[Garufa]

Hi, I'm trying to show these messages through modules, but I can't. Can anyone help me?

Here is the code:

PHP Code:

void Show_dHudMessage( edict_t *pEdict )
{
    const char message[ 128 ] = "Hello world!";

    // Green color
    unsigned long color = 160 << 8;

    MESSAGE_BEGIN( MSG_ONE, SVC_DIRECTOR, NULL, pEdict );

        WRITE_BYTE( strlen( message ) + 31 );
        WRITE_BYTE( DRC_CMD_MESSAGE );
        WRITE_BYTE( 1 );    // effect
        
        WRITE_LONG( color );    // color
        
        WRITE_LONG( FixedSigned16( -1.0, 1 << 13 ) );    // x
        WRITE_LONG( FixedSigned16( 0.25, 1 << 13 ) );    // y

        WRITE_LONG( FixedUnsigned16( 0.1, 1 << 8 ) );    // fadeintime
        WRITE_LONG( FixedUnsigned16( 1.5, 1 << 8 ) );    // fadeouttime
        WRITE_LONG( FixedUnsigned16( 3.0, 1 << 8 ) );    // holdtime
        WRITE_LONG( FixedUnsigned16( 6.0, 1 << 8 ) );    // fxtime

        WRITE_STRING( message );

    MESSAGE_END(  );
}

short FixedSigned16( float value, float scale )
{
    int output;

    output = (int) (value * scale);

    if ( output > 32767 )
        output = 32767;

    if ( output < -32768 )
        output = -32768;

    return (short)output;
}

unsigned short FixedUnsigned16( float value, float scale )
{
    int output;

    output = (int) (value * scale);
    if ( output < 0 )
        output = 0;
    if ( output > 0xFFFF )
        output = 0xFFFF;

    return (unsigned short)output;
} 


Thanks!

[Arkshine]

What is the problem exactly ?

[Garufa]

Originally Posted by Arkshine What is the problem exactly ?

It just does not show the message. I suspect the data types I'm using.

Also, I don't understand several things:

PHP Code:

WRITE_BYTE( strlen( message ) + 31 ); 


This should be the size of the message, but because it adds 31?

PHP Code:

WRITE_LONG( FixedSigned16( -1.0, 1 << 13 ) ); 


The scale is correct?. Why do not use either WRITE_ANGLE or WRITE_COORD?. Both write float values​​.

I look at the hud_spectator.cpp, specifically the parse of DRC_CMD_MESSAGE:

PHP Code:

msg->x = READ_FLOAT();    // x pos
msg->y = READ_FLOAT();    // y pos
                                                
msg->fadein = READ_FLOAT();    // fadein
msg->fadeout = READ_FLOAT();    // fadeout
msg->holdtime = READ_FLOAT();    // holdtime
msg->fxtime = READ_FLOAT();    // fxtime; 


As you see, it uses READ_FLOAT() to parse the values, but what should be used to write?

I'm confused.

PHP Code:

// From parsemsg.cpp

float READ_FLOAT( void )
{
    union
    {
        byte    b[4];
        float   f;
        int     l;
    } dat;
    
    dat.b[0] = gpBuf[giRead];
    dat.b[1] = gpBuf[giRead+1];
    dat.b[2] = gpBuf[giRead+2];
    dat.b[3] = gpBuf[giRead+3];
    giRead += 4;
    
//    dat.l = LittleLong (dat.l);

    return dat.f;   
} 


Thanks in advanced.

PD: Sorry for my english.

[Arkshine]

The first arg is the length of the message all the WRITE_*.

WRITE_FLOAT doesn't exist. WRITE_LONG works with float without problem.

I don't see what's wrong. Will try myself later.

[jim_yang]

pawn use cell present 4 bytes value in 32bit system, and add tag to parse it. so the mem form is just 32bit, depend on how you read it.
Float:a = 1.0;
binary is 0011 1111 1000 0000 0000 0000 0000 0000
if you untag the a (use _: ), then it will read the bits above as an int/long (32bits) form. then you will get 1065353216.
you can see that message use write_long to send this value, and parse it use read_float, so you should pass the value using its int form of that float value.

here is the transform function

long ftol(float x)
{
    __asm mov eax, x;
}

or the c format

long ftol(float x)
{
    return *(long *)(&x);
}

then you can correctly send this message, here is an example

	float x = -1.0;
	float y = 0.25;
	float fadein = 1.0;
	float fadeout = 1.0;
	float holdtime = 6.0;
	float fxtime = 1.0;

	char msg[] = "hello world!";
	MESSAGE_BEGIN(MSG_ALL, SVC_DIRECTOR);
	WRITE_BYTE(strlen(msg) + 31);
	WRITE_BYTE(6);
	WRITE_BYTE(1);
	WRITE_LONG(160<<8);
	WRITE_LONG(ftol(x));
	WRITE_LONG(ftol(y));
	WRITE_LONG(ftol(fadein));
	WRITE_LONG(ftol(fadeout));
	WRITE_LONG(ftol(holdtime));
	WRITE_LONG(ftol(fxtime));
	WRITE_STRING(msg);
	MESSAGE_END();

[Shadows In Rain]

Keep style :]

inline void WRITE_FLOAT(float val)
{
	WRITE_LONG(*reinterpret_cast<float*>(&val));
}

[Garufa]

Originally Posted by jim_yang pawn use cell present 4 bytes value in 32bit system, and add tag to parse it. so the mem form is just 32bit, depend on how you read it. Float:a = 1.0; binary is 0011 1111 1000 0000 0000 0000 0000 0000 if you untag the a (use _: ), then it will read the bits above as an int/long (32bits) form. then you will get 1065353216. you can see that message use write_long to send this value, and parse it use read_float, so you should pass the value using its int form of that float value. here is the transform function...

I knew the problem was in the format of the data. Thanks Jim, your explanation I unlock a great mystery, at least for me ;)