process.c
author Mahlon E. Smith <mahlon@laika.com>
Wed, 25 Jul 2012 11:47:12 -0700
changeset 30 5cc836e06759
parent 29 c5d00a24af56
child 32 6dc2d52e4b13
permissions -rw-r--r--
branch merge
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
10
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
     1
/* vim: set noet nosta sw=4 ts=4 ft=c : */
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
     2
/*
22
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
     3
Copyright (c) 2011-2012, Mahlon E. Smith <mahlon@martini.nu>
10
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
     4
All rights reserved.
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
     5
Redistribution and use in source and binary forms, with or without
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
     6
modification, are permitted provided that the following conditions are met:
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
     7
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
     8
    * Redistributions of source code must retain the above copyright
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
     9
      notice, this list of conditions and the following disclaimer.
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    10
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    11
    * Redistributions in binary form must reproduce the above copyright
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    12
      notice, this list of conditions and the following disclaimer in the
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    13
      documentation and/or other materials provided with the distribution.
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    14
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    15
    * Neither the name of Mahlon E. Smith nor the names of his
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    16
      contributors may be used to endorse or promote products derived
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    17
      from this software without specific prior written permission.
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    18
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    19
THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND ANY
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    20
EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    21
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    22
DISCLAIMED. IN NO EVENT SHALL THE REGENTS AND CONTRIBUTORS BE LIABLE FOR ANY
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    23
DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    24
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    25
LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    26
ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    27
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    28
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    29
*/
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    30
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    31
#include "volta.h"
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    32
#include "db.h"
22
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
    33
#include "lua.h"
10
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    34
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    35
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    36
/*
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    37
 * Given a redirect +line+ from squid, send it to the parser,
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    38
 * perform database lookups, and conditonally perform the rewrite.
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    39
 *
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    40
 */
10
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    41
void
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    42
process( char *line )
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    43
{
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    44
	parsed *p_request = parse_request( line ), *rule = NULL;
10
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    45
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    46
	/* count lines in debugmode */
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    47
	if ( v.debugmode > 2 ) v.timer.lines++;
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    48
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    49
	/* If request parsing failed, return a blank line to squid
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    50
	   to allow the request to pass through unmolested. */
29
c5d00a24af56 Add support for Squid's 'url_rewriter_concurrency' pipelining.
Mahlon E. Smith <mahlon@laika.com>
parents: 22
diff changeset
    51
	if ( p_request == NULL || p_request->valid == 0 )
18
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
    52
		return pass( p_request, rule );
10
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
    53
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    54
	/*
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    55
	 * Main rewrite logic.
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    56
	 *
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    57
	 * First, try and match the host exactly.
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    58
	 *
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    59
	 * Second, match the TLD of the host, so separate rules aren't needed for
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    60
	 * every possible subdomain of one particular domain.
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    61
	 *
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    62
	 * Finally, look for '*', if for some reason the rules provided don't care to
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    63
	 * match specific hosts, and instead just match on any path.
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    64
	 *
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    65
	 * If DB matches are found at any step above, the rules are tried in order
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    66
	 * to attempt a match against the path.  Exact string match attempted
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    67
	 * first, then fallback to regexp.
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    68
	 * 
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    69
	 * First rule match wins, and elements of the URL are rewritten based on
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    70
	 * what is present in the rule -- any missing parts just use the original
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    71
	 * URL element.  (this way, you can rewrite just the host and leave the
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    72
	 * path intact, or redir to https, for example.)
