httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
/*
|
2013-01-15 18:22:17 +01:00
|
|
|
* Copyright (C) 2003-2013 The Music Player Daemon Project
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
* http://www.musicpd.org
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
* the Free Software Foundation; either version 2 of the License, or
|
|
|
|
* (at your option) any later version.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public License along
|
|
|
|
* with this program; if not, write to the Free Software Foundation, Inc.,
|
|
|
|
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
|
|
|
|
*/
|
|
|
|
|
2009-11-12 09:12:38 +01:00
|
|
|
#include "config.h"
|
2013-01-15 18:22:17 +01:00
|
|
|
#include "HttpdOutputPlugin.hxx"
|
|
|
|
#include "HttpdInternal.hxx"
|
|
|
|
#include "HttpdClient.hxx"
|
2013-07-30 08:34:10 +02:00
|
|
|
#include "OutputAPI.hxx"
|
2013-07-30 09:04:05 +02:00
|
|
|
#include "EncoderPlugin.hxx"
|
2013-04-17 22:22:37 +02:00
|
|
|
#include "EncoderList.hxx"
|
2013-09-04 18:02:09 +02:00
|
|
|
#include "system/Resolver.hxx"
|
2013-01-30 09:18:52 +01:00
|
|
|
#include "Page.hxx"
|
2013-01-30 09:08:50 +01:00
|
|
|
#include "IcyMetaDataServer.hxx"
|
2013-08-07 10:31:31 +02:00
|
|
|
#include "system/fd_util.h"
|
2014-01-04 18:21:36 +01:00
|
|
|
#include "IOThread.hxx"
|
2013-12-31 16:31:36 +01:00
|
|
|
#include "event/Call.hxx"
|
2013-08-10 18:02:44 +02:00
|
|
|
#include "util/Error.hxx"
|
|
|
|
#include "util/Domain.hxx"
|
2013-09-27 22:31:24 +02:00
|
|
|
#include "Log.hxx"
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
|
|
|
#include <assert.h>
|
|
|
|
|
2009-07-06 14:40:06 +02:00
|
|
|
#include <sys/types.h>
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
#include <unistd.h>
|
2013-07-30 20:11:57 +02:00
|
|
|
#include <string.h>
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
#include <errno.h>
|
|
|
|
|
2010-06-06 21:53:24 +02:00
|
|
|
#ifdef HAVE_LIBWRAP
|
2011-03-18 19:44:12 +01:00
|
|
|
#include <sys/socket.h> /* needed for AF_UNIX */
|
2010-06-06 21:53:24 +02:00
|
|
|
#include <tcpd.h>
|
|
|
|
#endif
|
|
|
|
|
2013-09-27 22:31:24 +02:00
|
|
|
const Domain httpd_output_domain("httpd_output");
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-01-30 13:29:21 +01:00
|
|
|
inline
|
|
|
|
HttpdOutput::HttpdOutput(EventLoop &_loop)
|
2013-12-31 16:31:36 +01:00
|
|
|
:ServerSocket(_loop), DeferredMonitor(_loop),
|
2013-01-30 13:20:27 +01:00
|
|
|
encoder(nullptr), unflushed_input(0),
|
2013-01-30 13:29:21 +01:00
|
|
|
metadata(nullptr)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
|
|
|
HttpdOutput::~HttpdOutput()
|
|
|
|
{
|
|
|
|
if (metadata != nullptr)
|
|
|
|
metadata->Unref();
|
|
|
|
|
|
|
|
if (encoder != nullptr)
|
|
|
|
encoder_finish(encoder);
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
inline bool
|
2013-08-10 18:02:44 +02:00
|
|
|
HttpdOutput::Bind(Error &error)
|
2009-11-07 16:52:44 +01:00
|
|
|
{
|
2013-01-27 23:23:46 +01:00
|
|
|
open = false;
|
2009-11-07 16:52:44 +01:00
|
|
|
|
2014-01-04 19:29:37 +01:00
|
|
|
bool result = false;
|
|
|
|
BlockingCall(DeferredMonitor::GetEventLoop(), [this, &error, &result](){
|
|
|
|
result = ServerSocket::Open(error);
|
|
|
|
});
|
|
|
|
return result;
|
2009-11-07 16:52:44 +01:00
|
|
|
}
|
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
inline void
|
|
|
|
HttpdOutput::Unbind()
|
2009-11-07 16:52:44 +01:00
|
|
|
{
|
2013-01-27 23:23:46 +01:00
|
|
|
assert(!open);
|
2009-11-07 16:52:44 +01:00
|
|
|
|
2014-01-04 19:29:37 +01:00
|
|
|
BlockingCall(DeferredMonitor::GetEventLoop(), [this](){
|
|
|
|
ServerSocket::Close();
|
|
|
|
});
|
2009-11-07 16:52:44 +01:00
|
|
|
}
|
|
|
|
|
2013-01-30 13:29:21 +01:00
|
|
|
inline bool
|
2013-08-10 18:02:44 +02:00
|
|
|
