw_out = destination for newly malloc'd pointer to
+ * raw line data. May be NULL if you don't want it.
+ * 3 : prefix_out = destination for newly malloc'd pointer to
+ * comments. May be NULL if you don't want it.
+ * 4 : data_out = destination for newly malloc'd pointer to
+ * line data with comments and leading/trailing spaces
+ * removed, and line continuation performed. May be
+ * NULL if you don't want it.
+ * 5 : newline = Standard for newlines in the file.
+ * On input, set to value to use or NEWLINE_UNKNOWN.
+ * On output, may be changed from NEWLINE_UNKNOWN to
+ * actual convention in file. May be NULL if you
+ * don't want it.
+ * 6 : line_number = Line number in file. In "lines" as
+ * reported by a text editor, not lines containing data.
+ *
+ * Returns : JB_ERR_OK on success
+ * JB_ERR_MEMORY on out-of-memory
+ * JB_ERR_FILE on EOF.
*
*********************************************************************/
-char *read_config_line(char *buf, int buflen, FILE *fp)
+jb_err edit_read_line(FILE *fp,
+ char **raw_out,
+ char **prefix_out,
+ char **data_out,
+ int *newline,
+ unsigned long *line_number)
{
- char *p;
- char *src;
- char *dest;
- char linebuf[BUFFER_SIZE];
- int contflag = 0;
+ char *p; /* Temporary pointer */
+ char *linebuf; /* Line read from file */
+ char *linestart; /* Start of linebuf, usually first non-whitespace char */
+ int contflag = 0; /* Nonzero for line continuation - i.e. line ends '\' */
+ int is_empty = 1; /* Flag if not got any data yet */
+ char *raw = NULL; /* String to be stored in raw_out */
+ char *prefix = NULL; /* String to be stored in prefix_out */
+ char *data = NULL; /* String to be stored in data_out */
+ int scrapnewline; /* Used for (*newline) if newline==NULL */
+ jb_err rval = JB_ERR_OK;
+
+ assert(fp);
+ assert(raw_out || data_out);
+ assert(newline == NULL
+ || *newline == NEWLINE_UNKNOWN
+ || *newline == NEWLINE_UNIX
+ || *newline == NEWLINE_DOS
+ || *newline == NEWLINE_MAC);
+
+ if (newline == NULL)
+ {
+ scrapnewline = NEWLINE_UNKNOWN;
+ newline = &scrapnewline;
+ }
+
+ /* Set output parameters to NULL */
+ if (raw_out)
+ {
+ *raw_out = NULL;
+ }
+ if (prefix_out)
+ {
+ *prefix_out = NULL;
+ }
+ if (data_out)
+ {
+ *data_out = NULL;
+ }
- *buf = '\0';
+ /* Set string variables to new, empty strings. */
- while (fgets(linebuf, sizeof(linebuf), fp))
+ if (raw_out)
{
- /* Trim off newline */
- if ((p = strpbrk(linebuf, "\r\n")) != NULL)
+ raw = strdup("");
+ if (NULL == raw)
{
- *p = '\0';
+ return JB_ERR_MEMORY;
}
- else
+ }
+ if (prefix_out)
+ {
+ prefix = strdup("");
+ if (NULL == prefix)
+ {
+ freez(raw);
+ return JB_ERR_MEMORY;
+ }
+ }
+ if (data_out)
+ {
+ data = strdup("");
+ if (NULL == data)
{
- p = linebuf + strlen(linebuf);
+ freez(raw);
+ freez(prefix);
+ return JB_ERR_MEMORY;
+ }
+ }
+
+ /* Main loop. Loop while we need more data & it's not EOF. */
+
+ while ((contflag || is_empty)
+ && (JB_ERR_OK == (rval = simple_read_line(fp, &linebuf, newline))))
+ {
+ if (line_number)
+ {
+ (*line_number)++;
+ }
+ if (raw)
+ {
+ string_append(&raw,linebuf);
+ if (string_append(&raw,NEWLINE(*newline)))
+ {
+ freez(prefix);
+ freez(data);
+ free(linebuf);
+ return JB_ERR_MEMORY;
+ }
}
/* Line continuation? Trim escape and set flag. */
- if ((p != linebuf) && (*--p == '\\'))
+ p = linebuf + strlen(linebuf) - 1;
+ contflag = ((*linebuf != '\0') && (*p == '\\'));
+ if (contflag)
{
- contflag = 1;
*p = '\0';
}
- /* If there's a comment char.. */
- p = linebuf;
+ /* Trim leading spaces if we're at the start of the line */
+ linestart = linebuf;
+ assert(NULL != data);
+ if (*data == '\0')
+ {
+ /* Trim leading spaces */
+ while (*linestart && isspace((int)(unsigned char)*linestart))
+ {
+ linestart++;
+ }
+ }
+
+ /* Handle comment characters. */
+ p = linestart;
while ((p = strchr(p, '#')) != NULL)
{
- /* ..and it's escaped, left-shift the line over the escape. */
+ /* Found a comment char.. */
if ((p != linebuf) && (*(p-1) == '\\'))
{
- src = p;
- dest = p - 1;
- while ((*dest++ = *src++) != '\0')
+ /* ..and it's escaped, left-shift the line over the escape. */
+ char *q = p - 1;
+ while ((*q = *(q + 1)) != '\0')
{
- /* nop */
+ q++;
}
/* Now scan from just after the "#". */
}
- /* Else, chop off the rest of the line */
else
{
+ /* Real comment. Save it... */
+ if (p == linestart)
+ {
+ /* Special case: Line only contains a comment, so all the
+ * previous whitespace is considered part of the comment.
