Two CC Mode bug fixes into the release branch?

Two CC Mode bug fixes into the release branch?

[Alan Mackenzie]

Hello, Emacs.

I think it would be right to put two recent CC Mode bug fixes into the
release branch.

The first is bug #53605 (from Richard Copley in January) which has just
been reported again by somebody else on bug-cc-mode@gnu.org.  The
problem is that the C++ line

    operator""_hexstring()

gets parsed as

    operatoR""_hexstring()

and handled wrongly as a raw string.  The solution to this bug is to
bind case-fold-search to nil in one place.

This was fixed in the master branch in February.  It would seem that
this is likely to happen quite a lot, so maybe the fix should be copied
into the release branch now.


The second is bug #56256 where an unusual set of circumstances (a C++
lambda form being opened in a #define) causes Emacs to lock up hard.

The solution is simple (as these things often are), replacing a call to
c-backward-token-2 with a save-excursion and a goto-char.

This may not happen often, but the consequence would appear serious
enough to warrant fixing the bug in the emacs-28 branch.


Should I put either of these fixes into the release branch?

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: Two CC Mode bug fixes into the release branch?

[Eli Zaretskii]

> Date: Sat, 2 Jul 2022 14:06:12 +0000
> From: Alan Mackenzie <acm@muc.de>
> 
> The first is bug #53605 (from Richard Copley in January) which has just
> been reported again by somebody else on bug-cc-mode@gnu.org.  The
> problem is that the C++ line
> 
>     operator""_hexstring()
> 
> gets parsed as
> 
>     operatoR""_hexstring()
> 
> and handled wrongly as a raw string.  The solution to this bug is to
> bind case-fold-search to nil in one place.

Please tell which commit you want to cherry-pick.  I'd like to be sure
we are talking about the same thing.

> The second is bug #56256 where an unusual set of circumstances (a C++
> lambda form being opened in a #define) causes Emacs to lock up hard.
> 
> The solution is simple (as these things often are), replacing a call to
> c-backward-token-2 with a save-excursion and a goto-char.
> 
> This may not happen often, but the consequence would appear serious
> enough to warrant fixing the bug in the emacs-28 branch.

Is the change you want to install the same one shown in the discussion
of that bug?  If not, please show the change.

Re: Two CC Mode bug fixes into the release branch?

[Alan Mackenzie]

Hello, Eli.

On Sat, Jul 02, 2022 at 17:41:43 +0300, Eli Zaretskii wrote:
> > Date: Sat, 2 Jul 2022 14:06:12 +0000
> > From: Alan Mackenzie <acm@muc.de>

> > The first is bug #53605 (from Richard Copley in January) which has just
> > been reported again by somebody else on bug-cc-mode@gnu.org.  The
> > problem is that the C++ line

> >     operator""_hexstring()

> > gets parsed as

> >     operatoR""_hexstring()

> > and handled wrongly as a raw string.  The solution to this bug is to
> > bind case-fold-search to nil in one place.

> Please tell which commit you want to cherry-pick.  I'd like to be sure
> we are talking about the same thing.

OK, this commit:

commit cf9c40159a6bef5a23f7b23a45c03f0c6be57574
Author: Alan Mackenzie <acm@muc.de>
Date:   Mon Jan 31 17:44:59 2022 +0000

    * lisp/progmodes/cc-mode.el (c-common-init): Bind case-fold-search to nil

    This fixes bug #53605.

It is best viewed with a -b flag, as in $ git show -b cf9c40159a.


> > The second is bug #56256 where an unusual set of circumstances (a C++
> > lambda form being opened in a #define) causes Emacs to lock up hard.

> > The solution is simple (as these things often are), replacing a call to
> > c-backward-token-2 with a save-excursion and a goto-char.

> > This may not happen often, but the consequence would appear serious
> > enough to warrant fixing the bug in the emacs-28 branch.

> Is the change you want to install the same one shown in the discussion
> of that bug?  If not, please show the change.

Yes, it's the change I send off to the OP on 2022-06-29, asking him/her
to test it.

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: Two CC Mode bug fixes into the release branch?

[Eli Zaretskii]

> Date: Sat, 2 Jul 2022 15:18:04 +0000
> Cc: emacs-devel@gnu.org
> From: Alan Mackenzie <acm@muc.de>
> 
> > Please tell which commit you want to cherry-pick.  I'd like to be sure
> > we are talking about the same thing.
> 
> OK, this commit:
> 
> commit cf9c40159a6bef5a23f7b23a45c03f0c6be57574
> Author: Alan Mackenzie <acm@muc.de>
> Date:   Mon Jan 31 17:44:59 2022 +0000
> 
>     * lisp/progmodes/cc-mode.el (c-common-init): Bind case-fold-search to nil
> 
>     This fixes bug #53605.

This is okay for the release branch.

> > Is the change you want to install the same one shown in the discussion
> > of that bug?  If not, please show the change.
> 
> Yes, it's the change I send off to the OP on 2022-06-29, asking him/her
> to test it.

Also okay.

Re: Two CC Mode bug fixes into the release branch?

[Alan Mackenzie]

Hello, Eli.

On Sat, Jul 02, 2022 at 18:26:49 +0300, Eli Zaretskii wrote:
> > Date: Sat, 2 Jul 2022 15:18:04 +0000
> > Cc: emacs-devel@gnu.org
> > From: Alan Mackenzie <acm@muc.de>

> > > Please tell which commit you want to cherry-pick.  I'd like to be sure
> > > we are talking about the same thing.

> > OK, this commit:

> > commit cf9c40159a6bef5a23f7b23a45c03f0c6be57574
> > Author: Alan Mackenzie <acm@muc.de>
> > Date:   Mon Jan 31 17:44:59 2022 +0000

> >     * lisp/progmodes/cc-mode.el (c-common-init): Bind case-fold-search to nil

> >     This fixes bug #53605.

> This is okay for the release branch.

> > > Is the change you want to install the same one shown in the discussion
> > > of that bug?  If not, please show the change.

> > Yes, it's the change I send off to the OP on 2022-06-29, asking him/her
> > to test it.

> Also okay.

Thanks.  I'll do these this afternoon.

-- 
Alan Mackenzie (Nuremberg, Germany).