HttpdOutput::Configure(const config_param ¶m, Error &error)
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
|
|
|
/* read configuration */
|
2013-08-04 12:25:08 +02:00
|
|
|
name = param.GetBlockValue("name", "Set name in config");
|
|
|
|
genre = param.GetBlockValue("genre", "Set genre in config");
|
|
|
|
website = param.GetBlockValue("website", "Set website in config");
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-10-02 12:20:36 +02:00
|
|
|
unsigned port = param.GetBlockValue("port", 8000u);
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2012-10-01 23:52:40 +02:00
|
|
|
const char *encoder_name =
|
2013-08-04 12:25:08 +02:00
|
|
|
param.GetBlockValue("encoder", "vorbis");
|
2013-07-30 09:04:05 +02:00
|
|
|
const auto encoder_plugin = encoder_plugin_get(encoder_name);
|
2013-10-28 23:58:17 +01:00
|
|
|
if (encoder_plugin == nullptr) {
|
2013-08-10 18:02:44 +02:00
|
|
|
error.Format(httpd_output_domain,
|
|
|
|
"No such encoder: %s", encoder_name);
|
2013-01-30 13:29:21 +01:00
|
|
|
return false;
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
}
|
|
|
|
|
2013-08-04 12:25:08 +02:00
|
|
|
clients_max = param.GetBlockValue("max_clients", 0u);
|
2009-10-29 22:38:18 +01:00
|
|
|
|
2010-09-25 15:00:43 +02:00
|
|
|
/* set up bind_to_address */
|
2010-10-05 20:37:18 +02:00
|
|
|
|
2013-08-04 12:25:08 +02:00
|
|
|
const char *bind_to_address = param.GetBlockValue("bind_to_address");
|
2013-10-28 23:58:17 +01:00
|
|
|
bool success = bind_to_address != nullptr &&
|
2010-10-05 20:37:18 +02:00
|
|
|
strcmp(bind_to_address, "any") != 0
|
2013-08-10 18:02:44 +02:00
|
|
|
? AddHost(bind_to_address, port, error)
|
|
|
|
: AddPort(port, error);
|
2013-01-30 13:29:21 +01:00
|
|
|
if (!success)
|
|
|
|
return false;
|
2009-04-13 19:35:02 +02:00
|
|
|
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
/* initialize encoder */
|
|
|
|
|
2013-08-10 18:02:44 +02:00
|
|
|
encoder = encoder_init(*encoder_plugin, param, error);
|
2013-01-30 13:29:21 +01:00
|
|
|
if (encoder == nullptr)
|
|
|
|
return false;
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2009-12-03 20:27:08 +01:00
|
|
|
/* determine content type */
|
2013-01-30 13:29:21 +01:00
|
|
|
content_type = encoder_get_mime_type(encoder);
|
|
|
|
if (content_type == nullptr)
|
|
|
|
content_type = "application/octet-stream";
|
|
|
|
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2013-12-31 16:55:26 +01:00
|
|
|
inline bool
|
|
|
|
HttpdOutput::Init(const config_param ¶m, Error &error)
|
|
|
|
{
|
|
|
|
return ao_base_init(&base, &httpd_output_plugin, param, error);
|
|
|
|
}
|
|
|
|
|
2013-01-30 13:29:21 +01:00
|
|
|
static struct audio_output *
|
2013-08-10 18:02:44 +02:00
|
|
|
httpd_output_init(const config_param ¶m, Error &error)
|
2013-01-30 13:29:21 +01:00
|
|
|
{
|
2014-01-04 18:21:36 +01:00
|
|
|
HttpdOutput *httpd = new HttpdOutput(io_thread_get());
|
2013-01-30 13:29:21 +01:00
|
|
|
|
2013-12-31 16:55:26 +01:00
|
|
|
audio_output *result = httpd->InitAndConfigure(param, error);
|
|
|
|
if (result == nullptr)
|
2013-01-30 13:29:21 +01:00
|
|
|
delete httpd;
|
2009-12-03 20:27:08 +01:00
|
|
|
|
2013-12-31 16:55:26 +01:00
|
|
|
return result;
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2011-09-16 23:31:48 +02:00
|
|
|
httpd_output_finish(struct audio_output *ao)
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
2013-12-31 16:59:18 +01:00
|
|
|
HttpdOutput *httpd = HttpdOutput::Cast(ao);
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-12-31 16:55:26 +01:00
|
|
|
httpd->Finish();
|
2013-01-15 18:22:17 +01:00
|
|
|
delete httpd;
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2013-01-15 18:22:17 +01:00
|
|
|
* Creates a new #HttpdClient object and adds it into the
|
2013-01-27 23:21:39 +01:00
|
|
|
* HttpdOutput.clients linked list.
