unicorn 0.1.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/.document +12 -0
- data/.gitignore +12 -0
- data/CHANGELOG +19 -0
- data/CONTRIBUTORS +20 -0
- data/DESIGN +80 -0
- data/GNUmakefile +111 -0
- data/LICENSE +53 -0
- data/Manifest +41 -0
- data/README +73 -0
- data/Rakefile +37 -0
- data/SIGNALS +34 -0
- data/TODO +5 -0
- data/bin/unicorn +189 -0
- data/ext/unicorn/http11/ext_help.h +15 -0
- data/ext/unicorn/http11/extconf.rb +5 -0
- data/ext/unicorn/http11/http11.c +526 -0
- data/ext/unicorn/http11/http11_parser.c +1220 -0
- data/ext/unicorn/http11/http11_parser.h +45 -0
- data/ext/unicorn/http11/http11_parser.rl +153 -0
- data/ext/unicorn/http11/http11_parser_common.rl +55 -0
- data/lib/unicorn.rb +548 -0
- data/lib/unicorn/configurator.rb +253 -0
- data/lib/unicorn/const.rb +116 -0
- data/lib/unicorn/http_request.rb +178 -0
- data/lib/unicorn/http_response.rb +72 -0
- data/lib/unicorn/socket.rb +142 -0
- data/lib/unicorn/util.rb +40 -0
- data/setup.rb +1585 -0
- data/test/aggregate.rb +13 -0
- data/test/benchmark/previous.rb +11 -0
- data/test/benchmark/simple.rb +11 -0
- data/test/benchmark/utils.rb +82 -0
- data/test/exec/README +5 -0
- data/test/exec/test_exec.rb +570 -0
- data/test/test_helper.rb +103 -0
- data/test/tools/trickletest.rb +45 -0
- data/test/unit/test_configurator.rb +48 -0
- data/test/unit/test_http_parser.rb +161 -0
- data/test/unit/test_response.rb +45 -0
- data/test/unit/test_server.rb +96 -0
- data/test/unit/test_upload.rb +151 -0
- data/unicorn.gemspec +35 -0
- metadata +122 -0
data/SIGNALS
ADDED
@@ -0,0 +1,34 @@
|
|
1
|
+
== Signal handling
|
2
|
+
|
3
|
+
In general, signals need only be sent to the master process. However,
|
4
|
+
the signals unicorn uses internally to communicate with the worker
|
5
|
+
processes are documented here as well.
|
6
|
+
|
7
|
+
=== Master Process
|
8
|
+
|
9
|
+
* HUP - reload config file and gracefully restart all workers
|
10
|
+
|
11
|
+
* INT/TERM - quick shutdown, kills all workers immediately
|
12
|
+
|
13
|
+
* QUIT - graceful shutdown, waits for workers to finish their
|
14
|
+
current request before finishing.
|
15
|
+
|
16
|
+
* USR1 - reopen all logs owned by the master and all workers
|
17
|
+
See Unicorn::Util.reopen_logs for what is considered a log.
|
18
|
+
|
19
|
+
* USR2 - reexecute the running binary. A separate QUIT
|
20
|
+
should be sent to the original process once the child is verified to
|
21
|
+
be up and running.
|
22
|
+
|
23
|
+
=== Worker Processes
|
24
|
+
|
25
|
+
Sending signals directly to the worker processes should not normally be
|
26
|
+
needed. If the master process is running, any exited worker will be
|
27
|
+
automatically respawned.
|
28
|
+
|
29
|
+
* INT/TERM - quick shutdown, immediately exit
|
30
|
+
|
31
|
+
* QUIT - gracefully exit after finishing the current request
|
32
|
+
|
33
|
+
* USR1 - reopen all logs owned by the worker process
|
34
|
+
See Unicorn::Util.reopen_logs for what is considered a log.
