Class: Curl::Multi

Inherits:
Object
  • Object
show all
Defined in:
lib/curl/multi.rb,
ext/curb_multi.c

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

.Curl::Multi.default_timeout( = 4) ⇒ 4

Get the global default time out for all Curl::Multi Handles.

Returns:

  • (4)

132
133
134
# File 'ext/curb_multi.c', line 132

VALUE ruby_curl_multi_get_default_timeout(VALUE klass) {
  return LONG2NUM(cCurlMutiDefaulttimeout);
}

.Curl::Multi.default_timeout( = 4) ⇒ 4

Set the global default time out for all Curl::Multi Handles. This value is used when libcurl cannot determine a timeout value when calling curl_multi_timeout.

Returns:

  • (4)

120
121
122
123
# File 'ext/curb_multi.c', line 120

VALUE ruby_curl_multi_set_default_timeout(VALUE klass, VALUE timeout) {
  cCurlMutiDefaulttimeout = NUM2LONG(timeout);
  return timeout;
}

.download(urls, easy_options = {}, multi_options = {}, download_paths = nil, &blk) ⇒ Object

call-seq:

Curl::Multi.download([‘http://example.com/p/a/t/h/file1.txt’,‘http://example.com/p/a/t/h/file2.txt’]){|c|}

will create 2 new files file1.txt and file2.txt

2 files will be opened, and remain open until the call completes

when using the :post or :put method, urls should be a hash, including the individual post fields per post


183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
# File 'lib/curl/multi.rb', line 183

def download(urls,easy_options={},multi_options={},download_paths=nil,&blk)
  errors = []
  procs = []
  files = []
  urls_with_config = []
  url_to_download_paths = {}

  urls.each_with_index do|urlcfg,i|
    if urlcfg.is_a?(Hash)
      url = url[:url]
    else
      url = urlcfg
    end

    if download_paths and download_paths[i]
      download_path = download_paths[i]
    else
      download_path = File.basename(url)
    end

    file = lambda do|dp|
      file = File.open(dp,"wb")
      procs << (lambda {|data| file.write data; data.size })
      files << file
      file
    end.call(download_path)

    if urlcfg.is_a?(Hash)
      urls_with_config << urlcfg.merge({:on_body => procs.last}.merge(easy_options))
    else
      urls_with_config << {:url => url, :on_body => procs.last, :method => :get}.merge(easy_options)
    end
    url_to_download_paths[url] = {:path => download_path, :file => file} # store for later
  end

  if blk
    # when injecting the block, ensure file is closed before yielding
    Curl::Multi.http(urls_with_config, multi_options) do |c,code,method|
      info = url_to_download_paths[c.url]
      begin
        file = info[:file]
        files.reject!{|f| f == file }
        file.close
      rescue => e
        errors << e
      end
      blk.call(c,info[:path])
    end
  else
    Curl::Multi.http(urls_with_config, multi_options)
  end

ensure
  files.each {|f|
    begin
      f.close
    rescue => e
      errors << e
    end
  }
  raise errors unless errors.empty?
end

.get(urls, easy_options = {}, multi_options = {}, &blk) ⇒ Object

call-seq:

Curl::Multi.get(['url1','url2','url3','url4','url5'], :follow_location => true) do|easy|
  easy
end

Blocking call to fetch multiple url's in parallel.


11
12
13
14
15
16
17
# File 'lib/curl/multi.rb', line 11

def get(urls, easy_options={}, multi_options={}, &blk)
  url_confs = []
  urls.each do|url|
    url_confs << {:url => url, :method => :get}.merge(easy_options)
  end
  self.http(url_confs, multi_options) {|c,code,method| blk.call(c) if blk }
end

.http(urls_with_config, multi_options = {}, &blk) ⇒ Object

call-seq:

Curl::Multi.http( [

{ :url => 'url1', :method => :post,
  :post_fields => {'field1' => 'value1', 'field2' => 'value2'} },
{ :url => 'url2', :method => :get,
  :follow_location => true, :max_redirects => 3 },
{ :url => 'url3', :method => :put, :put_data => File.open('file.txt','rb') },
{ :url => 'url4', :method => :head }

], => Curl::CURLPIPE_HTTP1)

Blocking call to issue multiple HTTP requests with varying verb's.