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
*/
|
2013-01-27 23:23:46 +01:00
|
|
|
inline void
|
|
|
|
HttpdOutput::AddClient(int fd)
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
2013-12-31 16:31:36 +01:00
|
|
|
clients.emplace_front(*this, fd, ServerSocket::GetEventLoop(),
|
2013-07-30 09:04:05 +02:00
|
|
|
encoder->plugin.tag == nullptr);
|
2013-01-27 23:23:46 +01:00
|
|
|
++clients_cnt;
|
2009-04-13 19:35:02 +02:00
|
|
|
|
|
|
|
/* pass metadata to client */
|
2013-01-27 23:23:46 +01:00
|
|
|
if (metadata != nullptr)
|
|
|
|
clients.front().PushMetaData(metadata);
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
}
|
|
|
|
|
2013-12-31 16:31:36 +01:00
|
|
|
void
|
|
|
|
HttpdOutput::RunDeferred()
|
|
|
|
{
|
|
|
|
/* this method runs in the IOThread; it broadcasts pages from
|
|
|
|
our own queue to all clients */
|
|
|
|
|
|
|
|
const ScopeLock protect(mutex);
|
|
|
|
|
|
|
|
while (!pages.empty()) {
|
|
|
|
Page *page = pages.front();
|
|
|
|
pages.pop();
|
|
|
|
|
|
|
|
for (auto &client : clients)
|
|
|
|
client.PushPage(page);
|
|
|
|
|
|
|
|
page->Unref();
|
|
|
|
}
|
|
|
|
|
|
|
|
/* wake up the client that may be waiting for the queue to be
|
|
|
|
flushed */
|
|
|
|
cond.broadcast();
|
|
|
|
}
|
|
|
|
|
2013-01-30 13:20:27 +01:00
|
|
|
void
|
|
|
|
HttpdOutput::OnAccept(int fd, const sockaddr &address,
|
|
|
|
size_t address_length, gcc_unused int uid)
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
|
|
|
/* the listener socket has become readable - a client has
|
|
|
|
connected */
|
|
|
|
|
2010-06-06 21:53:24 +02:00
|
|
|
#ifdef HAVE_LIBWRAP
|
2013-01-30 13:20:27 +01:00
|
|
|
if (address.sa_family != AF_UNIX) {
|
2013-12-04 08:43:55 +01:00
|
|
|
const auto hostaddr = sockaddr_to_string(&address,
|
|
|
|
address_length);
|
2013-11-26 13:13:23 +01:00
|
|
|
// TODO: shall we obtain the program name from argv[0]?
