[eric@openbsd.org: c-mode syntactic analysis regression in emacs-22.1]

[eric@openbsd.org: c-mode syntactic analysis regression in emacs-22.1]

[Richard Stallman]

[I sent this message to bug-cc-modea week ago but did not get a
response, so now I am sending it to emacs-devel as well.]

Can you please DTRT and ack?

------- Start of forwarded message -------
X-Spam-Status: No, score=0.5 required=5.0 tests=DNS_FROM_RFC_ABUSE,
	UNPARSEABLE_RELAY autolearn=no version=3.1.0
Date: Sat, 24 Nov 2007 18:04:41 +0100
From: Eric Faurot <eric@openbsd.org>
To: bug-gnu-emacs@gnu.org
Message-ID: <20071124180441.1b1351cb@thor.ekyo.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Subject: c-mode syntactic analysis regression in emacs-22.1

Hello,

emacs-22.1 does not indent the following excerpt correctly because the syntactic
analyser tags the members appearing after "abaz" as k&r arguments.

==========
#define FOO(t) struct { struct t tv; struct t *tp; }

struct baz {
  int nothing;
};

struct bar {       /* ((topmost-intro 84)) */
  FOO(baz) abaz;   /* ((inclass 88) (topmost-intro 88)) */
	   int a;  /* ((knr-argdecl 145)) !!! */
};
=========

emacs-21 works ok; it gives ((inclass . 88) (topmost-intro . 134)) for "int a;".

It is quite annoying, and I don't really know how to fix it.


Eric.
------- End of forwarded message -------

Re: [eric@openbsd.org: c-mode syntactic analysis regression in emacs-22.1]

[Glenn Morris]

Hello Alan,

Are you able to look into this issue? It is one of the few bugs
remaining in FOR-RELEASE.


> From: Eric Faurot <eric@openbsd.org>
> Subject: c-mode syntactic analysis regression in emacs-22.1
> To: bug-gnu-emacs@gnu.org
> Date: Sat, 24 Nov 2007 18:04:41 +0100
> X-Sent: 6 weeks, 1 day, 2 hours, 33 minutes, 8 seconds ago
>
> Hello,
>
> emacs-22.1 does not indent the following excerpt correctly because the syntactic
> analyser tags the members appearing after "abaz" as k&r arguments.
>
> ==========
> #define FOO(t) struct { struct t tv; struct t *tp; }
>
> struct baz {
>   int nothing;
> };
>
> struct bar {       /* ((topmost-intro 84)) */
>   FOO(baz) abaz;   /* ((inclass 88) (topmost-intro 88)) */
> 	   int a;  /* ((knr-argdecl 145)) !!! */
> };
> =========
>
> emacs-21 works ok; it gives ((inclass . 88) (topmost-intro . 134)) for "int a;".

Re: [eric@openbsd.org: c-mode syntactic analysis regression in emacs-22.1]

[Alan Mackenzie]

Hello, Richard and Glenn,

Yes, I'm looking at this now!

On Sun, Jan 06, 2008 at 02:39:21PM -0500, Glenn Morris wrote:
> 
> Hello Alan,
> 
> Are you able to look into this issue? It is one of the few bugs
> remaining in FOR-RELEASE.
> 
> 
> > From: Eric Faurot <eric@openbsd.org>
> > Subject: c-mode syntactic analysis regression in emacs-22.1
> > To: bug-gnu-emacs@gnu.org
> > Date: Sat, 24 Nov 2007 18:04:41 +0100
> > X-Sent: 6 weeks, 1 day, 2 hours, 33 minutes, 8 seconds ago
> >
> > Hello,
> >
> > emacs-22.1 does not indent the following excerpt correctly because the syntactic
> > analyser tags the members appearing after "abaz" as k&r arguments.
> >
> > ==========
> > #define FOO(t) struct { struct t tv; struct t *tp; }
> >
> > struct baz {
> >   int nothing;
> > };
> >
> > struct bar {       /* ((topmost-intro 84)) */
> >   FOO(baz) abaz;   /* ((inclass 88) (topmost-intro 88)) */
> > 	   int a;  /* ((knr-argdecl 145)) !!! */
> > };
> > =========
> >
> > emacs-21 works ok; it gives ((inclass . 88) (topmost-intro . 134)) for "int a;".
> 

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: [eric@openbsd.org: c-mode syntactic analysis regression in emacs-22.1]

[Alan Mackenzie]

[No apologies for top-posting!]

