From cbfalconer at yahoo.com  Fri Apr  6 21:54:22 2007
From: cbfalconer at yahoo.com (CBFalconer)
Date: Fri Apr  6 21:53:30 2007
Subject: [splint-discuss] How does splint check a pointer to constant
References: <e628adfd0703060752x64403368x6f8e25f2264e737c@mail.gmail.com>
Message-ID: <461723FE.B4839BCC@yahoo.com>

Jorge Peixoto de Morais Neto wrote:
> 
... snip ...
> 
> #include <stdlib.h>
> int main(void){
>   /*@unused@*/ char const * const teststringarray[]={"a"};
>   return EXIT_SUCCESS;
> }
> , splint warns me that
> 
> test.c: (in function main)
> test.c:3:54: Read-only string literal storage used as initial
> value for
>                 unqualified storage: teststringarray[0] = "a"
>   A read-only string literal is assigned to a non-observer
> reference. (Use
>   -readonlytrans to inhibit warning)
> 
> , while gcc warns me of nothing (except that the variable
> teststringarray is unused).
> 
> The behavior in the program is bugging me (I have similar case in
> a real program). Why does this happen? There is nothing wrong or
> suspect in initializing with a string literal an array of const
> pointers to constants (const * const) How can I fix this? In the
> manual, I only found something about /*@observer@*/, but this
> seems absurd, since I am not dealing with an abstract type.

Add "-Wwrite-strings" to the gcc call, and it should warn you.

-- 
 <http://www.cs.auckland.ac.nz/~pgut001/pubs/vista_cost.txt>
 <http://www.securityfocus.com/columnists/423>
 
 "A man who is right every time is not likely to do very much."
                           -- Francis Crick, co-discover of DNA
 "There is nothing more amazing than stupidity in action."
                                             -- Thomas Matthews


From ibemed at gmail.com  Mon Apr  9 15:39:59 2007
From: ibemed at gmail.com (=?ISO-8859-1?Q?Ib=E9ria_Medeiros?=)
Date: Mon Apr  9 15:40:10 2007
Subject: [splint-discuss] Porting 32 to 64 bits
Message-ID: <80bfbdd10704091539r19e447d5l6448da79ec24e45b@mail.gmail.com>

Hi,

I'am newbie with splint... and i like use the splint to give me warnings
when i porting the 32 to 64 bits.
The lint tool has a flag -errchk=longptr64 to do that, where warning when
int = long
int = (cast) ptr

With splint, how i do that??
I know that -ignorequals flag in splint, catch the first case (int = long)
and what a flag to second case??

Example:

int a, c;
long b;

a = b; // -ignorequals warning here (int = long)
c = 10;
a = (int) &c; // what is the flag? (int = ptr)


any help
thanks,
-- 
Ib?ria Medeiros
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.Virginia.EDU/pipermail/splint-discuss/attachments/20070409/76050e59/attachment.htm
From ibemed at gmail.com  Mon Apr  9 15:49:09 2007
From: ibemed at gmail.com (=?ISO-8859-1?Q?Ib=E9ria_Medeiros?=)
Date: Mon Apr  9 15:49:20 2007
Subject: [splint-discuss] parse error
Message-ID: <80bfbdd10704091549p2c126c8aq4da4a6e215a347c6@mail.gmail.com>

hi,

when i use the splint in the next code

int a;
int c;
a = c;
int b;
b = a;

the splint return a parse error.... i supose, maybe, because the declaration
of _b_ is after a instruction (a = c). Because, if i declare _b_ before that
instruction, splint dont give me any parse error.
Why splint dont work with  declarations of variables in any part of the
program?? Exist any flag in splint to acept that??