+ * Undo the whitespace skipping, if any.
+ */
+ linestart = linebuf;
+ p = linestart;
+ }
+ if (prefix)
+ {
+ string_append(&prefix,p);
+ if (string_append(&prefix, NEWLINE(*newline)))
+ {
+ freez(raw);
+ freez(data);
+ free(linebuf);
+ return JB_ERR_MEMORY;
+ }
+ }
+
+ /* ... and chop off the rest of the line */
*p = '\0';
}
- }
+ } /* END while (there's a # character) */
/* Write to the buffer */
- if (*linebuf)
+ if (*linestart)
{
- strncat(buf, linebuf, buflen - strlen(buf));
+ is_empty = 0;
+ if (data)
+ {
+ if (string_append(&data, linestart))
+ {
+ freez(raw);
+ freez(prefix);
+ free(linebuf);
+ return JB_ERR_MEMORY;
+ }
+ }
}
- /* Continue? */
- if (contflag)
- {
- contflag = 0;
- continue;
- }
+ free(linebuf);
+ } /* END while(we need more data) */
+
+ /* Handle simple_read_line() errors - ignore EOF */
+ if ((rval != JB_ERR_OK) && (rval != JB_ERR_FILE))
+ {
+ freez(raw);
+ freez(prefix);
+ freez(data);
+ return rval;
+ }
+
+ if (raw ? (*raw == '\0') : is_empty)
+ {
+ /* EOF and no data there. (Definition of "data" depends on whether
+ * the caller cares about "raw" or just "data").
+ */
+
+ freez(raw);
+ freez(prefix);
+ freez(data);
+
+ return JB_ERR_FILE;
+ }
+ else
+ {
+ /* Got at least some data */
- /* Remove leading and trailing whitespace */
- chomp(buf);
+ /* Remove trailing whitespace */
+ chomp(data);
- if (*buf)
+ if (raw_out)
{
- return buf;
+ *raw_out = raw;
}
+ else
+ {
+ freez(raw);
+ }
+ if (prefix_out)
+ {
+ *prefix_out = prefix;
+ }
+ else
+ {
+ freez(prefix);
+ }
+ if (data_out)
+ {
+ *data_out = data;
+ }
+ else
+ {
+ freez(data);
+ }
+ return JB_ERR_OK;
}
+}
- /* EOF */
- return NULL;
+/*********************************************************************
+ *
+ * Function : read_config_line
+ *
+ * Description : Read a single non-empty line from a file and return
+ * it. Trims comments, leading and trailing whitespace
+ * and respects escaping of newline and comment char.
+ *
+ * Parameters :
+ * 1 : fp = File to read from
+ * 2 : linenum = linenumber in file
+ * 3 : buf = Pointer to a pointer to set to the data buffer.
+ *
+ * Returns : NULL on EOF or error
+ * Otherwise, returns buf.
