Assuming that char is signed

[anotherlin]

Assuming that char is signed

[anotherlin]

Hi everyone,

Thank you for reporting this bug.
Yes, I indeed assumed that char are signed, but I can't remember why.
My mistake, standard doesn't specify if char is signed or unsigned by default.
I will fix that ASAP.

Best regards,

[allender]

Hello. I have verified that this is a problem on compiler that assume char is unsigned. I think that the OP meant to indicate that the xtensa compiler assumes char is unsigned. I compiled my project with char as unsigned by default and also ran into the same shift issues. Just wanted to confirm that this is indeed a problem.

[mengstr]

Ah yes.... I got confused while writing this issue - sorry for that. :-) But you are understanding me correctly. The xtensa compiler have unsigned chars whereas the code assumes that chars are signed.

But my crc's was slightly different from yours:

<rlca,rrca,rla,rra>...........  OK
shf/rot (<ix,iy>+1)...........  ERROR **** crc expected:710034cb found:cb127686
shf/rot <b,c,d,e,h,l,(hl),a>..  ERROR **** crc expected:a4255833 found:b5821b2b
<set,res> n,<bcdehl(hl)a>.....  OK

[anotherlin]

Hi again,

Ok, I reproduced the problem using gcc and "-funsigned-char" option.
Your CRCs are what I found, Mark's is CRC for zextest, and SmallRoomLabs' is CRC for zexall.
Long ago, I tried z80emu on an old SGI using MipsPro and I had a similar CRC error.
Don't remember well and didn't have time to investigate but I guess it was the same bug.
Working on fixing that, thanks again for your reports.

Best regards,

[anotherlin]

Hi everyone,

Ok, I've fixed the code. And tested with "-funsigned-char".
Can you guys please give it a try, thank you?
If that's fine, I'll close this issue and update the README.txt.
You'll be credited and thanked, of course.

Best regards,

[mengstr]

Runs fine on myLX106 hardware now. You even found the RLC_INDIRECT_HL: which I had missed in my fixes. Strangely enough it seems like the zex'es doesn't cover that particular instruction fully enough to discover errors in the displacement.

Great job! Thanks for this.