Thanks,
-- 
Ib?ria Medeiros
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.Virginia.EDU/pipermail/splint-discuss/attachments/20070409/cfddf8b0/attachment.htm
From ai2097 at users.sourceforge.net  Mon Apr  9 17:38:07 2007
From: ai2097 at users.sourceforge.net (Travis Snoozy)
Date: Mon Apr  9 17:38:23 2007
Subject: [splint-discuss] parse error
In-Reply-To: <80bfbdd10704091549p2c126c8aq4da4a6e215a347c6@mail.gmail.com>
References: <80bfbdd10704091549p2c126c8aq4da4a6e215a347c6@mail.gmail.com>
Message-ID: <20070409173807.06538cd5@localhost>

On Mon, 9 Apr 2007 22:49:09 +0000, "Ib?ria Medeiros" <ibemed@gmail.com>
wrote:

<snip>
> the splint return a parse error.... i supose, maybe, because the
> declaration of _b_ is after a instruction (a = c). Because, if i
> declare _b_ before that instruction, splint dont give me any parse
> error. Why splint dont work with  declarations of variables in any
> part of the program?? Exist any flag in splint to acept that??
<snip>

In ANSI C (C89), variable declarations must come first in your code. If
you are using GCC, you can use the -ansi -pedantic set of
switches to flag this behavior. At present, splint works only with C89,
not C99.

-- 
Travis

From ai2097 at users.sourceforge.net  Mon Apr  9 17:38:07 2007
From: ai2097 at users.sourceforge.net (Travis Snoozy)
Date: Mon Apr  9 17:38:25 2007
Subject: [splint-discuss] parse error
In-Reply-To: <80bfbdd10704091549p2c126c8aq4da4a6e215a347c6@mail.gmail.com>
References: <80bfbdd10704091549p2c126c8aq4da4a6e215a347c6@mail.gmail.com>
Message-ID: <20070409173807.06538cd5@localhost>

On Mon, 9 Apr 2007 22:49:09 +0000, "Ib?ria Medeiros" <ibemed@gmail.com>
wrote:

<snip>
> the splint return a parse error.... i supose, maybe, because the
> declaration of _b_ is after a instruction (a = c). Because, if i
> declare _b_ before that instruction, splint dont give me any parse
> error. Why splint dont work with  declarations of variables in any
> part of the program?? Exist any flag in splint to acept that??
<snip>

In ANSI C (C89), variable declarations must come first in your code. If
you are using GCC, you can use the -ansi -pedantic set of
switches to flag this behavior. At present, splint works only with C89,
not C99.

-- 
Travis

From ibemed at gmail.com  Mon Apr  9 17:51:51 2007
From: ibemed at gmail.com (=?ISO-8859-1?Q?Ib=E9ria_Medeiros?=)
Date: Mon Apr  9 17:51:58 2007
Subject: [splint-discuss] parse error
In-Reply-To: <20070409173807.06538cd5@localhost>
References: <80bfbdd10704091549p2c126c8aq4da4a6e215a347c6@mail.gmail.com>
	<20070409173807.06538cd5@localhost>
Message-ID: <80bfbdd10704091751u47d158e9n3a759068c005ee6c@mail.gmail.com>

okay,
if i understand, in first place i have to compile the program with C89
specifications (-ansi -pedantic) and after use the splint

thanks,

On 4/10/07, Travis Snoozy <ai2097@users.sourceforge.net> wrote:
>
> On Mon, 9 Apr 2007 22:49:09 +0000, "Ib?ria Medeiros" <ibemed@gmail.com>
> wrote:
>
> <snip>
> > the splint return a parse error.... i supose, maybe, because the
> > declaration of _b_ is after a instruction (a = c). Because, if i
> > declare _b_ before that instruction, splint dont give me any parse
> > error. Why splint dont work with  declarations of variables in any
> > part of the program?? Exist any flag in splint to acept that??
> <snip>
>
> In ANSI C (C89), variable declarations must come first in your code. If
> you are using GCC, you can use the -ansi -pedantic set of
> switches to flag this behavior. At present, splint works only with C89,
> not C99.
>
> --
> Travis
>
> _______________________________________________
> splint-discuss mailing list
> splint-discuss@ares.cs.Virginia.EDU
> http://www.cs.Virginia.EDU/mailman-2.1.5/listinfo/splint-discuss
>



-- 
Ib?ria Medeiros
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.Virginia.EDU/pipermail/splint-discuss/attachments/20070410/9b565719/attachment.htm
From cbfalconer at yahoo.com  Mon Apr  9 17:08:40 2007
From: cbfalconer at yahoo.com (CBFalconer)
Date: Mon Apr  9 17:54:06 2007
Subject: [splint-discuss] parse error
References: <80bfbdd10704091549p2c126c8aq4da4a6e215a347c6@mail.gmail.com>
Message-ID: <461AD588.59CE0429@yahoo.com>

Ib?ria Medeiros wrote:
> 
>    Part 1.1.1    Type: Plain Text (text/plain)
>              Encoding: quoted-printable
> 
>    Part 1.2    Type: Plain Text (text/plain)
>            Encoding: 7bit

Please do not use html in newsgroups or mailing lists.