+ *
+ *********************************************************************/
+char *read_config_line(FILE *fp, unsigned long *linenum, char **buf)
+{
+ jb_err err;
+ err = edit_read_line(fp, NULL, NULL, buf, NULL, linenum);
+ if (err)
+ {
+ if (err == JB_ERR_MEMORY)
+ {
+ log_error(LOG_LEVEL_FATAL, "Out of memory loading a config file");
+ }
+ *buf = NULL;
+ }
+ return *buf;
}
*********************************************************************/
static void unload_trustfile(void *f)
{
- struct block_spec *b = (struct block_spec *)f;
- if (b == NULL) return;
+ struct block_spec *cur = (struct block_spec *)f;
+ struct block_spec *next;
+
+ while (cur != NULL)
+ {
+ next = cur->next;
+
+ free_url_spec(cur->url);
+ free(cur);
- unload_trustfile(b->next); /* Stack is cheap, isn't it? */
+ cur = next;
+ }
- free_url(b->url);
+}
- freez(b);
+#ifdef FEATURE_GRACEFUL_TERMINATION
+/*********************************************************************
+ *
+ * Function : unload_current_trust_file
+ *
+ * Description : Unloads current trust file - reset to state at
+ * beginning of program.
+ *
+ * Parameters : None
+ *
+ * Returns : N/A
+ *
+ *********************************************************************/
+void unload_current_trust_file(void)
+{
+ if (current_trustfile)
+ {
+ current_trustfile->unloader = unload_trustfile;
+ current_trustfile = NULL;
+ }
}
+#endif /* FEATURE_GRACEFUL_TERMINATION */
/*********************************************************************
struct block_spec *b, *bl;
struct url_spec **tl;
- char buf[BUFFER_SIZE], *p, *q;
+ char *buf = NULL;
int reject, trusted;
struct file_list *fs;
+ unsigned long linenum = 0;
+ int trusted_referrers = 0;
if (!check_file_changed(current_trustfile, csp->config->trustfile, &fs))
{
/* No need to load */
- if (csp)
- {
- csp->tlist = current_trustfile;
- }
+ csp->tlist = current_trustfile;
return(0);
}
if (!fs)
{
goto load_trustfile_error;
}
+ log_error(LOG_LEVEL_INFO, "Loading trust file: %s", csp->config->trustfile);
tl = csp->config->trust_list;
- while (read_config_line(buf, sizeof(buf), fp) != NULL)
+ while (read_config_line(fp, &linenum, &buf) != NULL)
{
trusted = 0;
reject = 1;
if (*buf == '~')
{
+ char *p;
+ char *q;
+
reject = 0;
p = buf;
q = p+1;
- while ((*p++ = *q++))
+ while ((*p++ = *q++) != '\0')
{
/* nop */
}
/* skip blank lines */
if (*buf == '\0')
{
+ freez(buf);
continue;
}
*/
if (trusted)
{
- *tl++ = b->url;
+ if (++trusted_referrers < MAX_TRUSTED_REFERRERS)
+ {
+ *tl++ = b->url;
+ }
}
+ freez(buf);
+ }
+
+ if (trusted_referrers >= MAX_TRUSTED_REFERRERS)
+ {
+ /*
+ * FIXME: ... after Privoxy 3.0.4 is out.
+ */
+ log_error(LOG_LEVEL_ERROR, "Too many trusted referrers. Current limit is %d, you are using %d.\n"
+ " Additional trusted referrers are treated like ordinary trusted URLs.\n"
+ " (You can increase this limit by changing MAX_TRUSTED_REFERRERS in project.h and recompiling).",
+ MAX_TRUSTED_REFERRERS, trusted_referrers);
}
*tl = NULL;
fs->next = files->next;
files->next = fs;
current_trustfile = fs;
-
- if (csp)
- {
- csp->tlist = fs;
- }
+ csp->tlist = fs;
return(0);
load_trustfile_error:
log_error(LOG_LEVEL_FATAL, "can't load trustfile '%s': %E",
- csp->config->trustfile);
+ csp->config->trustfile);
+ freez(buf);
return(-1);
}
*
* Function : unload_re_filterfile
*
- * Description : Unload the re_filter list.
+ * Description : Unload the re_filter list by freeing all chained
+ * re_filterfile specs and their data.
*
* Parameters :
* 1 : f = the data structure associated with the filterfile.
*********************************************************************/
static void unload_re_filterfile(void *f)
{
- struct re_filterfile_spec *b = (struct re_filterfile_spec *)f;
+ struct re_filterfile_spec *a, *b = (struct re_filterfile_spec *)f;
- if (b == NULL)
+ while (b != NULL)
{
- return;
+ a = b->next;
+
+ destroy_list(b->patterns);
+ pcrs_free_joblist(b->joblist);
+ freez(b->name);
+ freez(b->description);
+ freez(b);
+
+ b = a;
}
- destroy_list(b->patterns);
- pcrs_free_joblist(b->joblist);
- freez(b);
+ return;
+}
+
+/*********************************************************************
+ *
+ * Function : unload_forward_spec
+ *
+ * Description : Unload the forward spec settings by freeing all
+ * memory referenced by members and the memory for
+ * the spec itself.