urls_with_config: is a hash of url's pointing to the easy handle options as well as the special option :method, that can by one of [:get, :post, :put, :delete, :head], when no verb is provided e.g. :method => nil -> GET is used multi_options: options for the multi handle blk: a callback, that yeilds when a handle is completed


85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
# File 'lib/curl/multi.rb', line 85

def http(urls_with_config, multi_options={}, &blk)
  m = Curl::Multi.new

  # maintain a sane number of easy handles
  multi_options[:max_connects] = max_connects = multi_options.key?(:max_connects) ? multi_options[:max_connects] : 10

  free_handles = [] # keep a list of free easy handles

  # configure the multi handle
  multi_options.each { |k,v| m.send("#{k}=", v) }
  callbacks = [:on_progress,:on_debug,:on_failure,:on_success,:on_redirect,:on_body,:on_header]

  add_free_handle = proc do|conf, easy|
    c       = conf.dup # avoid being destructive to input
    url     = c.delete(:url)
    method  = c.delete(:method)
    headers = c.delete(:headers)

    easy    = Curl::Easy.new if easy.nil?

    easy.url = url

    # assign callbacks
    callbacks.each do |cb|
      cbproc = c.delete(cb)
      easy.send(cb,&cbproc) if cbproc
    end

    case method
    when :post
      fields = c.delete(:post_fields)
      # set the post post using the url fields
      easy.post_body = fields.map{|f,k| "#{easy.escape(f)}=#{easy.escape(k)}"}.join('&')
    when :put
      easy.put_data = c.delete(:put_data)
    when :head
      easy.head = true
    when :delete
      easy.delete = true
    when :get
    else
      # XXX: nil is treated like a GET
    end

    # headers is a special key
    headers.each {|k,v| easy.headers[k] = v } if headers
 
    #
    # use the remaining options as specific configuration to the easy handle
    # bad options should raise an undefined method error
    #
    c.each { |k,v| easy.send("#{k}=",v) }

    easy.on_complete {|curl|
      free_handles << curl
      blk.call(curl,curl.response_code,method) if blk
    }
    m.add(easy)
  end

  max_connects.times do
    conf = urls_with_config.pop
    add_free_handle.call conf, nil
    break if urls_with_config.empty?
  end

  consume_free_handles = proc do
    # as we idle consume free handles
    if urls_with_config.size > 0 && free_handles.size > 0
      easy = free_handles.pop
      conf = urls_with_config.pop
      add_free_handle.call conf, easy
    end
  end

  if urls_with_config.empty?
    m.perform
  else
    until urls_with_config.empty?
      m.perform do
        consume_free_handles.call
      end
      consume_free_handles.call
    end
    free_handles = nil
  end
end

.Curl::Multi.new#<Curl::Easy...>

Create a new Curl::Multi instance

Returns:


92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
# File 'ext/curb_multi.c', line 92

VALUE ruby_curl_multi_new(VALUE klass) {
  VALUE new_curlm;

  ruby_curl_multi *rbcm = ALLOC(ruby_curl_multi);

  rbcm->handle = curl_multi_init();
  if (!rbcm->handle) {
    rb_raise(mCurlErrFailedInit, "Failed to initialize multi handle");
  }

  rbcm->requests = rb_hash_new();

  rbcm->active = 0;
  rbcm->running = 0;

  new_curlm = Data_Wrap_Struct(klass, curl_multi_mark, curl_multi_free, rbcm);

  return new_curlm;
}

.post(urls_with_config, easy_options = {}, multi_options = {}, &blk) ⇒ Object

call-seq:

Curl::Multi.post([{:url => 'url1', :post_fields => {'field1' => 'value1', 'field2' => 'value2'}},
                  {:url => 'url2', :post_fields => {'field1' => 'value1', 'field2' => 'value2'}},
                  {:url => 'url3', :post_fields => {'field1' => 'value1', 'field2' => 'value2'}}],
                 { :follow_location => true, :multipart_form_post => true },
                 {:pipeline => Curl::CURLPIPE_HTTP1}) do|easy|
  easy_handle_on_request_complete
end

Blocking call to POST multiple form's in parallel.

urls_with_config: is a hash of url's pointing to the postfields to send easy_options: are a set of common options to set on all easy handles multi_options: options to set on the Curl::Multi handle