|
data/TODO
ADDED
data/bin/unicorn
ADDED
@@ -0,0 +1,189 @@
|
|
1
|
+
#!/home/ew/bin/ruby
|
2
|
+
$stdin.sync = $stdout.sync = $stderr.sync = true
|
3
|
+
require 'unicorn' # require this first to populate Unicorn::DEFAULT_START_CTX
|
4
|
+
require 'optparse'
|
5
|
+
|
6
|
+
env = "development"
|
7
|
+
daemonize = false
|
8
|
+
listeners = []
|
9
|
+
options = { :listeners => listeners }
|
10
|
+
host = Unicorn::Const::DEFAULT_HOST
|
11
|
+
port = Unicorn::Const::DEFAULT_PORT
|
12
|
+
|
13
|
+
opts = OptionParser.new("", 24, ' ') do |opts|
|
14
|
+
opts.banner = "Usage: #{File.basename($0)} " \
|
15
|
+
"[ruby options] [unicorn options] [rackup config file]"
|
16
|
+
|
17
|
+
opts.separator "Ruby options:"
|
18
|
+
|
19
|
+
lineno = 1
|
20
|
+
opts.on("-e", "--eval LINE", "evaluate a LINE of code") do |line|
|
21
|
+
eval line, TOPLEVEL_BINDING, "-e", lineno
|
22
|
+
lineno += 1
|
23
|
+
end
|
24
|
+
|
25
|
+
opts.on("-d", "--debug", "set debugging flags (set $DEBUG to true)") do
|
26
|
+
$DEBUG = true
|
27
|
+
end
|
28
|
+
|
29
|
+
opts.on("-w", "--warn", "turn warnings on for your script") do
|
30
|
+
$-w = true
|
31
|
+
end
|
32
|
+
|
33
|
+
opts.on("-I", "--include PATH",
|
34
|
+
"specify $LOAD_PATH (may be used more than once)") do |path|
|
35
|
+
$LOAD_PATH.unshift(*path.split(/:/))
|
36
|
+
end
|
37
|
+
|
38
|
+
opts.on("-r", "--require LIBRARY",
|
39
|
+
"require the library, before executing your script") do |library|
|
40
|
+
require library
|
41
|
+
end
|
42
|
+
|
43
|
+
opts.separator "Unicorn options:"
|
44
|
+
|
45
|
+
# some of these switches exist for rackup command-line compatibility,
|
46
|
+
|
47
|
+
opts.on("-o", "--host HOST",
|
48
|
+
"listen on HOST (default: #{Unicorn::Const::DEFAULT_HOST})") do |h|
|
49
|
+
host = h
|
50
|
+
end
|
51
|
+
|
52
|
+
opts.on("-p", "--port PORT",
|
53
|
+
"use PORT (default: #{Unicorn::Const::DEFAULT_PORT})") do |p|
|
54
|
+
port = p.to_i
|
55
|
+
end
|
56
|
+
|
57
|
+
opts.on("-E", "--env ENVIRONMENT",
|
58
|
+
"use ENVIRONMENT for defaults (default: development)") do |e|
|
59
|
+
env = e
|
60
|
+
end
|
61
|
+
|
62
|
+
opts.on("-D", "--daemonize", "run daemonized in the background") do |d|
|
63
|
+
daemonize = d ? true : false
|
64
|
+
end
|
65
|
+
|
66
|
+
opts.on("-P", "--pid FILE", "file to store PID (default: none)") do |f|
|
67
|
+
options[:pid] = File.expand_path(f)
|
68
|
+
end
|
69
|
+
|
70
|
+
opts.on("-s", "--server SERVER",
|
71
|
+
"this flag only exists for compatibility") do |s|
|
72
|
+
warn "-s/--server only exists for compatibility with rackup"
|
73
|
+
end
|
74
|
+
|
75
|
+
# Unicorn-specific stuff
|
76
|
+
opts.on("-l", "--listen {HOST:PORT|PATH}",
|
77
|
+
"listen on HOST:PORT or PATH",
|
78
|
+
"this may be specified multiple times",
|
79
|
+
"(default: #{Unicorn::Const::DEFAULT_LISTEN})") do |address|
|
80
|
+
listeners << address
|
81
|
+
end
|
82
|
+
|
83
|
+
opts.on("-c", "--config-file FILE", "Unicorn-specific config file") do |f|
|
84
|
+
options[:config_file] = File.expand_path(f)
|
85
|
+
end
|
86
|
+
|
87
|
+
# I'm avoiding Unicorn-specific config options on the command-line.
|
88
|
+
# IMNSHO, config options on the command-line are redundant given
|
89
|
+
# config files and make things unnecessarily complicated with multiple
|
90
|
+
# places to look for a config option.
|
91
|
+
|
92
|
+
opts.separator "Common options:"
|
93
|
+
|
94
|
+
opts.on_tail("-h", "--help", "Show this message") do
|
95
|
+
puts opts
|
96
|
+
exit
|
97
|
+
end
|
98
|
+
|
99
|
+
opts.on_tail("-v", "--version", "Show version") do
|
100
|
+
puts "unicorn v#{Unicorn::Const::UNICORN_VERSION}"
|
101
|
+
exit
|
102
|
+
end
|
103
|
+
|
104
|
+
opts.parse! ARGV
|
105
|
+
end
|
106
|
+
|
107
|
+
require 'pp' if $DEBUG
|
108
|
+
|
109
|
+
# require Rack as late as possible in case $LOAD_PATH is modified
|
110
|
+
# in config.ru or command-line
|
111
|
+
require 'rack'
|
112
|
+
|
113
|
+
config = ARGV[0] || "config.ru"
|
114
|
+
abort "configuration file #{config} not found" unless File.exist?(config)
|
115
|
+
|
116
|
+
inner_app = case config
|
117
|
+
when /\.ru$/
|
118
|
+
raw = File.open(config, "rb") { |fp| fp.sysread(fp.stat.size) }
|
119
|
+
# parse embedded command-line options in config.ru comments
|
120
|
+
if raw[/^#\\(.*)/]
|
121
|
+
opts.parse! $1.split(/\s+/)
|
122
|
+
end
|
123
|
+
lambda { || eval("Rack::Builder.new {(#{raw}\n)}.to_app", nil, config) }
|
124
|
+
else
|
125
|
+
lambda do ||
|
126
|
+
require config
|
127
|
+
Object.const_get(File.basename(config, '.rb').capitalize)
|
128
|
+
end
|
129
|
+
end
|
130
|
+
|
131
|
+
app = case env
|
132
|
+
when "development"
|
133
|
+
lambda do ||
|
134
|
+
Rack::Builder.new do
|
135
|
+
use Rack::CommonLogger, $stderr
|
136
|
+
use Rack::ShowExceptions
|
137
|
+
use Rack::Lint
|
138
|
+
run inner_app.call
|
139
|
+
end.to_app
|
140
|
+
end
|
141
|
+
when "deployment"
|
142
|
+
lambda do ||
|
143
|
+
Rack::Builder.new do
|
144
|
+
use Rack::CommonLogger, $stderr
|
145
|
+
run inner_app.call
|
146
|
+
end.to_app
|
147
|
+
end
|
148
|
+
else
|
149
|
+
inner_app
|
150
|
+
end
|
151
|
+
|
152
|
+
if listeners.empty?
