Partial update waveshare4in2gsc

static void update_display_bw_part(const FrameBufferType& frame_buffer,gfx::rect16 bounds,bool dithering) {
    //Serial.printf("(%d,%d)-(%d,%d)\r\n",bounds.x1,bounds.y1,bounds.x2,bounds.y2);
    // i've checked these constants over and over again. they're right:
    bus_driver::send_command(0x91);
    bus_driver::send_command(0x90);
    int x1=bounds.x1&0xF8;
    int x2=bounds.x2|0x7;
    bus_driver::send_data8(x1 >> 8);
    bus_driver::send_data8(x1 & 0xff);     // x should be the multiple of 8, the last 3 bit will always be ignored
    bus_driver::send_data8(x2 >> 8);
    bus_driver::send_data8(x2 & 0xff);
    bus_driver::send_data8(bounds.y1 >> 8);
    bus_driver::send_data8(bounds.y1 & 0xff);
    bus_driver::send_data8(bounds.y2 >> 8);
    bus_driver::send_data8(bounds.y2 & 0xff);
    bus_driver::send_data8(0x01);         // Gates scan both inside and outside of the partial window. (default)
    delay(2);
    bus_driver::send_command(0x13);
    if(dithering) {
        for(int y = bounds.y1;y<=bounds.y2;++y) {
            int row = y&15;
            for(int x=x1;x<=x2;x+=8) {
                uint8_t b = 0;
                for(int bx=0;bx<8;++bx) {
                    int xx = (bx+x);
                    int col = xx & 15;
                    typename FrameBufferType::pixel_type px;
                    frame_buffer.point(gfx::point16(xx,y),&px);
                    b|=(1<<(7-bx))*
                                !!(255.0 * px.template channelr<
                                            gfx::channel_name::L>() >
                                gfx::helpers::dither::bayer_16[col +
                                                                row * 16]);
                }
                bus_driver::send_data8(b);
            }
        }
    } else {
        // this is the one getting called
        // everything about this loop has been tested
        // in the update_display_bw() routine which contains
        // the same code
        for(int y = bounds.y1;y<=bounds.y2;++y) {
            for(int x=x1;x<=x2;x+=8) {
                uint8_t b = 0;
                for(int bx=0;bx<8;++bx) {
                    typename FrameBufferType::pixel_type px;
                    frame_buffer.point(gfx::point16(bx+x,y),&px);
                    b|=(1<<(7-bx))*!!px.native_value;
                }
                bus_driver::send_data8(b);
            }
        }
    }

    delay(2);

    bus_driver::send_command(0x92);
}