Difference between revisions of "GstWebRTC - Custom Signaling"
| Line 4: | Line 4: | ||
This page provides the basics on how to implement custom signalers. In particular, this page describes how to integrate the custom signaler with the project's build system. The [[GstWebRTC - API Reference | API Reference]] shows how to develop a custom signaler that interacts with GstWebRTC state machine. | This page provides the basics on how to implement custom signalers. In particular, this page describes how to integrate the custom signaler with the project's build system. The [[GstWebRTC - API Reference | API Reference]] shows how to develop a custom signaler that interacts with GstWebRTC state machine. | ||
| + | |||
==Custom Signaling Basics== | ==Custom Signaling Basics== | ||
| Line 9: | Line 10: | ||
Custom signalers are implemented by subclassing the GstBaseSignaler class. By doing so, the new signaler is integrated into GstWebRTC state machine logic. The diagram in Figure 1 summarizes the concept: | Custom signalers are implemented by subclassing the GstBaseSignaler class. By doing so, the new signaler is integrated into GstWebRTC state machine logic. The diagram in Figure 1 summarizes the concept: | ||
| − | [[File: | + | [[File:gstwebrtc-uml-signaler.png|thumb|Figure 1. UML Diagram of Concrete Signalers]] |
| + | |||
| + | |||
==Integrating a Custom Signaler== | ==Integrating a Custom Signaler== | ||
There are two ways a custom signaler can be integrated into the project. The first one is to build the new signaler along with the GstWebRTC project. In this case the signaler will be part of the project's binaries and the signaler's properties will be shown in the inspect output. The second way to integrate an external signaler is by doing it at runtime via the '''signaler-obj'''. In this case the signaler is built in an independent project and passed to the WebRTC state machine at runtime. The following subsections present further details. | There are two ways a custom signaler can be integrated into the project. The first one is to build the new signaler along with the GstWebRTC project. In this case the signaler will be part of the project's binaries and the signaler's properties will be shown in the inspect output. The second way to integrate an external signaler is by doing it at runtime via the '''signaler-obj'''. In this case the signaler is built in an independent project and passed to the WebRTC state machine at runtime. The following subsections present further details. | ||
| + | |||
| + | |||
===Integrating Into GstWebRTC Project=== | ===Integrating Into GstWebRTC Project=== | ||
Revision as of 14:08, 20 September 2017
| ← |
⌂ Home |
→ |
This page provides the basics on how to implement custom signalers. In particular, this page describes how to integrate the custom signaler with the project's build system. The API Reference shows how to develop a custom signaler that interacts with GstWebRTC state machine.
Contents
Custom Signaling Basics
Custom signalers are implemented by subclassing the GstBaseSignaler class. By doing so, the new signaler is integrated into GstWebRTC state machine logic. The diagram in Figure 1 summarizes the concept:
Integrating a Custom Signaler
There are two ways a custom signaler can be integrated into the project. The first one is to build the new signaler along with the GstWebRTC project. In this case the signaler will be part of the project's binaries and the signaler's properties will be shown in the inspect output. The second way to integrate an external signaler is by doing it at runtime via the signaler-obj. In this case the signaler is built in an independent project and passed to the WebRTC state machine at runtime. The following subsections present further details.
Integrating Into GstWebRTC Project
The easiest way to integrate a custom signaler is to extend the project's build system to construct the new code. The following points summarize the steps necessary to integrate an custom signaler named GstExampleSignaler.
1. Specify the new sources:
At the end of gst/webrtc/Makefile.am:
libgstwebrtc_la_SOURCES += gstexamplesignaler.c
noinst_HEADERS += gstexamplesignaler.h
2. Specify extra compile and link flags (if needed):
At the end of gst/webrtc/Makefile.am:
libgstwebrtc_la_CFLAGS += -I/usr/include/dependency
libgstwebrtc_la_LIBADD += -L/usr/lib/dependency -ldependency
Alternatively, for advanced users, if the dependencies export a pkg-config file, the PKG_CHECK_MODULES may be used in the configure.ac and use the generated flags in this makefile.
3. Regenerate Makefiles:
At the root of the project:
./autogen.sh
make
sudo make install
Integrating From ExternalProject
An alternative way to integrate an external signaler is to do it at runtime via the signaler-obj property. This is specially useful for signalers whose licensing is different and more restrictive than the one shipped with GstWebRTC. As such, the signaler can be distributed independently.
If both the signaler-obj and the built-in signaler properties are set, the last property set will be the one finally configured.
Programmatically, an application would set the new signaler in the GstWebRTC element similar to the following:
GstElement * webrtcbin;
GstExampleSignaler *examplesignaler;
...
g_object_set (webrtcbin, "signaler-obj", GST_BASE_SIGNALER(example signaler), NULL);
| ← |
⌂ Home |
→ |