|
153
|
+
listener = "#{host}:#{port}"
|
154
|
+
listeners << listener
|
155
|
+
end
|
156
|
+
|
157
|
+
if $DEBUG
|
158
|
+
pp({
|
159
|
+
:unicorn_options => options,
|
160
|
+
:app => app,
|
161
|
+
:inner_app => inner_app,
|
162
|
+
:daemonize => daemonize,
|
163
|
+
})
|
164
|
+
end
|
165
|
+
|
166
|
+
# only daemonize if we're not inheriting file descriptors from our parent
|
167
|
+
if daemonize
|
168
|
+
|
169
|
+
$stdin.reopen("/dev/null")
|
170
|
+
unless ENV['UNICORN_FD']
|
171
|
+
exit if fork
|
172
|
+
Process.setsid
|
173
|
+
exit if fork
|
174
|
+
end
|
175
|
+
|
176
|
+
# We don't do a lot of standard daemonization stuff:
|
177
|
+
# * $stderr/$stderr can/will be redirected separately
|
178
|
+
# * umask is whatever was set by the parent process at startup
|
179
|
+
# and can be set in config.ru and config_file, so making it
|
180
|
+
# 0000 and potentially exposing sensitive log data can be bad
|
181
|
+
# policy.
|
182
|
+
# * Don't bother to chdir here since Unicorn is designed to
|
183
|
+
# run inside APP_ROOT. Unicorn will also re-chdir() to
|
184
|
+
# the directory it was started in when being re-executed
|
185
|
+
# to pickup code changes if the original deployment directory
|
186
|
+
# is a symlink or otherwise got replaced.
|
187
|
+
end
|
188
|
+
|
189
|
+
Unicorn.run(app, options)
|
@@ -0,0 +1,15 @@
|
|
1
|
+
#ifndef ext_help_h
|
2
|
+
#define ext_help_h
|
3
|
+
|
4
|
+
#define RAISE_NOT_NULL(T) if(T == NULL) rb_raise(rb_eArgError, "NULL found for " # T " when shouldn't be.");
|
5
|
+
#define DATA_GET(from,type,name) Data_Get_Struct(from,type,name); RAISE_NOT_NULL(name);
|
6
|
+
#define REQUIRE_TYPE(V, T) if(TYPE(V) != T) rb_raise(rb_eTypeError, "Wrong argument type for " # V " required " # T);
|
7
|
+
#define ARRAY_SIZE(x) (sizeof(x)/sizeof(x[0]))
|
8
|
+
|
9
|
+
#ifdef DEBUG
|
10
|
+
#define TRACE() fprintf(stderr, "> %s:%d:%s\n", __FILE__, __LINE__, __FUNCTION__)
|
11
|
+
#else
|
12
|
+
#define TRACE()
|
13
|
+
#endif
|
14
|
+
|
15
|
+
#endif
|
@@ -0,0 +1,526 @@
|
|
1
|
+
/**
|
2
|
+
* Copyright (c) 2005 Zed A. Shaw
|
3
|
+
* You can redistribute it and/or modify it under the same terms as Ruby.
|
4
|
+
*/
|
5
|
+
#include "ruby.h"
|
6
|
+
#include "ext_help.h"
|
7
|
+
#include <assert.h>
|
8
|
+
#include <string.h>
|
9
|
+
#include "http11_parser.h"
|
10
|
+
|
11
|
+
#ifndef RSTRING_PTR
|
12
|
+
#define RSTRING_PTR(s) (RSTRING(s)->ptr)
|
13
|
+
#endif
|
14
|
+
#ifndef RSTRING_LEN
|
15
|
+
#define RSTRING_LEN(s) (RSTRING(s)->len)
|
16
|
+
#endif
|
17
|
+
|
18
|
+
static VALUE mUnicorn;
|
19
|
+
static VALUE cHttpParser;
|
20
|
+
static VALUE eHttpParserError;
|
21
|
+
|
22
|
+
#define HTTP_PREFIX "HTTP_"
|
23
|
+
#define HTTP_PREFIX_LEN (sizeof(HTTP_PREFIX) - 1)
|
24
|
+
|
25
|
+
static VALUE global_request_method;
|
26
|
+
static VALUE global_request_uri;
|
27
|
+
static VALUE global_fragment;
|
28
|
+
static VALUE global_query_string;
|
29
|
+
static VALUE global_http_version;
|
30
|
+
static VALUE global_content_length;
|
31
|
+
static VALUE global_http_content_length;
|
32
|
+
static VALUE global_request_path;
|
33
|
+
static VALUE global_content_type;
|
34
|
+
static VALUE global_http_content_type;
|
35
|
+
static VALUE global_http_body;
|
36
|
+
static VALUE global_gateway_interface;
|
37
|
+
static VALUE global_gateway_interface_value;
|
38
|
+
static VALUE global_server_name;
|
39
|
+
static VALUE global_server_port;
|
40
|
+
static VALUE global_server_protocol;
|
41
|
+
static VALUE global_server_protocol_value;
|
42
|
+
static VALUE global_http_host;
|
43
|
+
static VALUE global_port_80;
|
44
|
+
static VALUE global_localhost;
|
45
|
+
|
46
|
+
/** Defines common length and error messages for input length validation. */
|
47
|
+
#define DEF_MAX_LENGTH(N,length) const size_t MAX_##N##_LENGTH = length; const char *MAX_##N##_LENGTH_ERR = "HTTP element " # N " is longer than the " # length " allowed length."
