bug#41618: 28.0.50; Can't byte-compile an edebugged macro

[Philipp Stephani]

Am So., 31. Mai 2020 um 19:01 Uhr schrieb Alan Mackenzie <acm@muc.de>:
>
> Hello, Philipp.
>
> In article <mailman.730.1590857164.2541.bug-gnu-emacs@gnu.org> you wrote:
>
> > 1. Define some arbitrary macro:
>
> >   (defmacro foo ())
>
> > 2. Edebug it using C-u C-M-x.
>
> This was not actually relevant.  A simple evaluation with C-M-x produces
> the same error.
>
> > 3. Attempt to byte-compile it using M-: (byte-compile 'foo).
>
> > This produces an error:
>
> >   Wrong type argument: listp, #[0 "\300\207" [nil] 1]
>
> > The stack trace is
>
> > Debugger entered--Lisp error: (wrong-type-argument listp 
> > #f(compiled-function () #<bytecode 0x1e0000171e91>))
> >   eval((macro . #f(compiled-function () #<bytecode 0x1e0000171e91>)) t)
> >   #f(compiled-function (form) #<bytecode -0x16fada817203f185>)(foo)
> >   byte-compile(foo)
> >   eval((byte-compile 'foo) t)
> >   eval-expression((byte-compile 'foo) nil nil 127)
> >   funcall-interactively(eval-expression (byte-compile 'foo) nil nil 127)
> >   call-interactively(eval-expression nil nil)
> >   command-execute(eval-expression)
>
> This was quite a simple bug.  At the end of byte-compile, the code does
> two things:
> (i) If the argument to byte-compile is a symbol, the result is eval'd.
> (ii) If a macro is being compiled, 'macro is pushed onto the result.
>
> When both of these things were necessary, they were being done in the
> wrong order, throwing the error.
>
> I've committed a fix to the emacs-27 branch, and it should reach master
> the next time "somebody" copies the commits over.  In the mean time,
> here's that patch, should you want to apply it to your system now:
>


Thanks for the quick fix. Confirmed that it's fixed on the emacs-27 branch.