+ *
+ * Parameters :
+ * 1 : fwd = the forward spec.
+ *
+ * Returns : N/A
+ *
+ *********************************************************************/
+void unload_forward_spec(struct forward_spec *fwd)
+{
+ free_url_spec(fwd->url);
+ freez(fwd->gateway_host);
+ freez(fwd->forward_host);
+ free(fwd);
return;
}
+
+#ifdef FEATURE_GRACEFUL_TERMINATION
+/*********************************************************************
+ *
+ * Function : unload_current_re_filterfile
+ *
+ * Description : Unloads current re_filter file - reset to state at
+ * beginning of program.
+ *
+ * Parameters : None
+ *
+ * Returns : N/A
+ *
+ *********************************************************************/
+void unload_current_re_filterfile(void)
+{
+ int i;
+
+ for (i = 0; i < MAX_AF_FILES; i++)
+ {
+ if (current_re_filterfile[i])
+ {
+ current_re_filterfile[i]->unloader = unload_re_filterfile;
+ current_re_filterfile[i] = NULL;
+ }
+ }
+}
+#endif
+
+
/*********************************************************************
*
- * Function : load_re_filterfile
+ * Function : load_re_filterfiles
*
- * Description : Load the re_filterfile. Each non-comment, non-empty
- * line is instantly added to the joblist, which is
- * a chained list of pcrs_job structs.
+ * Description : Loads all the filterfiles.
+ * Generate a chained list of re_filterfile_spec's from
+ * the "FILTER: " blocks, compiling all their substitutions
+ * into chained lists of pcrs_job structs.
*
* Parameters :
* 1 : csp = Current client state (buffers, headers, etc...)
* Returns : 0 => Ok, everything else is an error.
*
*********************************************************************/
-int load_re_filterfile(struct client_state *csp)
+int load_re_filterfiles(struct client_state *csp)
+{
+ int i;
+ int result;
+
+ for (i = 0; i < MAX_AF_FILES; i++)
+ {
+ if (csp->config->re_filterfile[i])
+ {
+ result = load_one_re_filterfile(csp, i);
+ if (result)
+ {
+ return result;
+ }
+ }
+ else if (current_re_filterfile[i])
+ {
+ current_re_filterfile[i]->unloader = unload_re_filterfile;
+ current_re_filterfile[i] = NULL;
+ }
+ }
+
+ return 0;
+}
+
+
+/*********************************************************************
+ *
+ * Function : load_one_re_filterfile
+ *
+ * Description : Load a re_filterfile.
+ * Generate a chained list of re_filterfile_spec's from
+ * the "FILTER: " blocks, compiling all their substitutions
+ * into chained lists of pcrs_job structs.
+ *
+ * Parameters :
+ * 1 : csp = Current client state (buffers, headers, etc...)
+ *
+ * Returns : 0 => Ok, everything else is an error.
+ *
+ *********************************************************************/
+int load_one_re_filterfile(struct client_state *csp, int fileid)
{
FILE *fp;
- struct re_filterfile_spec *bl;
+ struct re_filterfile_spec *new_bl, *bl = NULL;
struct file_list *fs;
- char buf[BUFFER_SIZE];
+ char *buf = NULL;
int error;
- pcrs_job *dummy;
+ unsigned long linenum = 0;
+ pcrs_job *dummy, *lastjob = NULL;
- if (!check_file_changed(current_re_filterfile, csp->config->re_filterfile, &fs))
+ /*
+ * No need to reload if unchanged
+ */
+ if (!check_file_changed(current_re_filterfile[fileid], csp->config->re_filterfile[fileid], &fs))
{
- /* No need to load */
- if (csp)
- {
- csp->rlist = current_re_filterfile;
- }
+ csp->rlist[fileid] = current_re_filterfile[fileid];
return(0);
}
if (!fs)
goto load_re_filterfile_error;
}
- fs->f = bl = (struct re_filterfile_spec *)zalloc(sizeof(*bl));
- if (bl == NULL)
+ /*
+ * Open the file or fail
+ */
