OpenEuphoria

constant xGetMouseWheelMove = define_c_func(ray,GetMouseWheelMove,{},C_FLOAT) 

// File: test_ffi.c 
// compile with:- m32 for 32bit dll 
//gcc -shared -m64 -o test_ffi64.dll test_ffi.c   
//or 
//gcc -shared -m64 -s -o test_ffi64.dll test_ffi.c -Wl,--out-implib,libtest_ffi.a 
 
#include <windows.h> 
 
// Funktioniert in GCC/MinGW 
__attribute__((dllexport)) float GetTestFloat() { 
    return 123.456f; 
} 
 
__attribute__((dllexport)) float MultiplyFloats(float a, float b) { 
    return a * b; 
} 
 

sequence dll={"test_ffi64.dll","test_ffi32.dll"} 
atom result 
atom x=2.5,y=3.5 
atom lib=open_dll(dll) 
 
constant GetTestFloat= define_c_func(lib,"GetTestFloat",{},C_FLOAT) 
constant MultiplyFloats=define_c_func(lib,"MultiplyFloats",{C_FLOAT,C_FLOAT},C_FLOAT) 
 
 
printf(1,"%d Bit %s\n",{machine_bits(),"System"}) 
 
result=c_func(GetTestFloat,{}) 
printf(1,"%f should be 123.456\n",result) 
 
result=c_func(MultiplyFloats,{x,y}) 
printf(1,"%f should be %f\n",{result,x*y}) 
 
wait_key() 

2. Re: Phix: float32 return value

• Posted by petelomax 1 month ago
• 453 views

Oh my, that is an absolutely perfect error report, can I just say well done and thankyou!

The fix itself is pretty straightforward, pcfunc.e ~line 2219 should end up as

                cmp rdx,0x03000004  -- (C_FLOAT) 
--              je :cstorexmm0 
                jne @f 
                    sub rsp,8 
                    movss dword[rsp],xmm0 
                    fld dword[rsp] 
                    jmp :cstorest0 
            @@: 
                cmp rdx,0x03000008  -- (C_DOUBLE) 
                jne @f 
--          ::cstorexmm0 
-- 14/2/16: (certainly C_DOUBLE, not necessarily C_FLOAT?) [25/2, I think it's the same] 
                    sub rsp,8 
                    movsd qword[rsp],xmm0 
                    fld qword[rsp] 
            ::cstorest0 
                    add rsp,8 

However, 1.0.5 does not implement movss, which needs, before making the above changes:

pttree.e line 1289: 
global constant T_swi           = 6060  tt_stringF("swi",T_swi) 
global constant T_movss         = 6068  tt_stringF("movss",T_movss) 
 
pilasm.e line 4333: 
--          elsif ttidx=T_movsd then 
            elsif ttidx=T_movsd 
               or ttidx=T_movss then 
                op = iff(ttidx=T_movsd?0o362:0o363) 
and ten/twenty lines on: 
--                          s5 &= {0o362,0o017,0o020} 
                            s5 &= {op,0o017,0o020} 
 
--                          s5 &= {0o362,0o017,0o021} 
                            s5 &= {op,0o017,0o021} 

You'll then need to run p -c p, then make the changes to pcfunc.e, then run p -c p again.
Actually, probably best if you just change pttree.e first and quickly check by running "p p"
whether you get anything like "movss should be 6068(not 6064)", if not kill it and carry on.

Finally, a quick "p -test" should reassure you nothing else got broken.

Thanks again, I would probably never have found or fixed that without your help.

4. Re: Phix: float32 return value

• Posted by Icy_Viking 1 month ago
• 437 views

I do have a Raylib wrapper for Euphoria 4.1.0 Beta 2 [https://github.com/gAndy50/EuRayLib5]

It uses the FFI library to help with wrapping structs. It might help with making a Phix wrapper.
Forked into: Raylib Phix reasing.e for Icy_Viking

6. Re: Phix: float32 return value

• Posted by andreasWagner 1 month ago
• 429 views

Hallo,

i did this, T_swi is not in pttree.e (from the 1.0.5 release):

pttree.e 
begining with line 1271 this: 
global constant T_swi       = 5956  tt_stringF("swi",T_swi)  
global constant T_movss     = 5964  tt_stringF("movss",T_movss)  

p p worked.
then this in pilasm.e (also from the 1.0.5 release):

pilasm.e 
line 4337: 
--          elsif ttidx=T_movsd then  
            elsif ttidx=T_movsd  
               or ttidx=T_movss then  
                op = iff(ttidx=T_movsd?0o362:0o363) -- 64-bit move (as per movsb, movsw, movsd) 
 
 
line 4301:          --      s5 &= {0o362,0o017,0o020}  
                            s5 &= {op,0o017,0o020}  
 
line 4311           --      s5 &= {0o362,0o017,0o021}  
                            s5 &= {op,0o017,0o021} 

p p now gives this.