35
36
37
38
39
40
41
# File 'lib/curl/multi.rb', line 35

def post(urls_with_config, easy_options={}, multi_options={}, &blk)
  url_confs = []
  urls_with_config.each do|uconf|
    url_confs << uconf.merge(:method => :post).merge(easy_options)
  end
  self.http(url_confs, multi_options) {|c,code,method| blk.call(c) }
end

.put(urls_with_config, easy_options = {}, multi_options = {}, &blk) ⇒ Object

call-seq:

Curl::Multi.put([{:url => 'url1', :put_data => "some message"},
                 {:url => 'url2', :put_data => IO.read('filepath')},
                 {:url => 'url3', :put_data => "maybe another string or socket?"],
                 {:follow_location => true},
                 {:pipeline => Curl::CURLPIPE_HTTP1}) do|easy|
  easy_handle_on_request_complete
end

Blocking call to POST multiple form's in parallel.

urls_with_config: is a hash of url's pointing to the postfields to send easy_options: are a set of common options to set on all easy handles multi_options: options to set on the Curl::Multi handle


59
60
61
62
63
64
65
# File 'lib/curl/multi.rb', line 59

def put(urls_with_config, easy_options={}, multi_options={}, &blk)
  url_confs = []
  urls_with_config.each do|uconf|
    url_confs << uconf.merge(:method => :put).merge(easy_options)
  end
  self.http(url_confs, multi_options) {|c,code,method| blk.call(c) }
end

Instance Method Details

#add(easy) ⇒ Object

multi = Curl::Multi.new easy = Curl::Easy.new('url')

multi.add(easy)

Add an easy handle to the multi stack


242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
# File 'ext/curb_multi.c', line 242

VALUE ruby_curl_multi_add(VALUE self, VALUE easy) {
  CURLMcode mcode;
  VALUE r;
  ruby_curl_easy *rbce;
  ruby_curl_multi *rbcm;

  Data_Get_Struct(self, ruby_curl_multi, rbcm);
  Data_Get_Struct(easy, ruby_curl_easy, rbce);

  // check if this curl handle has been added before adding again
  r = rb_hash_aref(rbcm->requests, LONG2NUM((long)rbce->curl));
  if ( r != Qnil ) {
    return Qnil;
  }

  /* setup the easy handle */
  ruby_curl_easy_setup( rbce );

  mcode = curl_multi_add_handle(rbcm->handle, rbce->curl);
  if (mcode != CURLM_CALL_MULTI_PERFORM && mcode != CURLM_OK) {
    raise_curl_multi_error_exception(mcode);
  }

  rbcm->active++;

  /* Increase the running count, so that the perform loop keeps running.
   * If this number is not correct, the next call to curl_multi_perform will correct it. */
  rbcm->running++;

  /* track a reference to associated multi handle */
  rbce->multi = self;

  rb_hash_aset( rbcm->requests, LONG2NUM((long)rbce->curl), easy );

  return self;
}

#cancel!Object

Cancels all requests currently being made on this Curl::Multi handle.


344
345
346
347
348
349
350
351
352
353
# File 'ext/curb_multi.c', line 344

static VALUE ruby_curl_multi_cancel(VALUE self) {
  ruby_curl_multi *rbcm;

  Data_Get_Struct(self, ruby_curl_multi, rbcm);
  
  rb_hash_foreach( rbcm->requests, ruby_curl_multi_cancel_callback, (VALUE)rbcm );
  
  /* for chaining */
  return self;
}

#idle?Boolean

Returns whether or not this Curl::Multi handle is processing any requests. E.g. this returns true when multi.requests.length == 0.

Returns:

  • (Boolean)

Returns:

  • (Boolean)

170
171
172
173
174
175
176
177
178
179
180
# File 'ext/curb_multi.c', line 170

static VALUE ruby_curl_multi_idle(VALUE self) {
  ruby_curl_multi *rbcm;
  
  Data_Get_Struct(self, ruby_curl_multi, rbcm);
  
  if (RHASH_SIZE(rbcm->requests) == 0) {
    return Qtrue;
  } else {
    return Qfalse;
  }
}

#max_connects=(count) ⇒ Object

multi = Curl::Multi.new multi.max_connects = 800

Set the max connections in the cache for a multi handle


189
190
191
192
193
194
195
196
197
198
199
# File 'ext/curb_multi.c', line 189

static VALUE ruby_curl_multi_max_connects(VALUE self, VALUE count) {
#ifdef HAVE_CURLMOPT_MAXCONNECTS
  ruby_curl_multi *rbcm;