+ if ((fp = fopen(csp->config->re_filterfile[fileid], "r")) == NULL)
{
goto load_re_filterfile_error;
}
- /* Open the file or fail */
- if ((fp = fopen(csp->config->re_filterfile, "r")) == NULL)
- {
- goto load_re_filterfile_error;
- }
+ log_error(LOG_LEVEL_INFO, "Loading filter file: %s", csp->config->re_filterfile[fileid]);
- /* Read line by line */
- while (read_config_line(buf, sizeof(buf), fp) != NULL)
+ /*
+ * Read line by line
+ */
+ while (read_config_line(fp, &linenum, &buf) != NULL)
{
- enlist( bl->patterns, buf );
+ int new_filter = NO_NEW_FILTER;
+
+ if (strncmp(buf, "FILTER:", 7) == 0)
+ {
+ new_filter = FT_CONTENT_FILTER;
+ }
+ else if (strncmp(buf, "SERVER-HEADER-FILTER:", 21) == 0)
+ {
+ new_filter = FT_SERVER_HEADER_FILTER;
+ }
+ else if (strncmp(buf, "CLIENT-HEADER-FILTER:", 21) == 0)
+ {
+ new_filter = FT_CLIENT_HEADER_FILTER;
+ }
+ else if (strncmp(buf, "CLIENT-HEADER-TAGGER:", 21) == 0)
+ {
+ new_filter = FT_CLIENT_HEADER_TAGGER;
+ }
+ else if (strncmp(buf, "SERVER-HEADER-TAGGER:", 21) == 0)
+ {
+ new_filter = FT_SERVER_HEADER_TAGGER;
+ }
- /* We have a meaningful line -> make it a job */
- if ((dummy = pcrs_compile_command(buf, &error)) == NULL)
+ /*
+ * If this is the head of a new filter block, make it a
+ * re_filterfile spec of its own and chain it to the list:
+ */
+ if (new_filter != NO_NEW_FILTER)
{
- log_error(LOG_LEVEL_RE_FILTER,
- "Adding re_filter job %s failed with error %d.", buf, error);
+ new_bl = (struct re_filterfile_spec *)zalloc(sizeof(*bl));
+ if (new_bl == NULL)
+ {
+ goto load_re_filterfile_error;
+ }
+ if (new_filter == FT_CONTENT_FILTER)
+ {
+ new_bl->name = chomp(buf + 7);
+ }
+ else
+ {
+ new_bl->name = chomp(buf + 21);
+ }
+ new_bl->type = new_filter;
+
+ /*
+ * If a filter description is available,
+ * encode it to HTML and save it.
+ */
+ if (NULL != (new_bl->description = strpbrk(new_bl->name, " \t")))
+ {
+ *new_bl->description++ = '\0';
+ new_bl->description = html_encode(chomp(new_bl->description));
+ if (NULL == new_bl->description)
+ {
+ new_bl->description = strdup("Out of memory while encoding this filter's description to HTML");
+ }
+ }
+ else
+ {
+ new_bl->description = strdup("No description available for this filter");
+ }
+
+ new_bl->name = strdup(chomp(new_bl->name));
+
+ /*
+ * If this is the first filter block, chain it
+ * to the file_list rather than its (nonexistant)
+ * predecessor
+ */
+ if (fs->f == NULL)
+ {
+ fs->f = new_bl;
+ }
+ else
+ {
+ assert(NULL != bl);
+ bl->next = new_bl;
+ }
+ bl = new_bl;
+
+ log_error(LOG_LEVEL_RE_FILTER, "Reading in filter \"%s\" (\"%s\")", bl->name, bl->description);
+
+ freez(buf);
continue;
}
+
+ /*
+ * Else, save the expression, make it a pcrs_job
+ * and chain it into the current filter's joblist
+ */
+ if (bl != NULL)
+ {
+ error = enlist(bl->patterns, buf);
+ if (JB_ERR_MEMORY == error)
+ {
+ log_error(LOG_LEVEL_FATAL,
+ "Out of memory while enlisting re_filter job \'%s\' for filter %s.", buf, bl->name);
+ }
+ assert(JB_ERR_OK == error);
+
+ if (pcrs_job_is_dynamic(buf))
+ {
+ /*
+ * Dynamic pattern that might contain variables
+ * and has to be recompiled for every request
+ */
+ if (bl->joblist != NULL)
+ {
+ pcrs_free_joblist(bl->joblist);
+ bl->joblist = NULL;
+ }
+ bl->dynamic = 1;
+ log_error(LOG_LEVEL_RE_FILTER,
+ "Adding dynamic re_filter job \'%s\' to filter %s succeeded.", buf, bl->name);
+ freez(buf);
+ continue;
+ }
+ else if (bl->dynamic)
+ {
+ /*
+ * A previous job was dynamic and as we
+ * recompile the whole filter anyway, it
+ * makes no sense to compile this job now.