-- 
 <http://www.cs.auckland.ac.nz/~pgut001/pubs/vista_cost.txt>
 <http://www.securityfocus.com/columnists/423>
 <http://www.aaxnet.com/editor/edit043.html>
 
 "A man who is right every time is not likely to do very much."
                           -- Francis Crick, co-discover of DNA
 "There is nothing more amazing than stupidity in action."
                                             -- Thomas Matthews


From Michael.Wojcik at MicroFocus.com  Tue Apr 10 05:54:44 2007
From: Michael.Wojcik at MicroFocus.com (Michael Wojcik)
Date: Tue Apr 10 05:56:04 2007
Subject: [splint-discuss] parse error
In-Reply-To: <80bfbdd10704091751u47d158e9n3a759068c005ee6c@mail.gmail.com>
References: <80bfbdd10704091549p2c126c8aq4da4a6e215a347c6@mail.gmail.com><20070409173807.06538cd5@localhost>
	<80bfbdd10704091751u47d158e9n3a759068c005ee6c@mail.gmail.com>
Message-ID: <11352F9641010A418AD5057945A3A6598B41D5@MTV-EXCHANGE.microfocus.com>

> From: splint-discuss-bounces@cs.virginia.edu [mailto:splint-discuss-bounces@cs.virginia.edu] On Behalf Of Ib?ria Medeiros
> Sent: Monday, 09 April, 2007 20:52
	
> if i understand, in first place i have to compile the program with > C89 specifications (-ansi -pedantic) and after use the splint

No, you have to write your program to the C89 specification (ISO 9899:1989).  Splint doesn't care how you compile it.

The "-ansi" and "-pedantic" options for GCC will tell you if your program uses certain constructs that are not allowed in C89.  You may find them useful, if you're writing C, using splint, and compiling with GCC.  But they have no effect on splint itself.

-- 
Michael Wojcik
Principal Software Systems Developer, Micro Focus

From Michael.Wojcik at MicroFocus.com  Tue Apr 10 05:54:52 2007
From: Michael.Wojcik at MicroFocus.com (Michael Wojcik)
Date: Tue Apr 10 05:59:32 2007
Subject: [splint-discuss] Porting 32 to 64 bits
In-Reply-To: <80bfbdd10704091539r19e447d5l6448da79ec24e45b@mail.gmail.com>
References: <80bfbdd10704091539r19e447d5l6448da79ec24e45b@mail.gmail.com>
Message-ID: <11352F9641010A418AD5057945A3A6598B41D3@MTV-EXCHANGE.microfocus.com>

> From: splint-discuss-bounces@cs.virginia.edu [mailto:splint-discuss-bounces@cs.virginia.edu] On Behalf Of Ib?ria Medeiros
> 	Sent: Monday, 09 April, 2007 18:40
	
> I'am newbie with splint... and i like use the splint to give me warnings when i porting the 32 to 64 bits.
> The lint tool has a flag -errchk=longptr64 to do that, where warning when
> int = long
> int = (cast) ptr 

*Some* lint tools have that flag.  There are many lint implementations.
	
> With splint, how i do that??

As far as I know, there's currently no splint check for assignment of a pointer type to an integer type with a (proper) cast.  If you cast, splint generally assumes you know what you're doing.