  Data_Get_Struct(self, ruby_curl_multi, rbcm);

  curl_multi_setopt(rbcm->handle, CURLMOPT_MAXCONNECTS, NUM2LONG(count));
#endif

  return count;
}

#perform(*args) ⇒ Object

multi = Curl::Multi.new easy1 = Curl::Easy.new('url') easy2 = Curl::Easy.new('url')

multi.add(easy1) multi.add(easy2)

multi.perform do

# while idle other code my execute here

end

Run multi handles, looping selecting when data can be transfered


534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
# File 'ext/curb_multi.c', line 534

VALUE ruby_curl_multi_perform(int argc, VALUE *argv, VALUE self) {
  CURLMcode mcode;
  ruby_curl_multi *rbcm;
  int maxfd, rc = -1;
  fd_set fdread, fdwrite, fdexcep;
#ifdef _WIN32
  fd_set crt_fdread, crt_fdwrite, crt_fdexcep;
#endif
  long timeout_milliseconds;
  struct timeval tv = {0, 0};
  VALUE block = Qnil;
#if defined(HAVE_RB_THREAD_BLOCKING_REGION) || defined(HAVE_RB_THREAD_CALL_WITHOUT_GVL)
  struct _select_set fdset_args;
#endif

  rb_scan_args(argc, argv, "0&", &block);

  Data_Get_Struct(self, ruby_curl_multi, rbcm);

  timeout_milliseconds = cCurlMutiDefaulttimeout;

  rb_curl_multi_run( self, rbcm->handle, &(rbcm->running) );
  rb_curl_multi_read_info( self, rbcm->handle );
  if (block != Qnil) { rb_funcall(block, rb_intern("call"), 1, self);  }
 
  do {
    while (rbcm->running) {

#ifdef HAVE_CURL_MULTI_TIMEOUT
      /* get the curl suggested time out */
      mcode = curl_multi_timeout(rbcm->handle, &timeout_milliseconds);
      if (mcode != CURLM_OK) {
        raise_curl_multi_error_exception(mcode);
      }
#else
      /* libcurl doesn't have a timeout method defined, initialize to -1 we'll pick up the default later */
      timeout_milliseconds = -1;
#endif

      if (timeout_milliseconds == 0) { /* no delay */
        rb_curl_multi_run( self, rbcm->handle, &(rbcm->running) );
        rb_curl_multi_read_info( self, rbcm->handle );
        if (block != Qnil) { rb_funcall(block, rb_intern("call"), 1, self);  }
        continue;
      }

      if (timeout_milliseconds < 0 || timeout_milliseconds > cCurlMutiDefaulttimeout) {
        timeout_milliseconds = cCurlMutiDefaulttimeout; /* libcurl doesn't know how long to wait, use a default timeout */
                                                        /* or buggy versions libcurl sometimes reports huge timeouts... let's cap it */
      }

      tv.tv_sec  = 0; /* never wait longer than 1 second */
      tv.tv_usec = (int)(timeout_milliseconds * 1000); /* XXX: int is the right type for OSX, what about linux? */

      FD_ZERO(&fdread);
      FD_ZERO(&fdwrite);
      FD_ZERO(&fdexcep);

      /* load the fd sets from the multi handle */
      mcode = curl_multi_fdset(rbcm->handle, &fdread, &fdwrite, &fdexcep, &maxfd);
      if (mcode != CURLM_OK) {
        raise_curl_multi_error_exception(mcode);
      }

      if (maxfd == -1) {
        /* libcurl recommends sleeping for 100ms */
        rb_thread_wait_for(rb_time_timeval(DBL2NUM(0.1)));
        rb_curl_multi_run( self, rbcm->handle, &(rbcm->running) );
        rb_curl_multi_read_info( self, rbcm->handle );
        if (block != Qnil) { rb_funcall(block, rb_intern("call"), 1, self);  }
        continue;
      }

#ifdef _WIN32
      create_crt_fd(&fdread, &crt_fdread);
      create_crt_fd(&fdwrite, &crt_fdwrite);
      create_crt_fd(&fdexcep, &crt_fdexcep);
#endif