D:\testpool\ray_workplace\Phix\pilasm.e:4338 
            elsif ttidx=T_movsd or ttidx=T_movss then op = iff(ttidx=T_movsd?.. 
^ duplicate case value [3364, il offset:2791] 

I think your local copy of Phix is probably very different from the release version. But thanks for the quick response. It shows that Phix is still alive and being maintained.

Edit:I have found a temporary workaround. I have changed raylib; the return value is now int. That is sufficient for the two examples I have ported for phix and which use the function. Now I can continue until the next problems arise.

D:\testpool\ray_workplace\Phix\pilasm.e:4338 
            elsif ttidx=T_movsd or ttidx=T_movss then op = iff(ttidx=T_movsd?.. 
^ duplicate case value [3364, il offset:2791] 

8. Re: Phix: float32 return value

• Posted by andreasWagner 3 weeks ago
• 337 views

No, I'm not really happy. A float32 return value is of course needed more often, especially in raymath. And I wasn't able to apply the patch myself.

Perhaps the easiest thing would be for you to update your GitHub repo. Maybe even with the older patches, thank you.

I had an unpleasant experience. I referred a friend to my https://github.com/andizk4kx/ArduLink. Unfortunately, the program only seems to work with the phix version on my computer and probably yours. It doesn't work with the git repo or the official download. It was something to do with assert https://openeuphoria.org/forum/138715.wc#138715. In the case with my friend, I simply copied everything onto a USB stick and drove over to his place (he trusts me not to bring any viruses with me).

In addition, return values of type c_bool are unstable (I don't know how else to put it). You can try bunnymark test and https://openeuphoria.org/forum/139277.wc?last_id=0

and remove the and_bits() here

global function IsMouseButtonDown(atom button)  
        return and_bits(c_func(xIsMouseButtonDown,{button}),1)  
end function  

Thank you
and apologies for the grumbling and complaining

10. Re: Phix: float32 return value

• Posted by andreasWagner 3 weeks ago
• 274 views

Hi, First of all, thanks. I didn't really expect a response so quickly (I know it takes time and effort).

Of course, I only briefly tested everything I usually do when a new Phix comes out.

So, p -c p works, p64 -c p works too.
..\..\p -c edita.exw, okay
Edix looks promising.
The float32 bug is gone

The old bugs I remember are also gone, meaning my old programs are running again with the official version of Phix.
iup.chm and cd.chm don t work for me with Windows HTML Help; I only get the table of contents the individual contents aren t displayed.
However, I can read the files with the Calibre e-book viewer. It does report missing links in both files, though.

Before I forget, the new icon is much better (the old one always reminded me of Turbo Pascal for Windows from the last century)

12. Re: Phix: float32 return value

• Posted by andreasWagner 6 days ago
• 66 views

Just a few thoughts...

Unfortunately, I had to delve deeply into the MS-ABI (not the MS Win32 API) specifically the AMD64 MS-ABI.

Of course, whether you like it or not, the C programming language is the deciding factor here. The 'bool' data type is unknown in C and was therefore represented as an integer: 0 = false, 1 = true. With C99, a fixed definition was introduced via stdbool.h that did exactly that. This replaced the internal keyword _Bool.

stdbool.h has been deprecated since C23. I don't know at the moment how things will proceed from here.

A simple definition of bool (used in many PLC systems) is that if the least significant bit is 1, it is true; if it is 0, it is false.

The problem with all ABIs is that it's not guaranteed that the irrelevant bits will be cleared. A bool can be 1 byte, 2 bytes, 4 bytes, or even 8 bytes in 64-bit systems. But as far as I've seen, there's no specification anywhere that requires clearing the irrelevant bits.

I think 'and_bits(result, 1)' is the best solution available.

However, this also means that if 'result' were 2, the result would be false.

And many checks like 'if (result) then' would, understandably, no longer work.

These are just my thoughts.

And, since I use a translation tool, it's really a pain that the forum crashes every time I use a Unicode character.