|
|
|
|
const char *progname = "mpd";
|
2010-06-06 21:53:24 +02:00
|
|
|
|
|
|
|
struct request_info req;
|
|
|
|
request_init(&req, RQ_FILE, fd, RQ_DAEMON, progname, 0);
|
|
|
|
|
|
|
|
fromhost(&req);
|
|
|
|
|
|
|
|
if (!hosts_access(&req)) {
|
|
|
|
/* tcp wrappers says no */
|
2013-09-27 22:31:24 +02:00
|
|
|
FormatWarning(httpd_output_domain,
|
|
|
|
"libwrap refused connection (libwrap=%s) from %s",
|
2013-12-04 08:43:55 +01:00
|
|
|
progname, hostaddr.c_str());
|
2011-09-19 21:04:19 +02:00
|
|
|
close_socket(fd);
|
2010-10-05 20:37:18 +02:00
|
|
|
return;
|
2010-06-06 21:53:24 +02:00
|
|
|
}
|
|
|
|
}
|
2010-10-05 20:37:18 +02:00
|
|
|
#else
|
|
|
|
(void)address;
|
|
|
|
(void)address_length;
|
2010-06-06 21:53:24 +02:00
|
|
|
#endif /* HAVE_WRAP */
|
2010-10-05 20:37:18 +02:00
|
|
|
|
2013-01-30 13:20:27 +01:00
|
|
|
const ScopeLock protect(mutex);
|
2010-10-05 20:37:18 +02:00
|
|
|
|
2009-10-29 22:38:18 +01:00
|
|
|
if (fd >= 0) {
|
|
|
|
/* can we allow additional client */
|
2013-01-30 13:20:27 +01:00
|
|
|
if (open && (clients_max == 0 || clients_cnt < clients_max))
|
|
|
|
AddClient(fd);
|
2009-10-29 22:38:18 +01:00
|
|
|
else
|
2011-09-19 21:04:19 +02:00
|
|
|
close_socket(fd);
|
2009-10-29 22:38:18 +01:00
|
|
|
} else if (fd < 0 && errno != EINTR) {
|
2013-09-27 22:31:24 +02:00
|
|
|
LogErrno(httpd_output_domain, "accept() failed");
|
2009-10-29 22:38:18 +01:00
|
|
|
}
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
}
|
|
|
|
|
2013-01-30 09:18:52 +01:00
|
|
|
Page *
|
2013-01-27 23:23:46 +01:00
|
|
|
HttpdOutput::ReadPage()
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
2013-01-27 23:23:46 +01:00
|
|
|
if (unflushed_input >= 65536) {
|
2010-08-06 22:18:01 +02:00
|
|
|
/* we have fed a lot of input into the encoder, but it
|
|
|
|
didn't give anything back yet - flush now to avoid
|
|
|
|
buffer underruns */
|
2013-08-10 18:02:44 +02:00
|
|
|
encoder_flush(encoder, IgnoreError());
|
2013-01-27 23:23:46 +01:00
|
|
|
unflushed_input = 0;
|
2010-08-06 22:18:01 +02:00
|
|
|
}
|
|
|
|
|
2012-10-01 23:52:40 +02:00
|
|
|
size_t size = 0;
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
do {
|
2013-01-27 23:23:46 +01:00
|
|
|
size_t nbytes = encoder_read(encoder,
|
|
|
|
buffer + size,
|
|
|
|
sizeof(buffer) - size);
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
if (nbytes == 0)
|
|
|
|
break;
|
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
unflushed_input = 0;
|
2010-08-06 22:18:01 +02:00
|
|
|
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
size += nbytes;
|
2013-01-27 23:23:46 +01:00
|
|
|
} while (size < sizeof(buffer));
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
|
|
|
if (size == 0)
|
2013-10-28 23:58:17 +01:00
|
|
|
return nullptr;
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-01-30 09:18:52 +01:00
|
|
|
return Page::Copy(buffer, size);
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
static bool
|
2013-08-10 18:02:44 +02:00
|
|
|
httpd_output_enable(struct audio_output *ao, Error &error)
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
2013-12-31 16:59:18 +01:00
|
|
|
HttpdOutput *httpd = HttpdOutput::Cast(ao);
|
2013-01-27 23:23:46 +01:00
|
|
|
|
2013-08-10 18:02:44 +02:00
|
|
|
return httpd->Bind(error);
|
2013-01-27 23:23:46 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
