fix the sscanf usage in etags

fix the sscanf usage in etags

[Masatake YAMATO]

Could someone install this patch?

I have some troubles in my cvs settings.
Thanks in advance.
 
2006-08-07  Masatake YAMATO  <jet@gyve.org>

	* etags.c (readline): expect sscanf returns 2, 
	not 1.

--- orig/lib-src/etags.c
+++ mod/lib-src/etags.c
@@ -6296,7 +6296,7 @@
 	  int start, lno;
 
 	  if (DEBUG) start = 0;	/* shut up the compiler */
-	  if (sscanf (lbp->buffer, "#line %d \"%n", &lno, &start) == 1)
+	  if (sscanf (lbp->buffer, "#line %d \"%n", &lno, &start) == 2)
 	    {
 	      char *endp = lbp->buffer + start;

Re: fix the sscanf usage in etags

[Jan Djärv]

Masatake YAMATO skrev:
> Could someone install this patch?
> 

Done.

	Jan D.

> I have some troubles in my cvs settings.
> Thanks in advance.
>  
> 2006-08-07  Masatake YAMATO  <jet@gyve.org>
> 
> 	* etags.c (readline): expect sscanf returns 2, 
> 	not 1.
> 
> --- orig/lib-src/etags.c
> +++ mod/lib-src/etags.c
> @@ -6296,7 +6296,7 @@
>  	  int start, lno;
>  
>  	  if (DEBUG) start = 0;	/* shut up the compiler */
> -	  if (sscanf (lbp->buffer, "#line %d \"%n", &lno, &start) == 1)
> +	  if (sscanf (lbp->buffer, "#line %d \"%n", &lno, &start) == 2)
>  	    {
>  	      char *endp = lbp->buffer + start;
>  
> 
> 
> _______________________________________________
> Emacs-devel mailing list
> Emacs-devel@gnu.org
> http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: fix the sscanf usage in etags

[Kevin Ryde]

Masatake YAMATO <jet@gyve.org> writes:
>
> -	  if (sscanf (lbp->buffer, "#line %d \"%n", &lno, &start) == 1)
> +	  if (sscanf (lbp->buffer, "#line %d \"%n", &lno, &start) == 2)

I think that's wrong, I believe %n is not included in the return
count.  Eg. in the current glibc, and in the posix spec (under "n"),

    http://www.opengroup.org/onlinepubs/007904975/functions/scanf.html

It does look like there's a problem with degenerate input having
"start" used uninitialized though.  Eg. say

	#line 00000000000000000000000000000

Re: fix the sscanf usage in etags

[Jan Djärv]

Kevin Ryde skrev:
> Masatake YAMATO <jet@gyve.org> writes:
>> -	  if (sscanf (lbp->buffer, "#line %d \"%n", &lno, &start) == 1)
>> +	  if (sscanf (lbp->buffer, "#line %d \"%n", &lno, &start) == 2)
> 
> I think that's wrong, I believe %n is not included in the return
> count.  Eg. in the current glibc, and in the posix spec (under "n"),
> 
>     http://www.opengroup.org/onlinepubs/007904975/functions/scanf.html
> 

The man page on GNU/Linux says:
               The C standard says: "Execution of  a  %n  directive
               does  not increment the assignment count returned at the comple-
               tion of execution" but the Corrigendum seems to contradict this.
               Probably it is wise not to make any assumptions on the effect of
               %n conversions on the return value.

So I guess we safely can test for >= 1 instead for those implementations that 
do return 2.  I've checked in that change.

> It does look like there's a problem with degenerate input having
> "start" used uninitialized though.  Eg. say
> 
> 	#line 00000000000000000000000000000

We can switch places between \" and %n and check for ":

	  if (sscanf (lbp->buffer, "#line %d %n\"", &lno, &start) >= 1
	      && lbp->buffer[start] == '"')
	    {
	      char *endp = lbp->buffer + ++start;

	Jan D.