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    73
	 *
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    74
	 */
15
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
    75
	rule = find_rule( p_request->host, p_request );
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
    76
	if ( rule == NULL ) rule = find_rule( p_request->tld, p_request );
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
    77
	if ( rule == NULL ) rule = find_rule( "*", p_request );
13
23a242d7b7fa 1st iteration of volta actually doing something. Process the request,
Mahlon E. Smith <mahlon@laika.com>
parents: 10
diff changeset
    78
18
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
    79
	/* no matching rule still or negated rule?  no need to rewrite anything. */
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
    80
	if ( rule == NULL || rule->negate )
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
    81
		return pass( p_request, rule );
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
    82
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
    83
	/* avoid trivial redirect loops */
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
    84
	if (
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
    85
		( rule->redir ) &&
22
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
    86
		( rule->scheme == NULL || ( p_request->scheme && ( strcmp(p_request->scheme, rule->scheme) == 0) )) &&
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
    87
		( rule->path   == NULL || ( strcmp(p_request->path, rule->path) == 0) ) &&
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
    88
		( strcmp( p_request->host, rule->host ) == 0 )
18
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
    89
	   ) {
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
    90
		debug( 2, LOC, "Potential rewrite loop, skipping rewrite.\n" );
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
    91
		return pass( p_request, rule );
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
    92
	}
18
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
    93
22
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
    94
	/* At this point we know we'll be doing a rewrite. */
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
    95
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
    96
	/* Pass the request to lua for processing if we saw a 'lua:' tag. */
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
    97
	if ( rule->lua == 1 ) {
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
    98
		char *rewrite_string = luaV_run( p_request, rule->luapath );
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
    99
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   100
		/* the script returned nil, or otherwise had an error. */
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   101
		if ( rewrite_string == NULL ) return pass( p_request, rule );
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   102
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   103
		/* send squid the lua return value. */
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   104
		if ( v.debugmode < 5 ) {
29
c5d00a24af56 Add support for Squid's 'url_rewriter_concurrency' pipelining.
Mahlon E. Smith <mahlon@laika.com>
parents: 22
diff changeset
   105
			if ( p_request->chid ) printf( "%s", p_request->chid );
22
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   106
			puts( rewrite_string );
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   107
			fflush( stdout );
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   108
		}
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   109
	}
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   110
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   111
	/* otherwise, perform the rewrite internally. */
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   112
	else {
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   113
		rewrite( p_request, rule );
822094314703 Add the ability to optionally script rewrite logic using Lua.
Mahlon E. Smith <mahlon@martini.nu>
parents: 18
diff changeset
   114
	}
13
23a242d7b7fa 1st iteration of volta actually doing something. Process the request,
Mahlon E. Smith <mahlon@laika.com>
parents: 10
diff changeset
   115
15
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   116
	finish_parsed( rule );
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   117
	finish_parsed( p_request );
10
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
   118
	return;
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
   119
}
d07309450285 Get the ragel line parser properly tokenizing the input lines. Add a
Mahlon E. Smith <mahlon@martini.nu>
parents:
diff changeset
   120
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   121
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   122
/*
18
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   123
 * Allow the request to pass through without being rewritten.
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   124
 *
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   125
 */
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   126
void
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   127
pass( parsed *request, parsed *rule )
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   128
{
29
c5d00a24af56 Add support for Squid's 'url_rewriter_concurrency' pipelining.
Mahlon E. Smith <mahlon@laika.com>
parents: 22
diff changeset
   129
	if ( v.debugmode >= 5 ) {
c5d00a24af56 Add support for Squid's 'url_rewriter_concurrency' pipelining.
Mahlon E. Smith <mahlon@laika.com>
parents: 22
diff changeset
   130
		finish_parsed( rule );
c5d00a24af56 Add support for Squid's 'url_rewriter_concurrency' pipelining.
Mahlon E. Smith <mahlon@laika.com>
parents: 22
diff changeset
   131
		finish_parsed( request );
c5d00a24af56 Add support for Squid's 'url_rewriter_concurrency' pipelining.
Mahlon E. Smith <mahlon@laika.com>
parents: 22
diff changeset
   132
		return;
c5d00a24af56 Add support for Squid's 'url_rewriter_concurrency' pipelining.
Mahlon E. Smith <mahlon@laika.com>
parents: 22
diff changeset
   133
	}
18
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   134
29
c5d00a24af56 Add support for Squid's 'url_rewriter_concurrency' pipelining.
Mahlon E. Smith <mahlon@laika.com>
parents: 22
diff changeset
   135
	if ( request && request->chid ) printf( "%s", request->chid );
18
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   136
	printf( "\n" );
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   137
	fflush( stdout );
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   138
29
c5d00a24af56 Add support for Squid's 'url_rewriter_concurrency' pipelining.
