LightMessaging: Difference between revisions

From iPhone Development Wiki
 
(18 intermediate revisions by 2 users not shown)
Line 1: Line 1:
LightMessaging is an [[IPC|inter-process communication]] header-only library by [[User:Rpetrich|Ryan Petrich]]. According to its [https://github.com/rpetrich/LightMessaging GitHub page], it is a "simple low-level replacement for CFMessagePort".
LightMessaging is a work-in-progress [[IPC|inter-process communication]] header-only library by [[User:Rpetrich|Ryan Petrich]]. According to its [https://github.com/rpetrich/LightMessaging GitHub page], it is a "simple low-level replacement for CFMessagePort".


== How to use this library ==
== How to use this library ==
Line 17: Line 17:
<source lang="objc">
<source lang="objc">
#define LIGHTMESSAGING_USE_ROCKETBOOTSTRAP 0
#define LIGHTMESSAGING_USE_ROCKETBOOTSTRAP 0
</source>
=== Timeouts ===
Use this directive (before including the LM header) to set the timeout:
<source lang="objc">
#define LIGHTMESSAGING_TIMEOUT 500 // 500 ms timeout
</source>
</source>


== Server ==
== Server ==


This should be done in [[SpringBoard.app|SpringBoard]] or a [[Daemons|daemon]]. If you put the server in anywhere other than SpringBoard, then you must register the service in SpringBoard. [[RocketBootstrap#Usage_notes|Further explanation here]].
This must be done in a non-sandboxed process. Typically, this is done in [[SpringBoard.app|SpringBoard]] or a ''non-sandboxed'' [[Daemons|daemon]]. If you put the server anywhere other than SpringBoard, then you must register the service in SpringBoard using <code>rocketbootstrap_unlock</code>. [[RocketBootstrap#Usage_notes|More info here]].
 
'''Disclaimer''': The server interface is not finished yet; it may change in a future version.


=== The Callback ===
=== The Callback ===


<source lang="objc">
<source lang="objc">
void some_callback(CFMachPortRef port,
void some_callback(CFMachPortRef port, LMMessage *message, CFIndex size, void *info) {
                  LMMessage *request,
                  CFIndex size,
                  void *info)
{
// get the reply port
// get the reply port
mach_port_t replyPort = request->head.msgh_remote_port;
mach_port_t replyPort = message->head.msgh_remote_port;


// sanity check
// Check validity of message
if (size < sizeof(LMMessage)) {
if (!LMDataWithSizeIsValidMessage(message, size)) {
LMSendReply(replyPort, NULL, 0);
LMSendReply(replyPort, NULL, 0);
LMResponseBufferFree((LMResponseBuffer *)request);
LMResponseBufferFree((LMResponseBuffer *)message);
return;
return;
}
}
      
      
// get the data you recieved
// Get the data you received
void *data = LMMessageGetData(request);
void *data = LMMessageGetData(message);
size_t length = LMMessageGetDataLength(message);
// Make it into a CFDataRef object
CFDataRef cfdata = CFDataCreateWithBytesNoCopy(kCFAllocatorDefault, (const UInt8 *)data ?: (const UInt8 *)&data, length, kCFAllocatorNull);


// send some data back
// Send some data back
const char *msg = "lol";
const char *msg = "lol";
LMSendReply(replyPort, msg, strlen(msg) + 1);
LMSendReply(replyPort, msg, strlen(msg) + 1);


// free the response buffer??
// Free the CFDataRef object
LMResponseBufferFree((LMResponseBuffer *)request);
if (cfdata) {
CFRelease(cfdata);
}
// Free the response buffer
LMResponseBufferFree((LMResponseBuffer *)message);
}
}
</source>
</source>
Line 61: Line 74:
== Client ==
== Client ==


This can be run anywhere that has the appropriate permissions to send/receive data on a message port.
This can be run anywhere; like a command line program, sandboxed process, daemon, etc.