Hallo, Glenn,

I've just commited a patch to cc-engine.el (both in the trunk and
Emacs-22 branch) which fixes this.

How soon is Emacs 22.2 going to happen?  (I've got one or two other bug
fixes in CC Mode to transfer to savannah.)

On Sun, Jan 06, 2008 at 02:39:21PM -0500, Glenn Morris wrote:
> 
> Hello Alan,
> 
> Are you able to look into this issue? It is one of the few bugs
> remaining in FOR-RELEASE.
> 
> 
> > From: Eric Faurot <eric@openbsd.org>
> > Subject: c-mode syntactic analysis regression in emacs-22.1
> > To: bug-gnu-emacs@gnu.org
> > Date: Sat, 24 Nov 2007 18:04:41 +0100
> > X-Sent: 6 weeks, 1 day, 2 hours, 33 minutes, 8 seconds ago
> >
> > Hello,
> >
> > emacs-22.1 does not indent the following excerpt correctly because the syntactic
> > analyser tags the members appearing after "abaz" as k&r arguments.
> >
> > ==========
> > #define FOO(t) struct { struct t tv; struct t *tp; }
> >
> > struct baz {
> >   int nothing;
> > };
> >
> > struct bar {       /* ((topmost-intro 84)) */
> >   FOO(baz) abaz;   /* ((inclass 88) (topmost-intro 88)) */
> > 	   int a;  /* ((knr-argdecl 145)) !!! */
> > };
> > =========
> >
> > emacs-21 works ok; it gives ((inclass . 88) (topmost-intro . 134)) for "int a;".

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: [eric@openbsd.org: c-mode syntactic analysis regression in emacs-22.1]

[Glenn Morris]

Alan Mackenzie wrote:

> I've just commited a patch to cc-engine.el (both in the trunk and
> Emacs-22 branch) which fixes this.

Thank you very much, that's great.

> How soon is Emacs 22.2 going to happen?

It's impossible to say. It seems to depend on FOR-RELEASE becoming
empty of bugs before someone happens to report a new bug of any kind.

There is one remaining bug (`pos-visible-in-window-p at ellipsis')
that I have looked at but don't know how to fix. :( Someone familiar
with the display code would need to add ellipsis handling to
pos_visible_p in xdisp.c.

Re: [eric@openbsd.org: c-mode syntactic analysis regression in emacs-22.1]

[Richard Stallman]

    It's impossible to say. It seems to depend on FOR-RELEASE becoming
    empty of bugs before someone happens to report a new bug of any kind.

Very few bugs have been added in the past couple of months.
Unfortunately people have not fixed many either.  Would people
please turn their attention to fixing the Emacs 22 bugs?

Also, who would like to take charge of Emacs 22 pretesting?

Re: [eric@openbsd.org: c-mode syntactic analysis regression in emacs-22.1]

[Chong Yidong]

Richard Stallman <rms@gnu.org> writes:

>     It's impossible to say. It seems to depend on FOR-RELEASE becoming
>     empty of bugs before someone happens to report a new bug of any kind.
>
> Very few bugs have been added in the past couple of months.
> Unfortunately people have not fixed many either.  Would people
> please turn their attention to fixing the Emacs 22 bugs?
>
> Also, who would like to take charge of Emacs 22 pretesting?

I can do the Emacs 22.2 pretesting, whenever it begins.