httpd_output_disable(struct audio_output *ao)
|
|
|
|
{
|
2013-12-31 16:59:18 +01:00
|
|
|
HttpdOutput *httpd = HttpdOutput::Cast(ao);
|
2013-01-27 23:23:46 +01:00
|
|
|
|
|
|
|
httpd->Unbind();
|
|
|
|
}
|
|
|
|
|
|
|
|
inline bool
|
2013-08-10 18:02:44 +02:00
|
|
|
HttpdOutput::OpenEncoder(AudioFormat &audio_format, Error &error)
|
2013-01-27 23:23:46 +01:00
|
|
|
{
|
|
|
|
if (!encoder_open(encoder, audio_format, error))
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
return false;
|
|
|
|
|
|
|
|
/* we have to remember the encoder header, i.e. the first
|
|
|
|
bytes of encoder output after opening it, because it has to
|
|
|
|
be sent to every new client */
|
2013-01-27 23:23:46 +01:00
|
|
|
header = ReadPage();
|
2010-08-06 22:18:01 +02:00
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
unflushed_input = 0;
|
2010-08-06 22:18:01 +02:00
|
|
|
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
inline bool
|
2013-08-10 18:02:44 +02:00
|
|
|
HttpdOutput::Open(AudioFormat &audio_format, Error &error)
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
2013-01-27 23:23:46 +01:00
|
|
|
assert(!open);
|
|
|
|
assert(clients.empty());
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
/* open the encoder */
|
2009-11-05 23:47:29 +01:00
|
|
|
|
2013-08-10 18:02:44 +02:00
|
|
|
if (!OpenEncoder(audio_format, error))
|
2013-01-27 23:23:46 +01:00
|
|
|
return false;
|
|
|
|
|
|
|
|
/* initialize other attributes */
|
|
|
|
|
|
|
|
clients_cnt = 0;
|
2013-08-03 21:00:50 +02:00
|
|
|
timer = new Timer(audio_format);
|
2009-11-05 23:47:29 +01:00
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
open = true;
|
|
|
|
|
|
|
|
return true;
|
2009-11-05 23:47:29 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
static bool
|
2013-08-03 21:00:50 +02:00
|
|
|
httpd_output_open(struct audio_output *ao, AudioFormat &audio_format,
|
2013-08-10 18:02:44 +02:00
|
|
|
Error &error)
|
2009-11-05 23:47:29 +01:00
|
|
|
{
|
2013-12-31 16:59:18 +01:00
|
|
|
HttpdOutput *httpd = HttpdOutput::Cast(ao);
|
2013-01-15 18:22:17 +01:00
|
|
|
|
|
|
|
const ScopeLock protect(httpd->mutex);
|
2013-01-27 23:23:46 +01:00
|
|
|
return httpd->Open(audio_format, error);
|
|
|
|
}
|
2009-11-05 23:47:29 +01:00
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
inline void
|
|
|
|
HttpdOutput::Close()
|
|
|
|
{
|
|
|
|
assert(open);
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
open = false;
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-05-12 15:03:42 +02:00
|
|
|
delete timer;
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
clients.clear();
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-10-28 23:58:17 +01:00
|
|
|
if (header != nullptr)
|
2013-01-30 09:18:52 +01:00
|
|
|
header->Unref();
|
2009-11-05 23:47:29 +01:00
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
encoder_close(encoder);
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
}
|
|
|
|
|
2011-09-16 23:31:48 +02:00
|
|
|
static void
|
|
|
|
httpd_output_close(struct audio_output *ao)
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
2013-12-31 16:59:18 +01:00
|
|
|
HttpdOutput *httpd = HttpdOutput::Cast(ao);
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-01-15 18:22:17 +01:00
|
|
|
const ScopeLock protect(httpd->mutex);
|
2013-01-27 23:23:46 +01:00
|
|
|
httpd->Close();
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