-- 
Michael Wojcik
Principal Software Systems Developer, Micro Focus

From ibemed at gmail.com  Tue Apr 10 06:43:30 2007
From: ibemed at gmail.com (=?ISO-8859-1?Q?Ib=E9ria_Medeiros?=)
Date: Tue Apr 10 06:43:52 2007
Subject: [splint-discuss] Porting 32 to 64 bits
In-Reply-To: <11352F9641010A418AD5057945A3A6598B41D3@MTV-EXCHANGE.microfocus.com>
References: <80bfbdd10704091539r19e447d5l6448da79ec24e45b@mail.gmail.com>
	<11352F9641010A418AD5057945A3A6598B41D3@MTV-EXCHANGE.microfocus.com>
Message-ID: <80bfbdd10704100643n6441fe69k1ff7f790d250d23@mail.gmail.com>

On 4/10/07, Michael Wojcik <Michael.Wojcik@microfocus.com> wrote:
> > From: splint-discuss-bounces@cs.virginia.edu [mailto:splint-discuss-bounces@cs.virginia.edu] On Behalf Of Ib?ria Medeiros
> >       Sent: Monday, 09 April, 2007 18:40
>
> > I'am newbie with splint... and i like use the splint to give me warnings when i porting the 32 to 64 bits.
> > The lint tool has a flag -errchk=longptr64 to do that, where warning when
> > int = long
> > int = (cast) ptr
>
> *Some* lint tools have that flag.  There are many lint implementations.
>
> > With splint, how i do that??
>
> As far as I know, there's currently no splint check for assignment of a pointer type to an integer type with a (proper) cast.  If you cast, splint generally assumes you know what you're doing.

yes, this is true.. if iam development applications to 32 bits (data
model ILP32), where sizeof(int) = sizeof(long) = sizeof(ptr) = 32
bits.
If i build programs to 64 bits, to data model LP64, the size of
integral types are diferents... sizeof(int) = 32bits and sizeof(long)
= sizeof(ptr) = 64bits.

The problem is when i porting 32 bits to 64 bits, the instructions
int = long (32 = 64 bits, in LP64) lose bits, and splint with
-ignorequals set warning me
int = ptr or int = (cast)ptr (32 = 64 bits, in LP64), splint dont
warning! There is any flag in splint that detect this?

Thanks,
-- 
Ib?ria Medeiros

From Michael.Wojcik at MicroFocus.com  Tue Apr 10 07:44:59 2007
From: Michael.Wojcik at MicroFocus.com (Michael Wojcik)
Date: Tue Apr 10 07:46:37 2007
Subject: [splint-discuss] Porting 32 to 64 bits
In-Reply-To: <80bfbdd10704100643n6441fe69k1ff7f790d250d23@mail.gmail.com>
References: <80bfbdd10704091539r19e447d5l6448da79ec24e45b@mail.gmail.com><11352F9641010A418AD5057945A3A6598B41D3@MTV-EXCHANGE.microfocus.com>
	<80bfbdd10704100643n6441fe69k1ff7f790d250d23@mail.gmail.com>
Message-ID: <11352F9641010A418AD5057945A3A6598B41D8@MTV-EXCHANGE.microfocus.com>

> From: splint-discuss-bounces@cs.virginia.edu 
> [mailto:splint-discuss-bounces@cs.virginia.edu] On Behalf Of 
> Ib?ria Medeiros
> Sent: Tuesday, 10 April, 2007 09:44
> 
> On 4/10/07, Michael Wojcik <Michael.Wojcik@microfocus.com> wrote:
> > As far as I know, there's currently no splint check for 
> > assignment of a pointer type to an integer type with a 
> > (proper) cast.  If you cast, splint generally assumes you 
> > know what you're doing.
> 
> The problem is when i porting 32 bits to 64 bits, the instructions
> int = long (32 = 64 bits, in LP64) lose bits, and splint with
> -ignorequals set warning me

Yes, I understand all of that.  I'm well aware of the issues with converting integral types in various 64-bit models.

> int = ptr or int = (cast)ptr (32 = 64 bits, in LP64), splint dont
> warning! There is any flag in splint that detect this?

Not if you use the cast.  That was the point of my previous message.

Without the cast, splint will warn about the assignment by default.  You have to specify "-type" to get splint to not warn about it.