|
48
|
+
|
49
|
+
/** Validates the max length of given input and throws an HttpParserError exception if over. */
|
50
|
+
#define VALIDATE_MAX_LENGTH(len, N) if(len > MAX_##N##_LENGTH) { rb_raise(eHttpParserError, MAX_##N##_LENGTH_ERR); }
|
51
|
+
|
52
|
+
/** Defines global strings in the init method. */
|
53
|
+
#define DEF_GLOBAL(N, val) global_##N = rb_obj_freeze(rb_str_new2(val)); rb_global_variable(&global_##N)
|
54
|
+
|
55
|
+
|
56
|
+
/* Defines the maximum allowed lengths for various input elements.*/
|
57
|
+
DEF_MAX_LENGTH(FIELD_NAME, 256);
|
58
|
+
DEF_MAX_LENGTH(FIELD_VALUE, 80 * 1024);
|
59
|
+
DEF_MAX_LENGTH(REQUEST_URI, 1024 * 12);
|
60
|
+
DEF_MAX_LENGTH(FRAGMENT, 1024); /* Don't know if this length is specified somewhere or not */
|
61
|
+
DEF_MAX_LENGTH(REQUEST_PATH, 1024);
|
62
|
+
DEF_MAX_LENGTH(QUERY_STRING, (1024 * 10));
|
63
|
+
DEF_MAX_LENGTH(HEADER, (1024 * (80 + 32)));
|
64
|
+
|
65
|
+
struct common_field {
|
66
|
+
const signed long len;
|
67
|
+
const char *name;
|
68
|
+
VALUE value;
|
69
|
+
};
|
70
|
+
|
71
|
+
/*
|
72
|
+
* A list of common HTTP headers we expect to receive.
|
73
|
+
* This allows us to avoid repeatedly creating identical string
|
74
|
+
* objects to be used with rb_hash_aset().
|
75
|
+
*/
|
76
|
+
static struct common_field common_http_fields[] = {
|
77
|
+
# define f(N) { (sizeof(N) - 1), N, Qnil }
|
78
|
+
f("ACCEPT"),
|
79
|
+
f("ACCEPT_CHARSET"),
|
80
|
+
f("ACCEPT_ENCODING"),
|
81
|
+
f("ACCEPT_LANGUAGE"),
|
82
|
+
f("ALLOW"),
|
83
|
+
f("AUTHORIZATION"),
|
84
|
+
f("CACHE_CONTROL"),
|
85
|
+
f("CONNECTION"),
|
86
|
+
f("CONTENT_ENCODING"),
|
87
|
+
f("CONTENT_LENGTH"),
|
88
|
+
f("CONTENT_TYPE"),
|
89
|
+
f("COOKIE"),
|
90
|
+
f("DATE"),
|
91
|
+
f("EXPECT"),
|
92
|
+
f("FROM"),
|
93
|
+
f("HOST"),
|
94
|
+
f("IF_MATCH"),
|
95
|
+
f("IF_MODIFIED_SINCE"),
|
96
|
+
f("IF_NONE_MATCH"),
|
97
|
+
f("IF_RANGE"),
|
98
|
+
f("IF_UNMODIFIED_SINCE"),
|
99
|
+
f("KEEP_ALIVE"), /* Firefox sends this */
|
100
|
+
f("MAX_FORWARDS"),
|
101
|
+
f("PRAGMA"),
|
102
|
+
f("PROXY_AUTHORIZATION"),
|
103
|
+
f("RANGE"),
|
104
|
+
f("REFERER"),
|
105
|
+
f("TE"),
|
106
|
+
f("TRAILER"),
|
107
|
+
f("TRANSFER_ENCODING"),
|
108
|
+
f("UPGRADE"),
|
109
|
+
f("USER_AGENT"),
|
110
|
+
f("VIA"),
|
111
|
+
f("X_FORWARDED_FOR"), /* common for proxies */
|
112
|
+
f("X_REAL_IP"), /* common for proxies */
|
113
|
+
f("WARNING")
|
114
|
+
# undef f
|
115
|
+
};
|
116
|
+
|
117
|
+
/*
|
118
|
+
* qsort(3) and bsearch(3) improve average performance slightly, but may
|
119
|
+
* not be worth it for lack of portability to certain platforms...