void
|
2013-01-27 23:23:46 +01:00
|
|
|
HttpdOutput::RemoveClient(HttpdClient &client)
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
2013-01-27 23:23:46 +01:00
|
|
|
assert(clients_cnt > 0);
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
for (auto prev = clients.before_begin(), i = std::next(prev);;
|
2013-01-15 18:22:17 +01:00
|
|
|
prev = i, i = std::next(prev)) {
|
2013-01-27 23:23:46 +01:00
|
|
|
assert(i != clients.end());
|
|
|
|
if (&*i == &client) {
|
|
|
|
clients.erase_after(prev);
|
|
|
|
clients_cnt--;
|
2013-01-15 18:22:17 +01:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
void
|
2013-01-27 23:23:46 +01:00
|
|
|
HttpdOutput::SendHeader(HttpdClient &client) const
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
2013-10-28 23:58:17 +01:00
|
|
|
if (header != nullptr)
|
2013-01-27 23:23:46 +01:00
|
|
|
client.PushPage(header);
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
}
|
|
|
|
|
2013-12-31 17:01:08 +01:00
|
|
|
inline unsigned
|
|
|
|
HttpdOutput::Delay() const
|
2010-11-05 09:42:14 +01:00
|
|
|
{
|
2013-12-31 17:01:08 +01:00
|
|
|
if (!LockHasClients() && base.pause) {
|
2012-08-14 21:39:33 +02:00
|
|
|
/* if there's no client and this output is paused,
|
|
|
|
then httpd_output_pause() will not do anything, it
|
|
|
|
will not fill the buffer and it will not update the
|
|
|
|
timer; therefore, we reset the timer here */
|
2013-12-31 17:01:08 +01:00
|
|
|
timer->Reset();
|
2012-08-14 21:46:43 +02:00
|
|
|
|
|
|
|
/* some arbitrary delay that is long enough to avoid
|
|
|
|
consuming too much CPU, and short enough to notice
|
|
|
|
new clients quickly enough */
|
|
|
|
return 1000;
|
2012-08-14 21:39:33 +02:00
|
|
|
}
|
|
|
|
|
2013-12-31 17:01:08 +01:00
|
|
|
return timer->IsStarted()
|
|
|
|
? timer->GetDelay()
|
2010-11-05 09:42:14 +01:00
|
|
|
: 0;
|
|
|
|
}
|
|
|
|
|
2013-12-31 17:01:08 +01:00
|
|
|
static unsigned
|
|
|
|
httpd_output_delay(struct audio_output *ao)
|
|
|
|
{
|
|
|
|
HttpdOutput *httpd = HttpdOutput::Cast(ao);
|
|
|
|
|
|
|
|
return httpd->Delay();
|
|
|
|
}
|
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
void
|
2013-01-30 09:18:52 +01:00
|
|
|
HttpdOutput::BroadcastPage(Page *page)
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
2013-10-28 23:58:17 +01:00
|
|
|
assert(page != nullptr);
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-12-31 16:31:36 +01:00
|
|
|
mutex.lock();
|
|
|
|
pages.push(page);
|
|
|
|
page->Ref();
|
|
|
|
mutex.unlock();
|
|
|
|
|
|
|
|
DeferredMonitor::Schedule();
|
2009-05-05 22:43:28 +02:00
|
|
|
}
|
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
void
|
|
|
|
HttpdOutput::BroadcastFromEncoder()
|
2009-05-05 22:43:28 +02:00
|
|
|
{
|
2013-12-31 16:31:36 +01:00
|
|
|
/* synchronize with the IOThread */
|
|
|
|
mutex.lock();
|
|
|
|
while (!pages.empty())
|
|
|
|
cond.wait(mutex);
|
|
|
|
|
2013-01-30 09:18:52 +01:00
|
|
|
Page *page;
|
2013-12-31 16:31:36 +01:00
|
|
|
while ((page = ReadPage()) != nullptr)
|
|
|
|
pages.push(page);
|
|
|
|
|
|
|
|
mutex.unlock();
|
|
|
|
|
|
|
|
DeferredMonitor::Schedule();
|
2009-05-05 22:43:28 +02:00
|
|
|
}
|
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
inline bool
|
2013-08-10 18:02:44 +02:00
|
|
|
HttpdOutput::EncodeAndPlay(const void *chunk, size_t size, Error &error)
|
2009-05-05 22:43:28 +02:00
