GNU bug report logs - #55230
CC Mode 5.35.1 (C++//l); Cache inconsistency

Previous Next

Full log

Message #14 received at 55230 <at> debbugs.gnu.org (full text, mbox):

From: Michael Welsh Duggan <mwd <at> md5i.com>
To: Alan Mackenzie <acm <at> muc.de>
Cc: Michael Welsh Duggan <mwd <at> md5i.com>, 55230 <at> debbugs.gnu.org
Subject: Re: bug#55230: CC Mode 5.35.1 (C++//l); Cache inconsistency
Date: Thu, 19 May 2022 00:24:18 -0400

Alan Mackenzie <acm <at> muc.de> writes:

> Hello again, Michael.
>
> On Tue, May 03, 2022 at 18:10:36 +0000, Alan Mackenzie wrote:
>> On Mon, May 02, 2022 at 15:30:09 -0400, Michael Welsh Duggan wrote:
>> > Package: cc-mode
>
>> > From emacs -Q:
>
>> > C-x C-f /tmp/foo.cpp RET
>> > M-x c-toggle-parse-state-debug RET
>> > "{}
>
>> > After typing the close brace, I get the following:
>> > c-parse-state inconsistency at 3: using cache: (2), from scratch:
>> > nil.  POINT-MIN: 1
>> > Old state:
>> > (setq c-state-cache '(2) c-state-cache-good-pos 1
>> > c-state-nonlit-pos-cache nil c-state-nonlit-pos-cache-limit 0
>> > c-state-brace-pair-desert nil c-state-point-min 1
>> > c-state-point-min-lit-type nil c-state-point-min-lit-start nil
>> > c-state-min-scan-pos 1 c-state-old-cpp-beg nil c-state-old-cpp-end
>> > nil c-parse-state-point 3)
>
>> > I have not noticed any indentation errors due to this, but any cache
>> > inconsistency might lead to errors down the line, hence this report.
>
>> Many thanks for the bug report.
>
>> This time it is not the cache mechanism itself which is at fault.
>> Instead, it is the first character ", which since it is unbalanced by
>> another ", has had its syntax is temporarily "commented out" by giving
>> it a syntax-table property of "punctuation".  The parse-state cache was
>> generated whilst in this state, but used when it no longer was, hence
>> the inconsistency.
>
>> There are somewhere between 10 and 20 occurrences of c-parse-state in
>> the source code which will need to be "protected" against this fault.  I
>> hope to be able to finish this work in the next two or three days, when
>> I'll get back to you again.
>
> There were actually 32 such occurrences, and they have generated a rather
> large patch which I am taking the liberty of just sending to you.  The
> vast bulk of this patch is just putting (c-with-string-fences ....)
> around these functions, reindenting them.
>
> I've attached the patch as a separate file because it is so big.  It
> should apply cleanly to the CC Mode in the Emacs savannah master version.
> Would you please apply it and try it out a bit, and let me know how well
> it works.  Thanks!

Unfortunately, it has some major problems.  If you follow the original
recipe, after the first brace is input cc-mode errors out with "Marker
does not point anywhere".  Backtrace follows:

Debugger entered--Lisp error: (error "Marker does not point anywhere")
  c-restore-string-fences()
  c-electric-brace(nil)
  funcall-interactively(c-electric-brace nil)
  command-execute(c-electric-brace)


>> > Emacs  : GNU Emacs 29.0.50 (build 1, x86_64-pc-linux-gnu, X toolkit, cairo version 1.16.0)
>> >  of 2022-04-09
>> > Package: CC Mode 5.35.1 (C++//l)
>> > Buffer Style: gnu
>> > c-emacs-features: (pps-extended-state col-0-paren posix-char-classes gen-string-delim gen-comment-delim syntax-properties 1-bit)
>
>> > -- 
>> > Michael Welsh Duggan
>> > (md5i <at> md5i.com)

-- 
Michael Welsh Duggan
(md5i <at> md5i.com)

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.