Launching programmes in Win32 & Linux/Unix/Posix

I've got this:

Code:

---

PAW_SHARED struct paw_app* paw_boot_app( void *ud, char const * const path, char *argv[], char *envp[] )
{
#ifdef PAW_WIN32
        /* Not supported yet */
        return NULL;
#else
        int pid = fork();
       
        if ( pid == 0 )
                (void)execve( path, argv, envp );
                /* According to docs the new app will take over the pid and replace
                * our address space with it's own as it initiates, we therefore
                * should assume that if execve returned we failed to launch the app,
                * either way we do not want the child to return a valid object for
                * itself or the root system process which is probably the kernel and
                * not what any normal user will want to use, anyone that does want to
                * use it will not use such a convoluded method as going through this
                * function unless they just want to see what happens, in which case
                * they will have downloaded the project, modified it and compiled that
                * copy instead */
        else if ( pid > 0 )
                /* Instead of duplicating code let's just pass the pid to one that
                * assumes it already exists */
                return paw_load_app( ud, pid );
       
        return NULL;
#endif
}

---

But I'm struggling to find the right way to do it on Win32, also is this the correct way for linux etc?

Windows has CreateProcess() function on Win32 API.

Quote:

---

Originally Posted by flp1969

Windows has CreateProcess() function on Win32 API.

---

ty :)

This is the new state of the function, I'm not expecting there to be issues but posting it on the off chance I overlooked something.

Code:

---

PAW_SHARED struct paw_app* paw_boot_app( void *ud, char const * const path, char *argv[], char *envp[] )
{
#ifdef PAW_WIN32
        struct paw_app *app;
        char *cmd, *dir, *tok = NULL;
        PROCESS_INFORMATION pi = {NULL};
        int created = 0;
        size_t size = strlen( path ) + 1, want = size * 2, done = 0;
       
        for ( int i = 0; argv[i]; ++i )
                want += strlen(argv[i]) + 2;
               
        dir = paw.alloc( ud, NULL, 0, want );
       
        if ( !dir )
                return NULL;
       
        cmd = dir + size;
        memset( dir, 0, want );
       
        for ( int i = 0; argv[i]; ++i )
        {
                char *arg = cmd + done;
                size_t len = strlen(argv[i]);
                memcpy( arg, argv[i], len );
                arg[len] = ' ';
                done += len + 1;
        }
        cmd[done-1] = 0;
       
        created = CreateProcess(
                path, cmd, NULL, NULL, FALSE, 0, (void*)envp, dir, NULL, &pi );
       
        paw.alloc( ud, dir, want, 0 );
        app = created ? paw_load_app( ud, pi.dwProcessId ) : NULL;
       
        if ( app )
        {
                app->hook = pi.hProcess;
                app->tracing = true;
        }
       
        return app;
#else
        int pid = fork();
       
        if ( pid == 0 )
                (void)execve( path, argv, envp );
                /* According to docs the new app will take over the pid and replace
                * our address space with it's own as it initiates, we therefore
                * should assume that if execve returned we failed to launch the app,
                * either way we do not want the child to return a valid object for
                * itself or the root system process which is probably the kernel and
                * not what any normal user will want to use, anyone that does want to
                * use it will not use such a convoluded method as going through this
                * function unless they just want to see what happens, in which case
                * they will have downloaded the project, modified it and compiled that
                * copy instead */
        else if ( pid > 0 )
                /* Instead of duplicating code let's just pass the pid to one that
                * assumes it already exists */
                return paw_load_app( ud, pid );
       
        return NULL;
#endif
}

---

Edit: Minor modification to the win32 code:

Code:

---

        ...
        dir = paw.alloc( ud, NULL, 0, want );
       
        if ( !dir )
                return NULL;
       
        memset( dir, 0, want );
        memcpy( dir, path, size - 1 );
        PathRemoveFileSpecA( dir );
       
        cmd = dir + size;
        ...

---

Sooner or later I'm gonna have to get round to checking for the death of traced & potentially hooked processes, what would be the suggestions for when I get to that point for both systems