|
120
|
+
*/
|
121
|
+
#if defined(HAVE_QSORT_BSEARCH)
|
122
|
+
/* sort by length, then by name if there's a tie */
|
123
|
+
static int common_field_cmp(const void *a, const void *b)
|
124
|
+
{
|
125
|
+
struct common_field *cfa = (struct common_field *)a;
|
126
|
+
struct common_field *cfb = (struct common_field *)b;
|
127
|
+
signed long diff = cfa->len - cfb->len;
|
128
|
+
return diff ? diff : memcmp(cfa->name, cfb->name, cfa->len);
|
129
|
+
}
|
130
|
+
#endif /* HAVE_QSORT_BSEARCH */
|
131
|
+
|
132
|
+
static void init_common_fields(void)
|
133
|
+
{
|
134
|
+
int i;
|
135
|
+
struct common_field *cf = common_http_fields;
|
136
|
+
char tmp[256]; /* MAX_FIELD_NAME_LENGTH */
|
137
|
+
memcpy(tmp, HTTP_PREFIX, HTTP_PREFIX_LEN);
|
138
|
+
|
139
|
+
for(i = 0; i < ARRAY_SIZE(common_http_fields); cf++, i++) {
|
140
|
+
memcpy(tmp + HTTP_PREFIX_LEN, cf->name, cf->len + 1);
|
141
|
+
cf->value = rb_obj_freeze(rb_str_new(tmp, HTTP_PREFIX_LEN + cf->len));
|
142
|
+
rb_global_variable(&cf->value);
|
143
|
+
}
|
144
|
+
|
145
|
+
#if defined(HAVE_QSORT_BSEARCH)
|
146
|
+
qsort(common_http_fields,
|
147
|
+
ARRAY_SIZE(common_http_fields),
|
148
|
+
sizeof(struct common_field),
|
149
|
+
common_field_cmp);
|
150
|
+
#endif /* HAVE_QSORT_BSEARCH */
|
151
|
+
}
|
152
|
+
|
153
|
+
static VALUE find_common_field_value(const char *field, size_t flen)
|
154
|
+
{
|
155
|
+
#if defined(HAVE_QSORT_BSEARCH)
|
156
|
+
struct common_field key;
|
157
|
+
struct common_field *found;
|
158
|
+
key.name = field;
|
159
|
+
key.len = (signed long)flen;
|
160
|
+
found = (struct common_field *)bsearch(&key, common_http_fields,
|
161
|
+
ARRAY_SIZE(common_http_fields),
|
162
|
+
sizeof(struct common_field),
|
163
|
+
common_field_cmp);
|
164
|
+
return found ? found->value : Qnil;
|
165
|
+
#else /* !HAVE_QSORT_BSEARCH */
|
166
|
+
int i;
|
167
|
+
struct common_field *cf = common_http_fields;
|
168
|
+
for(i = 0; i < ARRAY_SIZE(common_http_fields); i++, cf++) {
|
169
|
+
if (cf->len == flen && !memcmp(cf->name, field, flen))
|
170
|
+
return cf->value;
|
171
|
+
}
|
172
|
+
return Qnil;
|
173
|
+
#endif /* !HAVE_QSORT_BSEARCH */
|
174
|
+
}
|
175
|
+
|
176
|
+
static void http_field(void *data, const char *field,
|
177
|
+
size_t flen, const char *value, size_t vlen)
|
178
|
+
{
|
179
|
+
VALUE req = (VALUE)data;
|
180
|
+
VALUE v = Qnil;
|
181
|
+
VALUE f = Qnil;
|
182
|
+
|
183
|
+
VALIDATE_MAX_LENGTH(flen, FIELD_NAME);
|
184
|
+
VALIDATE_MAX_LENGTH(vlen, FIELD_VALUE);
|
185
|
+
|
186
|
+
v = rb_str_new(value, vlen);
|
187
|
+
|
188
|
+
f = find_common_field_value(field, flen);
|
189
|
+
|
190
|
+
if (f == Qnil) {
|
191
|
+
/*
|
192
|
+
* We got a strange header that we don't have a memoized value for.
|
193
|
+
* Fallback to creating a new string to use as a hash key.
|
194
|
+
*
|
195
|
+
* using rb_str_new(NULL, len) here is faster than rb_str_buf_new(len)
|
196
|
+
* in my testing, because: there's no minimum allocation length (and
|
197
|
+
* no check for it, either), RSTRING_LEN(f) does not need to be
|
198
|
+
* written twice, and and RSTRING_PTR(f) will already be
|
199
|
+
* null-terminated for us.
