address@hidden: Re: address@hidden: TAB (c-indent-command) signals a peculiar error]]

[Richard Stallman]

Would someone like to investigate this bug?
Please send me *a reply to this message* if you work on it.


------- Start of forwarded message -------
To: address@hidden
Cc: address@hidden, address@hidden, address@hidden,
        Kenichi Handa <address@hidden>
Subject: Re: address@hidden: TAB (c-indent-command) signals a peculiar
 error]
Reply-To: address@hidden
From: Martin Stjernholm <address@hidden>
Date: Tue, 17 Aug 2004 04:26:47 +0200
In-Reply-To: <address@hidden> (Richard Stallman's
 message of "Wed, 04 Aug 2004 00:32:52 -0400")
X-Virus-Scanned: by amavisd 0.1
X-Spam-Status: No, hits=-2.5 required=5.0
        tests=EMAIL_ATTRIBUTION,IN_REP_TO,QUOTED_EMAIL_TEXT,RCVD_IN_ORBS,
              REFERENCES,REPLY_WITH_QUOTES,USER_AGENT_GNUS_UA
        version=2.55
X-Spam-Level: 
X-Spam-Checker-Version: SpamAssassin 2.55 (1.174.2.19-2003-05-19-exp)

Richard Stallman <address@hidden> wrote:

> Would you please take a look at this?
>
> From: Kenichi Handa <address@hidden>
> Subject: TAB (c-indent-command) signals a peculiar error
> To: address@hidden
> Cc: address@hidden
> Date: Tue Aug  3 07:58:38 2004 +0200
>
> I got this bug report (testcase is simplified from the
> original report).
>
> With the latest CVS HEAD code, when you visit the attached
> file, move point to the beginning of 5th line ("int y;"),
> and hit TAB, the following error is signaled.

This is a bug in the syntax functions in the Emacs core. I've further
minimized the test case to this:

1.  Start Emacs with -q -no-site-file.
2.  Create a new buffer in C mode.
3.  Enter: À RET (
4.  Narrow the buffer to show only the open paren.
5.  Evaluate: (forward-comment -1)

An error "Error in syntax_table logic for to-the-end intervals" is
signalled. I expected the function to not move the point and return
nil, as it does if the buffer isn't narrowed.

Furthermore, I've no idea what the error means and how the caller is
supposed to handle it. Perhaps it should be an internal assertion
instead.

> When you move the cursor one line up, hit TAB, move the
> cursor one line down, and hit TAB, then it works.
>
> This doesn't happen when you are using global-font-lock-mode.

The various caches in CC Mode probably avoids the error in these
cases.
------- End of forwarded message -------