Mahlon E. Smith <mahlon@laika.com>
parents: 22
diff changeset
   139
	finish_parsed( rule );
c5d00a24af56 Add support for Squid's 'url_rewriter_concurrency' pipelining.
Mahlon E. Smith <mahlon@laika.com>
parents: 22
diff changeset
   140
	finish_parsed( request );
18
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   141
	return;
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   142
}
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   143
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   144
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   145
/*
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   146
 * Output a rewritten URL for squid.
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   147
 *
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   148
 */
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   149
void
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   150
rewrite( parsed *request, parsed *rule )
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   151
{
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   152
	if ( rule == NULL || v.debugmode >= 5 ) return;
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   153
29
c5d00a24af56 Add support for Squid's 'url_rewriter_concurrency' pipelining.
Mahlon E. Smith <mahlon@laika.com>
parents: 22
diff changeset
   154
	if ( request->chid ) printf( "%s", request->chid );
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   155
	if ( rule->redir ) printf( "%s:", rule->redir );
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   156
	printf( "%s%s", (rule->scheme ? rule->scheme : request->scheme), rule->host );
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   157
	if ( rule->port ) printf( ":%s", rule->port );
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   158
	printf( "%s", rule->path ? rule->path : request->path );
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   159
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   160
	printf("\n");
16
e6a640ad2cc2 Ensure that all output is flushed immediately.
Mahlon E. Smith <mahlon@martini.nu>
parents: 15
diff changeset
   161
	fflush( stdout );
18
d4ce82194b64 - 1st pass at documentation with the README.
Mahlon E. Smith <mahlon@laika.com>
parents: 16
diff changeset
   162
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   163
	return;
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   164
}
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   165
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   166
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   167
/*
15
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   168
 * Compare a parsed +rule+ against the +request+.
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   169
 * Returns 1 on a match, 0 otherwise.
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   170
 *
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   171
 */
15
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   172
unsigned short int
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   173
check_rule( parsed *rule, parsed *p_request )
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   174
{
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   175
	int re_rv;
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   176
	regex_t re;
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   177
	char re_err[128];
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   178
15
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   179
	if ( rule == NULL || p_request->path == NULL ) return( 0 );
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   180
15
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   181
	/* quick comparison */
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   182
	if ( (strcasecmp( rule->path_re, p_request->path ) == 0) ||
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   183
			(strcmp( rule->path_re, "*" ) == 0) ) {
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   184
		debug( 4, LOC, "Rule match \"%s\" (non regexp)\n", rule->path_re );
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   185
		return( 1 );
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   186
	}
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   187
15
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   188
	/* compile the regexp */
16
e6a640ad2cc2 Ensure that all output is flushed immediately.
Mahlon E. Smith <mahlon@martini.nu>
parents: 15
diff changeset
   189
	if ( (re_rv = regcomp( &re, rule->path_re, REG_EXTENDED | REG_NOSUB | REG_ICASE )) != 0 ) {
15
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   190
		regerror( re_rv, &re, re_err, 128 );
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   191
		debug( 4, LOC, "Invalid regex: \"%s\": %s\n", rule->path_re, re_err );
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   192
		regfree( &re );
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   193
		return( 0 );
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   194
	}
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   195
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   196
	/* compare! */
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   197
	if ( (regexec( &re, p_request->path, 0, NULL, 0 )) == 0 ) {
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   198
		debug( 4, LOC, "Rule match \"%s\" (regexp)\n", rule->path_re );
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   199
		regfree( &re );
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   200
		return( 1 );
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   201
	}
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   202
	else {
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   203
		regfree( &re );
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   204
		return( 0 );
2706fc514dea Add whitelisting rules, to negate other matches if they come first in
Mahlon E. Smith <mahlon@martini.nu>
parents: 14
diff changeset
   205
	}
14
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   206
}
51eb85ae4de4 There isn't a fast way to look up ( value exists or null ) for every
Mahlon E. Smith <mahlon@martini.nu>
parents: 13
diff changeset
   207