|
200
|
+
*/
|
201
|
+
f = rb_str_new(NULL, HTTP_PREFIX_LEN + flen);
|
202
|
+
memcpy(RSTRING_PTR(f), HTTP_PREFIX, HTTP_PREFIX_LEN);
|
203
|
+
memcpy(RSTRING_PTR(f) + HTTP_PREFIX_LEN, field, flen);
|
204
|
+
assert(*(RSTRING_PTR(f) + RSTRING_LEN(f)) == '\0'); /* paranoia */
|
205
|
+
/* fprintf(stderr, "UNKNOWN HEADER <%s>\n", RSTRING_PTR(f)); */
|
206
|
+
}
|
207
|
+
|
208
|
+
rb_hash_aset(req, f, v);
|
209
|
+
}
|
210
|
+
|
211
|
+
static void request_method(void *data, const char *at, size_t length)
|
212
|
+
{
|
213
|
+
VALUE req = (VALUE)data;
|
214
|
+
VALUE val = Qnil;
|
215
|
+
|
216
|
+
val = rb_str_new(at, length);
|
217
|
+
rb_hash_aset(req, global_request_method, val);
|
218
|
+
}
|
219
|
+
|
220
|
+
static void request_uri(void *data, const char *at, size_t length)
|
221
|
+
{
|
222
|
+
VALUE req = (VALUE)data;
|
223
|
+
VALUE val = Qnil;
|
224
|
+
|
225
|
+
VALIDATE_MAX_LENGTH(length, REQUEST_URI);
|
226
|
+
|
227
|
+
val = rb_str_new(at, length);
|
228
|
+
rb_hash_aset(req, global_request_uri, val);
|
229
|
+
}
|
230
|
+
|
231
|
+
static void fragment(void *data, const char *at, size_t length)
|
232
|
+
{
|
233
|
+
VALUE req = (VALUE)data;
|
234
|
+
VALUE val = Qnil;
|
235
|
+
|
236
|
+
VALIDATE_MAX_LENGTH(length, FRAGMENT);
|
237
|
+
|
238
|
+
val = rb_str_new(at, length);
|
239
|
+
rb_hash_aset(req, global_fragment, val);
|
240
|
+
}
|
241
|
+
|
242
|
+
static void request_path(void *data, const char *at, size_t length)
|
243
|
+
{
|
244
|
+
VALUE req = (VALUE)data;
|
245
|
+
VALUE val = Qnil;
|
246
|
+
|
247
|
+
VALIDATE_MAX_LENGTH(length, REQUEST_PATH);
|
248
|
+
|
249
|
+
val = rb_str_new(at, length);
|
250
|
+
rb_hash_aset(req, global_request_path, val);
|
251
|
+
}
|
252
|
+
|
253
|
+
static void query_string(void *data, const char *at, size_t length)
|
254
|
+
{
|
255
|
+
VALUE req = (VALUE)data;
|
256
|
+
VALUE val = Qnil;
|
257
|
+
|
258
|
+
VALIDATE_MAX_LENGTH(length, QUERY_STRING);
|
259
|
+
|
260
|
+
val = rb_str_new(at, length);
|
261
|
+
rb_hash_aset(req, global_query_string, val);
|
262
|
+
}
|
263
|
+
|
264
|
+
static void http_version(void *data, const char *at, size_t length)
|
265
|
+
{
|
266
|
+
VALUE req = (VALUE)data;
|
267
|
+
VALUE val = rb_str_new(at, length);
|
268
|
+
rb_hash_aset(req, global_http_version, val);
|
269
|
+
}
|
270
|
+
|
271
|
+
/** Finalizes the request header to have a bunch of stuff that's
|
272
|
+
needed. */
|
273
|
+
|
274
|
+
static void header_done(void *data, const char *at, size_t length)
|
275
|
+
{
|
276
|
+
VALUE req = (VALUE)data;
|
277
|
+
VALUE temp = Qnil;
|
278
|
+
VALUE ctype = Qnil;
|
279
|
+
VALUE clen = Qnil;
|
280
|
+
char *colon = NULL;
|
281
|
+
|
282
|
+
clen = rb_hash_aref(req, global_http_content_length);
|
283
|
+
if(clen != Qnil) {
|
284
|
+
rb_hash_aset(req, global_content_length, clen);
|
285
|
+
}
|
286
|
+
|
287
|
+
ctype = rb_hash_aref(req, global_http_content_type);
|
288
|
+
if(ctype != Qnil) {
|
289
|
+
rb_hash_aset(req, global_content_type, ctype);
|
290
|
+
}
|
291
|
+
|
292
|
+
rb_hash_aset(req, global_gateway_interface, global_gateway_interface_value);
|
293
|
+
if((temp = rb_hash_aref(req, global_http_host)) != Qnil) {
|
294
|
+
colon = memchr(RSTRING_PTR(temp), ':', RSTRING_LEN(temp));
|
295
|
+
if(colon != NULL) {
|
296
|
+
rb_hash_aset(req, global_server_name, rb_str_substr(temp, 0, colon - RSTRING_PTR(temp)));
|
297
|
+
rb_hash_aset(req, global_server_port,
|
298
|
+
rb_str_substr(temp, colon - RSTRING_PTR(temp)+1,
|
299
|
+
RSTRING_LEN(temp)));
|
300
|
+
} else {
|
301
|
+
rb_hash_aset(req, global_server_name, temp);
|
302
|
+
rb_hash_aset(req, global_server_port, global_port_80);
|
303
|
+
}
|
304
|
+
} else {
|
305
|
+
rb_hash_aset(req, global_server_name, global_localhost);
|
306
|
+
rb_hash_aset(req, global_server_port, global_port_80);
|
307
|
+
}
|
308
|
+
|
309
|
+
/* grab the initial body and stuff it into the hash */
|
310
|
+
rb_hash_aset(req, global_http_body, rb_str_new(at, length));
|
311
|
+
rb_hash_aset(req, global_server_protocol, global_server_protocol_value);
|
312
|
+
}
|
313
|
+
|
314
|
+
static void HttpParser_free(void *data) {
|
315
|
+
TRACE();
|
316
|
+
|
317
|
+
if(data) {
|
318
|
+
free(data);
|
319
|
+
}
|
320
|
+
}
|
321
|
+
|
322
|
+
|
323
|
+
static VALUE HttpParser_alloc(VALUE klass)
|
324
|
+
{
|
325
|
+
VALUE obj;
|
326
|
+
http_parser *hp = ALLOC_N(http_parser, 1);
|
327
|
+
TRACE();
|
328
|
+
hp->http_field = http_field;
|
329
|
+
hp->request_method = request_method;
|
330
|
+
hp->request_uri = request_uri;
|
331
|
+
hp->fragment = fragment;
|
332
|
+
hp->request_path = request_path;
|
333
|
+
hp->query_string = query_string;
|
334
|
+
hp->http_version = http_version;
|
335
|
+
hp->header_done = header_done;
|
336
|
+
http_parser_init(hp);
|
337
|
+
|
338
|
+
obj = Data_Wrap_Struct(klass, NULL, HttpParser_free, hp);
|
339
|
+
|
340
|
+
return obj;
|
341
|
+
}
|
342
|
+
|
343
|
+
|
344
|
+
/**
|
345
|
+
* call-seq:
|
346
|
+
* parser.new -> parser
|
347
|
+
*
|
348
|
+
* Creates a new parser.