Re: fix the sscanf usage in etags

[Masatake YAMATO]

This is the file what I got trouble without my patch.

boost_1_33_1/libs/wave/test/testwave/testfiles/t_6_062.cpp:
/*=============================================================================
    Boost.Wave: A Standard compliant C++ preprocessor library
    http://www.boost.org/

    Copyright (c) 2001-2005 Hartmut Kaiser. Distributed under the Boost
    Software License, Version 1.0. (See accompanying file
    LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)

    The tests included in this file were initially taken from the mcpp V2.5
    preprocessor validation suite and were modified to fit into the Boost.Wave 
    unit test requirements.
    The original files of the mcpp preprocessor are distributed under the 
    license reproduced at the end of this file.
=============================================================================*/

// Tests error reporting: #line error.

// 7.4: string literal in #line directive shall be a character string
//      literal.
//E t_6_062.cpp(21): warning: ill formed #line directive: 123 L"wide"
#line 123 L"wide"

/*-
 * Copyright (c) 1998, 2002-2005 Kiyoshi Matsui <kmatsui@t3.rim.or.jp>
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 */

Re: fix the sscanf usage in etags

[Jan Djärv]

Masatake YAMATO skrev:
> This is the file what I got trouble without my patch.
> 

...

> // 7.4: string literal in #line directive shall be a character string
> //      literal.
> //E t_6_062.cpp(21): warning: ill formed #line directive: 123 L"wide"
> #line 123 L"wide"
> 

I don't know if etags supports wide characters at all.  A line directive like 
that is not allowed by ANSI-C as the comment says.

	Jan D.

Re: fix the sscanf usage in etags

[Masatake YAMATO]

> Masatake YAMATO skrev:
> > This is the file what I got trouble without my patch.
> > 
> 
> ...
> 
> > // 7.4: string literal in #line directive shall be a character string
> > //      literal.
> > //E t_6_062.cpp(21): warning: ill formed #line directive: 123 L"wide"
> > #line 123 L"wide"
> > 
> 
> I don't know if etags supports wide characters at all.  A line directive like 
> that is not allowed by ANSI-C as the comment says.

I see. In this case, etags should ignore the line.
The problem I found was that etags crashed when `#line 123 L"wide"' was given
as input. If the crashing can be suppressed, it is enough.

> We can switch places between \" and %n and check for ":
> 
> 	  if (sscanf (lbp->buffer, "#line %d %n\"", &lno, &start) >= 1
> 	      && lbp->buffer[start] == '"')
> 	    {
> 	      char *endp = lbp->buffer + ++start;

This works fine.
I tested with:

const char * buffer = "#line 123                                      L\"wide\"";
const char * buffer = "#line 123 L\"wide\"";
const char * buffer = "#line 123L\"wide\"";

Masatake YAMATO

Re: fix the sscanf usage in etags

[Kevin Ryde]

Masatake YAMATO <jet@gyve.org> writes:
>
> The problem I found was that etags crashed when `#line 123 L"wide"' was given
> as input. If the crashing can be suppressed, it is enough.

That's the only bad case, you don't have some strange libc that
incorrectly counts %n?

Jan Djärv <jan.h.d@swipnet.se> writes:
>
> 	  if (sscanf (lbp->buffer, "#line %d %n\"", &lno, &start) >= 1
> 	      && lbp->buffer[start] == '"')

I think that still uses start uninitialized on say "#line 123 ".  I'd
suggest (untested) something like

	start = -1;
	if (sscanf (lbp->buffer, "#line %d \"%n", &lno, &start) == 1
	    && start != -1)
	  ...

Re: fix the sscanf usage in etags

[Jan Djärv]

Kevin Ryde skrev:
> Masatake YAMATO <jet@gyve.org> writes:
>> The problem I found was that etags crashed when `#line 123 L"wide"' was given
>> as input. If the crashing can be suppressed, it is enough.
> 
> That's the only bad case, you don't have some strange libc that
> incorrectly counts %n?
> 
> Jan Djärv <jan.h.d@swipnet.se> writes:
>> 	  if (sscanf (lbp->buffer, "#line %d %n\"", &lno, &start) >= 1
>> 	      && lbp->buffer[start] == '"')
> 
> I think that still uses start uninitialized on say "#line 123 ".

No, in that case start will be 10. If the %d matches, sscanf will return 1 and 
in that case %n must match (whitespace is optional).

	Jan D.