view test/testhread.c @ 1321:e05bc5d315e3

Date: Wed, 1 Feb 2006 18:20:33 -0800 From: Eric Wing Subject: SDL/Universal Binary updates Hi Sam, Attached is a big set of changes to the Xcode projects to support Universal Binaries. I have also included .dmgs that include the prebuilt frameworks. Ryan, I also updated SMPEG which is also in the package. The SDL and smpeg binaries were built against the CVS version (pulled maybe a month ago?). I do not have an Intel Mac to test on so I have no idea if this stuff actually works. However, Christian Walther has been a big help in testing 10.2.8 and 10.3.9 so I'm fairly confident we got the build settings correct for at least PPC. I have attempted to document the important things for producing these Universal Binaries. Documentation is somewhat scattered through out everything, but there is a big centralized piece of documentation in the UniversalBinaryNotes.rtf in the SDL.dmg. As far as Universal Binaries are concerned, the big things were: - Build with gcc 3.3 on PPC, 4.0 on Intel. - We couldn't get any of the MMX/SSE code to compile/link (SDL and smpeg). - All 3rd party dependencies had to be rebuilt as Universal There were also a bunch of non-Universal things that have been updated: - I converted the SDL-satellites to create .dmg's instead of .pkg installers - Updated all 3rd party static libraries with current versions. (I think libpng was the most dramatic going from 1.0.? to 1.2.8 with API breakage. I haven't found any problems so far in doing this.) - Changed some compiler optimization settings - Finally updated the exports list for SDL_mixer - Tried to include a static smpeg in SDL_mixer (multiple build variants in Xcode project now) - Enabled Altivec in SDL (we forgot to add the flags to Xcode last time) - More documentation Since so many things have changed, there might be new problems introduced. The big issue I've found so far is with SDL_mixer. As I mentioned on the mailing list, MP3's produce an assertion failure. And the MikMod problem reported on Bugzilla continues to persist. There's probably a bunch of other stuff I'm forgetting. There really were hundreds of little things I mucked with so it's hard to remember them all. If you have any questions, feel free to ask. Thanks, Eric
author Sam Lantinga <slouken@libsdl.org>
date Thu, 02 Feb 2006 06:26:39 +0000
parents be9c9c8f6d53
children 14717b52abc0
line wrap: on
line source


/* Simple test of the SDL threading code */

#include <stdio.h>
#include <stdlib.h>
#include <signal.h>

#include "SDL.h"
#include "SDL_thread.h"

static int alive = 0;

/* Call this instead of exit(), so we can clean up SDL: atexit() is evil. */
static void quit(int rc)
{
	SDL_Quit();
	exit(rc);
}

int ThreadFunc(void *data)
{
	printf("Started thread %s: My thread id is %u\n",
				(char *)data, SDL_ThreadID());
	while ( alive ) {
		printf("Thread '%s' is alive!\n", (char *)data);
		SDL_Delay(1*1000);
	}
	printf("Thread '%s' exiting!\n", (char *)data);
	return(0);
}

static void killed(int sig)
{
	printf("Killed with SIGTERM, waiting 5 seconds to exit\n");
	SDL_Delay(5*1000);
	alive = 0;
	quit(0);
}

int main(int argc, char *argv[])
{
	SDL_Thread *thread;

	/* Load the SDL library */
	if ( SDL_Init(0) < 0 ) {
		fprintf(stderr, "Couldn't initialize SDL: %s\n",SDL_GetError());
		return(1);
	}

	alive = 1;
	thread = SDL_CreateThread(ThreadFunc, "#1");
	if ( thread == NULL ) {
		fprintf(stderr, "Couldn't create thread: %s\n", SDL_GetError());
		quit(1);
	}
	SDL_Delay(5*1000);
	printf("Waiting for thread #1\n");
	alive = 0;
	SDL_WaitThread(thread, NULL);

	alive = 1;
	thread = SDL_CreateThread(ThreadFunc, "#2");
	if ( thread == NULL ) {
		fprintf(stderr, "Couldn't create thread: %s\n", SDL_GetError());
		quit(1);
	}
	SDL_Delay(5*1000);
	printf("Killing thread #2\n");
	SDL_KillThread(thread);

	alive = 1;
	signal(SIGTERM, killed);
	thread = SDL_CreateThread(ThreadFunc, "#3");
	if ( thread == NULL ) {
		fprintf(stderr, "Couldn't create thread: %s\n", SDL_GetError());
		quit(1);
	}
	raise(SIGTERM);

	SDL_Quit();	/* Never reached */
	return(0);	/* Never reached */
}