|
349
|
+
*/
|
350
|
+
static VALUE HttpParser_init(VALUE self)
|
351
|
+
{
|
352
|
+
http_parser *http = NULL;
|
353
|
+
DATA_GET(self, http_parser, http);
|
354
|
+
http_parser_init(http);
|
355
|
+
|
356
|
+
return self;
|
357
|
+
}
|
358
|
+
|
359
|
+
|
360
|
+
/**
|
361
|
+
* call-seq:
|
362
|
+
* parser.reset -> nil
|
363
|
+
*
|
364
|
+
* Resets the parser to it's initial state so that you can reuse it
|
365
|
+
* rather than making new ones.
|
366
|
+
*/
|
367
|
+
static VALUE HttpParser_reset(VALUE self)
|
368
|
+
{
|
369
|
+
http_parser *http = NULL;
|
370
|
+
DATA_GET(self, http_parser, http);
|
371
|
+
http_parser_init(http);
|
372
|
+
|
373
|
+
return Qnil;
|
374
|
+
}
|
375
|
+
|
376
|
+
|
377
|
+
/**
|
378
|
+
* call-seq:
|
379
|
+
* parser.finish -> true/false
|
380
|
+
*
|
381
|
+
* Finishes a parser early which could put in a "good" or bad state.
|
382
|
+
* You should call reset after finish it or bad things will happen.
|
383
|
+
*/
|
384
|
+
static VALUE HttpParser_finish(VALUE self)
|
385
|
+
{
|
386
|
+
http_parser *http = NULL;
|
387
|
+
DATA_GET(self, http_parser, http);
|
388
|
+
http_parser_finish(http);
|
389
|
+
|
390
|
+
return http_parser_is_finished(http) ? Qtrue : Qfalse;
|
391
|
+
}
|
392
|
+
|
393
|
+
|
394
|
+
/**
|
395
|
+
* call-seq:
|
396
|
+
* parser.execute(req_hash, data, start) -> Integer
|
397
|
+
*
|
398
|
+
* Takes a Hash and a String of data, parses the String of data filling in the Hash
|
399
|
+
* returning an Integer to indicate how much of the data has been read. No matter
|
400
|
+
* what the return value, you should call HttpParser#finished? and HttpParser#error?
|
401
|
+
* to figure out if it's done parsing or there was an error.
|
402
|
+
*
|
403
|
+
* This function now throws an exception when there is a parsing error. This makes
|
404
|
+
* the logic for working with the parser much easier. You can still test for an
|
405
|
+
* error, but now you need to wrap the parser with an exception handling block.
|
406
|
+
*
|
407
|
+
* The third argument allows for parsing a partial request and then continuing
|
408
|
+
* the parsing from that position. It needs all of the original data as well
|
409
|
+
* so you have to append to the data buffer as you read.
|
410
|
+
*/
|
411
|
+
static VALUE HttpParser_execute(VALUE self, VALUE req_hash,
|
412
|
+
VALUE data, VALUE start)
|
413
|
+
{
|
414
|
+
http_parser *http = NULL;
|
415
|
+
int from = 0;
|
416
|
+
char *dptr = NULL;
|
417
|
+
long dlen = 0;
|
418
|
+
|
419
|
+
DATA_GET(self, http_parser, http);
|
420
|
+
|
421
|
+
from = FIX2INT(start);
|
422
|
+
dptr = RSTRING_PTR(data);
|
423
|
+
dlen = RSTRING_LEN(data);
|
424
|
+
|
425
|
+
if(from >= dlen) {
|
426
|
+
rb_raise(eHttpParserError, "Requested start is after data buffer end.");
|
427
|
+
} else {
|
428
|
+
http->data = (void *)req_hash;
|
429
|
+
http_parser_execute(http, dptr, dlen, from);
|
430
|
+
|
431
|
+
VALIDATE_MAX_LENGTH(http_parser_nread(http), HEADER);
|
432
|
+
|
433
|
+
if(http_parser_has_error(http)) {
|
434
|
+
rb_raise(eHttpParserError, "Invalid HTTP format, parsing fails.");
|
435
|
+
} else {
|
436
|
+
return INT2FIX(http_parser_nread(http));
|
437
|
+
}
|
438
|
+
}
|
439
|
+
}
|
440
|
+
|
441
|
+
|
442
|
+
|
443
|
+
/**
|
444
|
+
* call-seq:
|
445
|
+
* parser.error? -> true/false
|
446
|
+
*
|
447
|
+
* Tells you whether the parser is in an error state.