|
|
|
{
|
2013-08-10 18:02:44 +02:00
|
|
|
if (!encoder_write(encoder, chunk, size, error))
|
2009-05-05 22:43:28 +02:00
|
|
|
return false;
|
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
unflushed_input += size;
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
BroadcastFromEncoder();
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2013-12-31 17:01:08 +01:00
|
|
|
inline size_t
|
|
|
|
HttpdOutput::Play(const void *chunk, size_t size, Error &error)
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
2013-12-31 17:01:08 +01:00
|
|
|
if (LockHasClients()) {
|
|
|
|
if (!EncodeAndPlay(chunk, size, error))
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2013-12-31 17:01:08 +01:00
|
|
|
if (!timer->IsStarted())
|
|
|
|
timer->Start();
|
|
|
|
timer->Add(size);
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
|
|
|
|
return size;
|
|
|
|
}
|
|
|
|
|
2013-12-31 17:01:08 +01:00
|
|
|
static size_t
|
|
|
|
httpd_output_play(struct audio_output *ao, const void *chunk, size_t size,
|
|
|
|
Error &error)
|
|
|
|
{
|
|
|
|
HttpdOutput *httpd = HttpdOutput::Cast(ao);
|
|
|
|
|
|
|
|
return httpd->Play(chunk, size, error);
|
|
|
|
}
|
|
|
|
|
2010-08-06 07:23:17 +02:00
|
|
|
static bool
|
2011-09-16 23:31:48 +02:00
|
|
|
httpd_output_pause(struct audio_output *ao)
|
2010-08-06 07:23:17 +02:00
|
|
|
{
|
2013-12-31 16:59:18 +01:00
|
|
|
HttpdOutput *httpd = HttpdOutput::Cast(ao);
|
2010-08-06 07:23:17 +02:00
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
if (httpd->LockHasClients()) {
|
2013-01-15 18:22:17 +01:00
|
|
|
static const char silence[1020] = { 0 };
|
2011-09-16 23:31:48 +02:00
|
|
|
return httpd_output_play(ao, silence, sizeof(silence),
|
2013-08-10 18:02:44 +02:00
|
|
|
IgnoreError()) > 0;
|
2010-08-06 07:23:17 +02:00
|
|
|
} else {
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
inline void
|
2013-07-30 20:11:57 +02:00
|
|
|
HttpdOutput::SendTag(const Tag *tag)
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
2013-10-28 23:58:17 +01:00
|
|
|
assert(tag != nullptr);
|
2009-05-05 22:36:44 +02:00
|
|
|
|
2013-07-30 09:04:05 +02:00
|
|
|
if (encoder->plugin.tag != nullptr) {
|
2009-05-05 22:41:36 +02:00
|
|
|
/* embed encoder tags */
|
2009-05-05 22:51:17 +02:00
|
|
|
|
|
|
|
/* flush the current stream, and end it */
|
|
|
|
|
2013-08-10 18:02:44 +02:00
|
|
|
encoder_pre_tag(encoder, IgnoreError());
|
2013-01-27 23:23:46 +01:00
|
|
|
BroadcastFromEncoder();
|
2009-05-05 22:51:17 +02:00
|
|
|
|
|
|
|
/* send the tag to the encoder - which starts a new
|
|
|
|
stream now */
|
2009-05-05 22:41:36 +02:00
|
|
|
|
2013-08-10 18:02:44 +02:00
|
|
|
encoder_tag(encoder, tag, IgnoreError());
|
2009-05-05 22:51:17 +02:00
|
|
|
|
|
|
|
/* the first page generated by the encoder will now be
|
|
|
|
used as the new "header" page, which is sent to all
|
|
|
|
new clients */
|
|
|
|
|
2013-01-30 09:18:52 +01:00
|
|
|
Page *page = ReadPage();
|
2013-10-28 23:58:17 +01:00
|
|
|
if (page != nullptr) {
|
|
|
|
if (header != nullptr)
|
2013-01-30 09:18:52 +01:00
|
|
|
header->Unref();
|
2013-01-27 23:23:46 +01:00
|
|
|
header = page;
|
|
|
|
BroadcastPage(page);
|
2009-05-05 22:51:17 +02:00
|
|
|
}
|
2009-05-05 22:41:36 +02:00
|
|
|
} else {
|
|
|
|
/* use Icy-Metadata */
|
|
|
|
|
2013-10-28 23:58:17 +01:00
|
|
|
if (metadata != nullptr)
|
2013-01-30 09:18:52 +01:00
