Warning: c-font-lock-declarators called with 3 arguments, but requires 4-5

[HinTak]

This is emitted by both the latest (v0.9.2) and also the last csharp-mode I used (Jun 4 2016) against emacs 26.1, whenever I try editing any C# source file now - it causes emacs to split screen with a separate screen for this warning. I can dismiss it with ctrl-x-1 but not very nice...

Warning: c-font-lock-declarators called with
    3 arguments, but requires 4-5

Fedora 28 started shipping emacs 26 mid-release 6 weeks ago.

[josteink]

As far as I know this error should only be emitted during installation and byte-compilation.

Are you saying you get this for every c# buffer? If so, how did you install it?

[HinTak]

I just copy the csharp-mode.el fiLe to my site directory and have my emacrc associate cs files with it...

[josteink]

Try deleting any associated csharp-mode.elc-file you can find.

My guess is it needs to be recompiled.

[josteink]

Also try using latest from git master.

[HinTak]

There are no elc lying around - also, tried latest git master (v0.9.2) also as I wrote above...

[josteink]

Can you reproduce the issue if you start with a clean emacs (emacs -Q) and install this package from MELPA?

Basically, this exact issue should be fixed, and it should be fixed months ago.

This is really strange.

[josteink]

So I don't have Fedora installed anywhere right now, but on my laptop I have Ubuntu 18.04 and run Emacs built from source. I've now done a rebuild from the emacs-26.1 tag, but I'm not able to reproduce the behaviour your describing.

Yes, I get that warning during initial compilation, but I never get it later, and definitely not when opening C# files.

Can you outline a "from scratch" way of reproducing this on your system?

[HinTak]

Hmm, I don't use csharp-mode in the compiled form - as I said, the el source is loaded every time via emacs rc...

[josteink]

To be clear, I just loaded the el file into a buffer, evaluated it and tried to visit a c#-file.

Like I said, I really can’t help unless you demonstrate a step by step, from scratch, way of reproducing this bug.

Could you try to do that?

[HinTak]

Here is section of my .emacs file:

;; Import C# mode
(add-to-list 'load-path "/home/Hin-Tak/.emacs.site-lisp")
(autoload 'csharp-mode "csharp-mode" "Major mode for editing C# code." t)
(setq auto-mode-alist
(append '(("\\.cs$" . csharp-mode)) auto-mode-alist))

my ~/.emacs.site-lisp/ contains csharp-mode.el . There is no elc files lying around. That's all.

[josteink]

Oh I see. That's certainly different from how I do things. I'll do some testing and see if I can reproduce from that.

I can't make any promises though: If we want to remain compatible with earlier emacs-versions, we will have to try to invoke both variants of this underlying cc-mode function.

(Unless we find another, more reliable way to detect which version we should invoke, that is)

[josteink]

I've now tested your init code using a clean environment (emacs -Q) and I can confirm that, yes, we get the compilation error upon visiting the first C# buffer.

But I still cannot reproduce the behaviour you claim to have: That this error pops up for every C#-buffer.

Can you please retest and confirm?

[josteink]

I've also pushed some changes to a new branch: https://github.com/josteink/csharp-mode/tree/bugfix/emacs26-compatibility

Do these changes work better for you?

[HinTak]

No, same problem. Here is a more detailed trace of make:

"/usr/bin/emacs" -Q -batch -L . -l csharp-mode-tests.el -f ert-run-tests-batch-and-exit
csharp-mode.el:Warning: c-font-lock-declarators called with 3 arguments, but
    requires 4-5
Running 29 tests (2018-08-15 21:43:32+0100)
   passed   1/29  activating-mode-doesnt-cause-failure
   passed   2/29  activating-mode-doesnt-clobber-global-adaptive-fill-regexp
Undefined style: defaultc#
Undefined style: defaultc#fromlist
...
Test fontification-of-method-names backtrace:
  signal(ert-test-failed (((should (assess-face-at= "using System;\n\n
  ert-fail(((should (assess-face-at= "using System;\n\n\n/// <summary>
  (if (unwind-protect (setq value-107 (apply fn-105 args-106)) (setq f
  (let (form-description-109) (if (unwind-protect (setq value-107 (app
  (let ((value-107 'ert-form-evaluation-aborted-108)) (let (form-descr
  (let* ((fn-105 (function assess-face-at=)) (args-106 (condition-case
  (progn (require 'assess) (let* ((fn-105 (function assess-face-at=)) 
  (progn (require 'assess) (let* ((fn-100 (function assess-face-at=)) 
  (lambda nil (progn (require 'assess) (let* ((fn-100 (function assess
  ert--run-test-internal(#s(ert--test-execution-info :test #s(ert-test
  ert-run-test(#s(ert-test :name fontification-of-method-names :docume
  ert-run-or-rerun-test(#s(ert--stats :selector t :tests [#s(ert-test 
  ert-run-tests(t #f(compiled-function (event-type &rest event-args) #
  ert-run-tests-batch(nil)
  ert-run-tests-batch-and-exit()
  command-line-1(("-L" "." "-l" "csharp-mode-tests.el" "-f" "ert-run-t
  command-line()
  normal-top-level()
Test fontification-of-method-names condition:
    (ert-test-failed
     ((should
       (assess-face-at= "using System;


...
" csharp-mode "ToString" font-lock-function-name-face)
      :value nil :explanation
      #("Face does not match expected value
	Expected: font-lock-function-name-face
	Actual: font-lock-type-face
	Location: 321
	Line Context: }

	bol Position: 299
" 134 135
(c-type c-decl-id-start))))
   FAILED  11/29  fontification-of-method-names
...
Ran 29 tests, 28 results as expected, 1 unexpected (2018-08-15 21:43:34+0100)

1 unexpected results:
   FAILED  fontification-of-method-names

make: *** [makefile:12: test] Error 1