+ */
+ log_error(LOG_LEVEL_RE_FILTER,
+ "Adding static re_filter job \'%s\' to dynamic filter %s succeeded.", buf, bl->name);
+ freez(buf);
+ continue;
+ }
+
+ if ((dummy = pcrs_compile_command(buf, &error)) == NULL)
+ {
+ log_error(LOG_LEVEL_ERROR,
+ "Adding re_filter job \'%s\' to filter %s failed with error %d.", buf, bl->name, error);
+ freez(buf);
+ continue;
+ }
+ else
+ {
+ if (bl->joblist == NULL)
+ {
+ bl->joblist = dummy;
+ }
+ else if (NULL != lastjob)
+ {
+ lastjob->next = dummy;
+ }
+ lastjob = dummy;
+ log_error(LOG_LEVEL_RE_FILTER, "Adding re_filter job \'%s\' to filter %s succeeded.", buf, bl->name);
+ }
+ }
else
{
- dummy->next = bl->joblist;
- bl->joblist = dummy;
- log_error(LOG_LEVEL_RE_FILTER, "Adding re_filter job %s succeeded.", buf);
+ log_error(LOG_LEVEL_ERROR, "Ignoring job %s outside filter block in %s, line %d",
+ buf, csp->config->re_filterfile[fileid], linenum);
}
+ freez(buf);
}
fclose(fp);
- /* the old one is now obsolete */
- if ( NULL != current_re_filterfile )
+ /*
+ * Schedule the now-obsolete old data for unloading
+ */
+ if (NULL != current_re_filterfile[fileid])
{
- current_re_filterfile->unloader = unload_re_filterfile;
+ current_re_filterfile[fileid]->unloader = unload_re_filterfile;
}
+ /*
+ * Chain this file into the global list of loaded files
+ */
fs->next = files->next;
files->next = fs;
- current_re_filterfile = fs;
+ current_re_filterfile[fileid] = fs;
+ csp->rlist[fileid] = fs;
- if (csp)
- {
- csp->rlist = fs;
- }
-
- return( 0 );
+ return(0);
load_re_filterfile_error:
log_error(LOG_LEVEL_FATAL, "can't load re_filterfile '%s': %E",
- csp->config->re_filterfile);
+ csp->config->re_filterfile[fileid]);
return(-1);
}
{
int i;
- for (i=0; i < NLOADERS; i++)
+ for (i = 0; i < NLOADERS; i++)
{
if (config->loaders[i] == NULL)
{
int ret = 0;
int i;
- for (i=0; i < NLOADERS; i++)
+ for (i = 0; i < NLOADERS; i++)
{
if (csp->config->loaders[i] == NULL)
{
}
+/*********************************************************************
+ *
+ * Function : file_has_been_modified
+ *
+ * Description : Helper function to check if a file has been changed
+ *
+ * Parameters :
+ * 1 : filename = The name of the file to check
+ * 2 : last_known_modification = The time of the last known
+ * modification
+ *
+ * Returns : TRUE if the file has been changed,
+ * FALSE otherwise.
+ *
+ *********************************************************************/
+static int file_has_been_modified(const char *filename, time_t last_know_modification)
+{
+ struct stat statbuf[1];
+
+ if (stat(filename, statbuf) < 0)
+ {
+ /* Error, probably file not found which counts as change. */
+ return 1;
+ }
+
+ return (last_know_modification != statbuf->st_mtime);
+}
+
+
+/*********************************************************************
+ *
+ * Function : any_loaded_file_changed
+ *
+ * Description : Helper function to check if any loaded file has been
+ * changed since the time it has been loaded.
+ *
+ * XXX: Should we cache the return value for x seconds?
+ *
+ * Parameters :
+ * 1 : files_to_check = List of files to check
+ *
+ * Returns : TRUE if any file has been changed,
+ * FALSE otherwise.
+ *
+ *********************************************************************/
+int any_loaded_file_changed(const struct file_list *files_to_check)
+{
+ const struct file_list *file_to_check = files_to_check;
+
+ while (file_to_check != NULL)
+ {
+ if (file_has_been_modified(file_to_check->filename, file_to_check->lastmodified))
+ {
+ return TRUE;
+ }
+ file_to_check = file_to_check->next;
+ }
+ return FALSE;
+}
+
/*
Local Variables:
tab-width: 3
end:
-*/
\ No newline at end of file
+*/