|
448
|
+
*/
|
449
|
+
static VALUE HttpParser_has_error(VALUE self)
|
450
|
+
{
|
451
|
+
http_parser *http = NULL;
|
452
|
+
DATA_GET(self, http_parser, http);
|
453
|
+
|
454
|
+
return http_parser_has_error(http) ? Qtrue : Qfalse;
|
455
|
+
}
|
456
|
+
|
457
|
+
|
458
|
+
/**
|
459
|
+
* call-seq:
|
460
|
+
* parser.finished? -> true/false
|
461
|
+
*
|
462
|
+
* Tells you whether the parser is finished or not and in a good state.
|
463
|
+
*/
|
464
|
+
static VALUE HttpParser_is_finished(VALUE self)
|
465
|
+
{
|
466
|
+
http_parser *http = NULL;
|
467
|
+
DATA_GET(self, http_parser, http);
|
468
|
+
|
469
|
+
return http_parser_is_finished(http) ? Qtrue : Qfalse;
|
470
|
+
}
|
471
|
+
|
472
|
+
|
473
|
+
/**
|
474
|
+
* call-seq:
|
475
|
+
* parser.nread -> Integer
|
476
|
+
*
|
477
|
+
* Returns the amount of data processed so far during this processing cycle. It is
|
478
|
+
* set to 0 on initialize or reset calls and is incremented each time execute is called.
|
479
|
+
*/
|
480
|
+
static VALUE HttpParser_nread(VALUE self)
|
481
|
+
{
|
482
|
+
http_parser *http = NULL;
|
483
|
+
DATA_GET(self, http_parser, http);
|
484
|
+
|
485
|
+
return INT2FIX(http->nread);
|
486
|
+
}
|
487
|
+
|
488
|
+
void Init_http11()
|
489
|
+
{
|
490
|
+
|
491
|
+
mUnicorn = rb_define_module("Unicorn");
|
492
|
+
|
493
|
+
DEF_GLOBAL(request_method, "REQUEST_METHOD");
|
494
|
+
DEF_GLOBAL(request_uri, "REQUEST_URI");
|
495
|
+
DEF_GLOBAL(fragment, "FRAGMENT");
|
496
|
+
DEF_GLOBAL(query_string, "QUERY_STRING");
|
497
|
+
DEF_GLOBAL(http_version, "HTTP_VERSION");
|
498
|
+
DEF_GLOBAL(request_path, "REQUEST_PATH");
|
499
|
+
DEF_GLOBAL(content_length, "CONTENT_LENGTH");
|
500
|
+
DEF_GLOBAL(http_content_length, "HTTP_CONTENT_LENGTH");
|
501
|
+
DEF_GLOBAL(http_body, "HTTP_BODY");
|
502
|
+
DEF_GLOBAL(content_type, "CONTENT_TYPE");
|
503
|
+
DEF_GLOBAL(http_content_type, "HTTP_CONTENT_TYPE");
|
504
|
+
DEF_GLOBAL(gateway_interface, "GATEWAY_INTERFACE");
|
505
|
+
DEF_GLOBAL(gateway_interface_value, "CGI/1.2");
|
506
|
+
DEF_GLOBAL(server_name, "SERVER_NAME");
|
507
|
+
DEF_GLOBAL(server_port, "SERVER_PORT");
|
508
|
+
DEF_GLOBAL(server_protocol, "SERVER_PROTOCOL");
|
509
|
+
DEF_GLOBAL(server_protocol_value, "HTTP/1.1");
|
510
|
+
DEF_GLOBAL(http_host, "HTTP_HOST");
|
511
|
+
DEF_GLOBAL(port_80, "80");
|
512
|
+
DEF_GLOBAL(localhost, "localhost");
|
513
|
+
|
514
|
+
eHttpParserError = rb_define_class_under(mUnicorn, "HttpParserError", rb_eIOError);
|
515
|
+
|
516
|
+
cHttpParser = rb_define_class_under(mUnicorn, "HttpParser", rb_cObject);
|
517
|
+
rb_define_alloc_func(cHttpParser, HttpParser_alloc);
|
518
|
+
rb_define_method(cHttpParser, "initialize", HttpParser_init,0);
|
519
|
+
rb_define_method(cHttpParser, "reset", HttpParser_reset,0);
|
520
|
+
rb_define_method(cHttpParser, "finish", HttpParser_finish,0);
|
521
|
+
rb_define_method(cHttpParser, "execute", HttpParser_execute,3);
|
522
|
+
rb_define_method(cHttpParser, "error?", HttpParser_has_error,0);
|
523
|
+
rb_define_method(cHttpParser, "finished?", HttpParser_is_finished,0);
|
524
|
+
rb_define_method(cHttpParser, "nread", HttpParser_nread,0);
|
525
|
+
init_common_fields();
|
526
|
+
}
|