#if defined(HAVE_RB_THREAD_BLOCKING_REGION) || defined(HAVE_RB_THREAD_CALL_WITHOUT_GVL)
      fdset_args.maxfd = maxfd+1;
      fdset_args.fdread = &fdread;
      fdset_args.fdwrite = &fdwrite;
      fdset_args.fdexcep = &fdexcep;
      fdset_args.tv = &tv;
#ifdef HAVE_RB_THREAD_CALL_WITHOUT_GVL
      rc = (int)(VALUE) rb_thread_call_without_gvl((void *(*)(void *))curb_select, &fdset_args, RUBY_UBF_IO, 0);
#elif HAVE_RB_THREAD_BLOCKING_REGION
      rc = rb_thread_blocking_region(curb_select, &fdset_args, RUBY_UBF_IO, 0);
#elif HAVE_RB_THREAD_FD_SELECT
      rc = rb_thread_fd_select(maxfd+1, &fdread, &fdwrite, &fdexcep, &tv);
#else
      rc = rb_thread_select(maxfd+1, &fdread, &fdwrite, &fdexcep, &tv);
#endif

#endif

#ifdef _WIN32
      cleanup_crt_fd(&fdread, &crt_fdread);
      cleanup_crt_fd(&fdwrite, &crt_fdwrite);
      cleanup_crt_fd(&fdexcep, &crt_fdexcep);
#endif

      switch(rc) {
      case -1:
        if(errno != EINTR) {
          rb_raise(rb_eRuntimeError, "select(): %s", strerror(errno));
          break;
        }
      case 0: /* timeout */
      default: /* action */
        rb_curl_multi_run( self, rbcm->handle, &(rbcm->running) );
        rb_curl_multi_read_info( self, rbcm->handle );
        if (block != Qnil) { rb_funcall(block, rb_intern("call"), 1, self);  }
        break;
      }
    }

  } while( rbcm->running );

  rb_curl_multi_read_info( self, rbcm->handle );
  if (block != Qnil) { rb_funcall(block, rb_intern("call"), 1, self);  }
    
  return Qtrue;
}

#pipeline=(method) ⇒ Object

multi = Curl::Multi.new multi.pipeline = true

Pass a long set to 1 for HTTP/1.1 pipelining, 2 for HTTP/2 multiplexing, or 0 to disable.

Enabling pipelining on a multi handle will make it attempt to perform HTTP Pipelining as

far as possible for transfers using this handle. This means that if you add a second request that can use an already existing connection, the second request will be “piped” on the same connection rather than being executed in parallel. (Added in 7.16.0, multiplex added in 7.43.0)


213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
# File 'ext/curb_multi.c', line 213

static VALUE ruby_curl_multi_pipeline(VALUE self, VALUE method) {
#ifdef HAVE_CURLMOPT_PIPELINING
  ruby_curl_multi *rbcm;

  long value;

  if (method == Qtrue) {
    value = 1;
  } else if (method == Qfalse) {
    value  = 0;
  } else {
    value = NUM2LONG(method);
  } 

  Data_Get_Struct(self, ruby_curl_multi, rbcm);
  curl_multi_setopt(rbcm->handle, CURLMOPT_PIPELINING, value);
#endif
  return method == Qtrue ? 1 : 0;
}

#remove(easy) ⇒ Object

multi = Curl::Multi.new easy = Curl::Easy.new('url')

multi.add(easy)

# sometime later multi.remove(easy)

Remove an easy handle from a multi stack.

Will raise an exception if the easy handle is not found


293
294
295
296
297
298
299
300
301
# File 'ext/curb_multi.c', line 293

VALUE ruby_curl_multi_remove(VALUE self, VALUE easy) {
  ruby_curl_multi *rbcm;

  Data_Get_Struct(self, ruby_curl_multi, rbcm);

  rb_curl_multi_remove(rbcm,easy);

  return self;
}

#requestsArray

Returns an array containing all the active requests on this Curl::Multi object.

Returns:

  • (Array)

149
150
151
152
153
154
155
156
157
158
159
160
161
# File 'ext/curb_multi.c', line 149

static VALUE ruby_curl_multi_requests(VALUE self) {
  ruby_curl_multi *rbcm;
  VALUE result_array;

  Data_Get_Struct(self, ruby_curl_multi, rbcm);
  
  result_array = rb_ary_new();
  
  /* iterate over the requests hash, and stuff references into the array. */
  rb_hash_foreach(rbcm->requests, ruby_curl_multi_requests_callback, result_array);
  
  return result_array;
}