|
|
|
metadata->Unref();
|
2013-01-27 23:23:46 +01:00
|
|
|
|
2013-10-20 13:32:59 +02:00
|
|
|
static constexpr TagType types[] = {
|
2013-01-30 09:13:46 +01:00
|
|
|
TAG_ALBUM, TAG_ARTIST, TAG_TITLE,
|
|
|
|
TAG_NUM_OF_ITEM_TYPES
|
|
|
|
};
|
|
|
|
|
2013-07-30 20:11:57 +02:00
|
|
|
metadata = icy_server_metadata_page(*tag, &types[0]);
|
2013-10-28 23:58:17 +01:00
|
|
|
if (metadata != nullptr) {
|
2013-01-27 23:23:46 +01:00
|
|
|
const ScopeLock protect(mutex);
|
|
|
|
for (auto &client : clients)
|
|
|
|
client.PushMetaData(metadata);
|
2009-05-05 22:41:36 +02:00
|
|
|
}
|
2009-04-13 19:35:02 +02:00
|
|
|
}
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
}
|
|
|
|
|
2013-01-27 23:23:46 +01:00
|
|
|
static void
|
2013-07-30 20:11:57 +02:00
|
|
|
httpd_output_tag(struct audio_output *ao, const Tag *tag)
|
2013-01-27 23:23:46 +01:00
|
|
|
{
|
2013-12-31 16:59:18 +01:00
|
|
|
HttpdOutput *httpd = HttpdOutput::Cast(ao);
|
2013-01-27 23:23:46 +01:00
|
|
|
|
|
|
|
httpd->SendTag(tag);
|
|
|
|
}
|
|
|
|
|
2013-12-31 17:01:08 +01:00
|
|
|
inline void
|
|
|
|
HttpdOutput::CancelAllClients()
|
|
|
|
{
|
|
|
|
const ScopeLock protect(mutex);
|
2013-12-31 16:31:36 +01:00
|
|
|
|
|
|
|
while (!pages.empty()) {
|
|
|
|
Page *page = pages.front();
|
|
|
|
pages.pop();
|
|
|
|
page->Unref();
|
|
|
|
}
|
|
|
|
|
2013-12-31 17:01:08 +01:00
|
|
|
for (auto &client : clients)
|
|
|
|
client.CancelQueue();
|
2013-12-31 16:31:36 +01:00
|
|
|
|
|
|
|
cond.broadcast();
|
2013-12-31 17:01:08 +01:00
|
|
|
}
|
|
|
|
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
static void
|
2011-09-16 23:31:48 +02:00
|
|
|
httpd_output_cancel(struct audio_output *ao)
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
{
|
2013-12-31 16:59:18 +01:00
|
|
|
HttpdOutput *httpd = HttpdOutput::Cast(ao);
|
2013-12-31 16:31:36 +01:00
|
|
|
|
|
|
|
BlockingCall(io_thread_get(), [httpd](){
|
|
|
|
httpd->CancelAllClients();
|
|
|
|
});
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
const struct audio_output_plugin httpd_output_plugin = {
|
2013-01-15 18:22:17 +01:00
|
|
|
"httpd",
|
|
|
|
nullptr,
|
|
|
|
httpd_output_init,
|
|
|
|
httpd_output_finish,
|
|
|
|
httpd_output_enable,
|
|
|
|
httpd_output_disable,
|
|
|
|
httpd_output_open,
|
|
|
|
httpd_output_close,
|
|
|
|
httpd_output_delay,
|
|
|
|
httpd_output_tag,
|
|
|
|
httpd_output_play,
|
|
|
|
nullptr,
|
|
|
|
httpd_output_cancel,
|
|
|
|
httpd_output_pause,
|
|
|
|
nullptr,
|
httpd: new output plugin to replace "shout"
Let's get rid of the "shout" plugin, and the awfully complicated
icecast daemon setup! MPD can do better if it's doing the HTTP server
stuff on its own. This new plugin has several advantages:
- easier to set up - only one daemon, no password settings, no mount
settings
- MPD controls the encoder and thus already knows the packet
boundaries - icecast has to parse them
- MPD doesn't bother to encode data while nobody is listening
This implementation is very experimental (no header parsing, ignores
request URI, no icy-metadata, ...). It should be able to suport
several encoders in parallel in the future (with different bit rates,
different codec, ...), to make MPD the perfect streaming server. Once
MPD gets multi-player support, we can even mount several different
radio stations on one server.
2009-03-15 03:32:34 +01:00
|
|
|
};
|