[Nasm-bugs] [Bug 3392406] 2.13 regression in (8086) "call far" for 16-bit DOS .OBJ ??

[no-reply at bugzilla-nasm.gorcunov.org]

https://bugzilla.nasm.us/show_bug.cgi?id=3392406

rugxulo at gmail.com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rugxulo at gmail.com

--- Comment #2 from rugxulo at gmail.com ---
> Could you please verify that this generate incorrect code
> past the linker stage?

The FD kernel build is a bit convoluted. Something like
making an .EXE, running EXEFLAT on it, and UPX'ing that.
(Note that I'm not a kernel developer, so I don't truly
understand every last detail.)

So yes, the binary differs, that's how I found this "bug"
at all.

I can point you to a VM (extremely small, FOSS), if you
think it'll help. It's very easy to reproduce.

> Also, a minimal test case would be incredibly helpful,
> both to track down the bug and avoid future regressions.

I thought that was minimal (isolated)! But I guess you
mean even simpler? I'll try, but no promises.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are watching all bug changes.