<source lang="objc">
<source lang="objc">
Line 79: Line 92:
// do whatever you want !!!!
// do whatever you want !!!!
const char *data = LMMessageGetData(response);
const char *data = LMMessageGetData(response);
//cleanup
LMResponseBufferFree(&buffer);
</source>
</source>


Line 91: Line 107:
* [https://github.com/r-plus/libcanopenurl/blob/master/LCCanOpenURL.x libcanopenurl]
* [https://github.com/r-plus/libcanopenurl/blob/master/LCCanOpenURL.x libcanopenurl]
* [https://github.com/rpetrich/AppList/blob/master/ALApplicationList.x AppList]
* [https://github.com/rpetrich/AppList/blob/master/ALApplicationList.x AppList]
* [https://github.com/a3tweaks/Flipswitch/blob/master/FSSwitchMainPanel.m#L750 FlipSwitch]


== External links ==
== External links ==

Latest revision as of 19:00, 24 March 2017

LightMessaging is a work-in-progress inter-process communication header-only library by Ryan Petrich. According to its GitHub page, it is a "simple low-level replacement for CFMessagePort".

How to use this library

Headers are available from LightMessaging's GitHub project. If using Theos, place the headers in $THEOS/include/LightMessaging.

Include directive

#import <LightMessaging/LightMessaging.h>

RocketBootstrap

LightMessaging uses RocketBootstrap by default, so be sure to have it set up (instructions here). To have LightMessaging not use RBS, add the following directive before including the LM header:

#define LIGHTMESSAGING_USE_ROCKETBOOTSTRAP 0

Timeouts

Use this directive (before including the LM header) to set the timeout:

#define LIGHTMESSAGING_TIMEOUT 500 // 500 ms timeout

Server

This must be done in a non-sandboxed process. Typically, this is done in SpringBoard or a non-sandboxed daemon. If you put the server anywhere other than SpringBoard, then you must register the service in SpringBoard using rocketbootstrap_unlock. More info here.

Disclaimer: The server interface is not finished yet; it may change in a future version.

The Callback

void some_callback(CFMachPortRef port, LMMessage *message, CFIndex size, void *info) {
	// get the reply port
	mach_port_t replyPort = message->head.msgh_remote_port;

	// Check validity of message
	if (!LMDataWithSizeIsValidMessage(message, size)) {
		LMSendReply(replyPort, NULL, 0);
		LMResponseBufferFree((LMResponseBuffer *)message);
		return;
	}
    
	// Get the data you received
	void *data = LMMessageGetData(message);
	size_t length = LMMessageGetDataLength(message);
	// Make it into a CFDataRef object
	CFDataRef cfdata = CFDataCreateWithBytesNoCopy(kCFAllocatorDefault, (const UInt8 *)data ?: (const UInt8 *)&data, length, kCFAllocatorNull);

	// Send some data back
	const char *msg = "lol";
	LMSendReply(replyPort, msg, strlen(msg) + 1);

	// Free the CFDataRef object
	if (cfdata) {
		CFRelease(cfdata);
	}
	// Free the response buffer
	LMResponseBufferFree((LMResponseBuffer *)message);
}

Starting the Server

LMStartService("net.iphonedevwiki.some.server", CFRunLoopGetCurrent(), (CFMachPortCallBack)some_callback);

Client

This can be run anywhere; like a command line program, sandboxed process, daemon, etc.

// setup connection
LMConnection connection = {
	MACH_PORT_NULL,
	"net.iphonedevwiki.some.server"
};

//send message
LMResponseBuffer buffer;
const char *msg = "lol wtf";
SInt32 messageId = 0x1111; // this is arbitrary i think
LMConnectionSendTwoWay(&connection, messageId, msg, strlen(msg) + 1, &buffer);
LMMessage *response = &(buffer.message);

// do whatever you want !!!!
const char *data = LMMessageGetData(response);

//cleanup
LMResponseBufferFree(&buffer);

Type helpers

There are some cool helper functions for if you want to send Images / PropertyLists / etc instead of just strings. Look at the header file or other examples to see how to use them.

(this section should be expanded)

Examples

External links