1. Phix: Strange Behavior

Hallo,

While working with maze3d.ex from this thread: https://openeuphoria.org/forum/139479.wc?last_id=139488

Something strange happened to me; I tried to incorporate it into a small program to illustrate the issue. Of course, this causes the program to crash under Phix.
The function is'nt actually necessary in principle.

Maybe it's just me, and I'm missing something.

--/* 
include std/console.e 
--*/ 
constant RAYWHITE={245,245,245,0} 
atom COLOR_IS_SEQ = sequence(RAYWHITE) 
?COLOR_IS_SEQ 
function mzRGB(integer r, integer g, integer b, integer a = 255)  
    if COLOR_IS_SEQ then  
        return {r, g, b, a}  
    else  
        return r + g * 256 + b * 65536 + a * 16777216  
    end if 
end function  
 
?RAYWHITE 
?sequence(RAYWHITE) 
?mzRGB(1,2,3) 
COLOR_IS_SEQ = 1 
?mzRGB(1,2,3) 
wait_key() 

Result with Phix:(unexpected)

1 
{245,245,245,0} 
1 
4278387201 
{1,2,3,255} 

Result with Euphoria:(expected)

1 
{245,245,245,0} 
1 
{1,2,3,255} 
{1,2,3,255} 

Andreas

new topic     » topic index » view message » categorize

2. Re: Phix: Strange Behavior

constant RAYWHITE={245,245,245,0}  
atom COLOR_IS_SEQ = sequence(RAYWHITE) 
?COLOR_IS_SEQ  
?COLOR_IS_SEQ 

Result:
1
0

new topic     » goto parent     » topic index » view message » categorize

3. Re: Phix: Strange Behavior

Did you forget to include raylib.e in your program file?

new topic     » goto parent     » topic index » view message » categorize

4. Re: Phix: Strange Behavior

Icy_Viking said...

Did you forget to include raylib.e in your program file?

This has nothing to do with raylib.e - the variable that receives the result of the function sequence() mysteriously changes its value after it is printed ...

constant s = {} 
atom x 
x = sequence(s) 
? x 
? x 

Result:
1
0

new topic     » goto parent     » topic index » view message » categorize

5. Re: Phix: Strange Behavior

Wow! Thanks for finding that!

The vast majority of programs rely exclusively on opSq0, which is fine and leaves e/rdx set to 0/1, however
opSq, which stores that 0/1 in a variable, is quite astonishingly simply backwards and always has been!
In many cases the compiler mitigates that by knowing what it should be rather than reloading from memory.
I even left myself the correct 32-bit code commented out and incorrectly optimised away...

Replace two uses of opAtomContinue in builtins\VM\pType.e from line 192 onwards with the two lines following it.

    :%opSq                          -- [e/rdi]=sequence(e/rax)  (NB: opSeq is dest=(p1=p2)) 
---------- 
        [32] 
            -- calling convention 
            --  lea edi,[dest] 
            --  mov eax,[src]       -- ref of src 
            --  mov ecx,src         -- var no of src 
            --  call opSq           -- [edi]:=sequence(eax) 
--          xor ebx,ebx 
            cmp eax,h4 
            mov edx,[edi] 
            jl :edi0 
          :!opSqe92                     -- exception mapped to e94vhnbaavecxfeh [DEV] 
            test byte[ebx+eax*4-1],0x80 
--          jmp :opAtomContinue // ***WRONG*** (1/11/2026!) 
            jz :edi0 
            jmp :edi1 
--  jz edi0 
--  mov dword[edi],1 
--  cmp edx,h4 
--  jle @f 
--      sub dword[ebx+edx*4-8],1 
--      jz :%pDealloc 
--  @@: 
--  ret 
        [64] 
            -- calling convention 
            --  lea rdi,[dest] 
            --  mov rax,[src]       -- ref of src 
            --  mov rcx,src         -- var no of src 
            --  call opSq           -- [rdi]:=sequence(rax) 
            mov r15,h4 
            mov rdx,[rdi] 
            cmp rax,r15 
            jl :edi0 
          :!opSqe92                     -- exception mapped to e94vhnbaavecxfeh [DEV] 
            test byte[rbx+rax*4-1],0x80 
--          jmp :opAtomContinue // ***WRONG*** (1/11/2026!) 
            jz :edi0 
            jmp :edi1 
        [] 

You will of course need to run "p -c p" for these changes to take proper effect.

new topic     » goto parent     » topic index » view message » categorize

6. Re: Phix: Strange Behavior

Yeah, works! Thank you!

new topic     » goto parent     » topic index » view message » categorize

7. Re: Phix: Strange Behavior

Thank you,

that fixed the problem.

Andreas

new topic     » goto parent     » topic index » view message » categorize

Search



Quick Links

User menu

Not signed in.

Misc Menu