Re: master 5252c45: Initialise unread buffer

[Lars Ingebrigtsen <larsi@gnus.org>]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Mattias Engdegård <mattiase@savannah.gnu.org> writes:
>
>>     The reader has an extra 1-char unread buffer that was incorrectly
>>     initialised to 0, which means that the first character read would
>>     always be NUL.  As this is often the code that looks for the
>>     lexical-binding cookie, the first loaded source module would be
>>     treated as dynamically bound.  During bootstrapping this is loadup.el
>>     and so its local variables got dumped into the global environment.
>
> Wow.  How did you manage to debug this?  :-)

But it seems like this leads to random breakages:

Debugger entered--Lisp error: (error "Invalid byte opcode: op=0, ptr=6")
  signal(error ("Invalid byte opcode: op=0, ptr=6"))
  apply(signal (error ("Invalid byte opcode: op=0, ptr=6")))
  (setq value-198 (apply fn-196 args-197))

This is when evalling the body of this test:

(ert-deftest eval-tests-accept-empty-optional-rest ()
  "Check that Emacs accepts empty &optional and &rest arglists.
Bug#24912."
  (dolist (args '((&optional) (&rest) (&optional &rest)
                  (&optional &rest a) (&optional a &rest)))
    (let ((fun `(lambda ,args 'ok)))
      (ert-info ("eval")
        (should (eq (funcall (eval fun t)) 'ok)))
      (ert-info ("byte comp check")
        (byte-compile-check-lambda-list args))
      (ert-info ("bytecomp")
        (let ((byte-compile-debug t))
          (should (eq (funcall (byte-compile fun)) 'ok))))))

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no