-- 
Michael Wojcik
Principal Software Systems Developer, Micro Focus

From spam_account at sympatico.ca  Tue Apr 10 18:16:42 2007
From: spam_account at sympatico.ca (spam_account@sympatico.ca)
Date: Tue Apr 10 18:59:11 2007
Subject: [splint-discuss] CVS access to sourceforge? 
Message-ID: <m2d52b7ls5.fsf@sympatico.ca>


Could I have CVS access to source forge to accept patches posted to
this mailing list?

-- 
Schrodinger on QED: I don't like it, and I'm sorry I ever had anything
to do with it.

From elfring at users.sourceforge.net  Wed Apr 11 09:12:01 2007
From: elfring at users.sourceforge.net (SF Markus Elfring)
Date: Wed Apr 11 09:12:20 2007
Subject: [splint-discuss] CVS access to sourceforge?
In-Reply-To: <m2d52b7ls5.fsf@sympatico.ca>
References: <m2d52b7ls5.fsf@sympatico.ca>
Message-ID: <461D08D1.2090000@users.sourceforge.net>


> Could I have CVS access to source forge to accept patches posted to
> this mailing list?

The published link 
"http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/splint/splint" does not 
work any more.

Will a Subversion repository become available to improve the software 
situation?

Regards,
Markus
From spam_account at sympatico.ca  Wed Apr 11 12:37:50 2007
From: spam_account at sympatico.ca (Bill Pringlemeir)
Date: Wed Apr 11 12:38:33 2007
Subject: [splint-discuss] CVS access to sourceforge?
In-Reply-To: <461D08D1.2090000@users.sourceforge.net>
References: <m2d52b7ls5.fsf@sympatico.ca>
	<461D08D1.2090000@users.sourceforge.net>
Message-ID: <m2lkgyya5t.fsf@sympatico.ca>

On 11 Apr 2007, elfring@users.sourceforge.net wrote:

> The published link
> "http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/splint/splint" does
> not work any more.

Source forge is re-arranging things constantly.  The link is here,

  "http://splint.cvs.sourceforge.net/splint/"

This needs to change at "http://www.splint.org/".  Sourceforge does
not have rights to this server.

-- 
The end of our foundation is knowledge of causes, and secret motions
of things; and the enlarging of the bounds of human empire, to the
effecting of all things possible.  - Francis Bacon, New Atlantis

From keithp at marvell.com  Wed Apr 11 15:06:39 2007
From: keithp at marvell.com (Keith Prickett)
Date: Wed Apr 11 15:24:30 2007
Subject: [splint-discuss] Splint and Vim Errorformat
Message-ID: <D8CE9CFA707C894180BAC8CDD154F8460CF2BC01@msiexch01.marvell.com>

I am using Splint 3.1.1 and attempting to integrate it with gVIM on my
Windows box.  I have tried multiple errorformats but can't seem to find
one that parses the Splint output properly.  Has anyone had any
experience with this that can lend some insight?

I tried ":set efm=%f(%l\,%c):\ %m" among others.

This doesn't seem to work.

Thanks in advance,

--
Keith Prickett


From lholzheid at bihl-wiedemann.de  Wed Apr 11 15:42:06 2007
From: lholzheid at bihl-wiedemann.de (Ludolf Holzheid)
Date: Wed Apr 11 15:44:23 2007
Subject: [splint-discuss] Splint and Vim Errorformat
In-Reply-To: <D8CE9CFA707C894180BAC8CDD154F8460CF2BC01@msiexch01.marvell.com>
References: <D8CE9CFA707C894180BAC8CDD154F8460CF2BC01@msiexch01.marvell.com>
Message-ID: <20070411224206.GA832@svr5.bihl-wiedemann.de>

On Wed, 2007-04-11 15:06:39 -0700, Keith Prickett wrote:
> I am using Splint 3.1.1 and attempting to integrate it with gVIM on my
> Windows box.  I have tried multiple errorformats but can't seem to find
> one that parses the Splint output properly.  Has anyone had any
> experience with this that can lend some insight?

Hello Keith,

I don't know anything about vi. However, I expect vi to be able to
parse GNU-C-style error messages properly (see the
"-paren-file-format" flag to splint).

You might also experiment with "-show-column" and "-show-func".

Ludolf

-- 

---------------------------------------------------------------
Ludolf Holzheid             Tel:    +49 621 339960
Bihl+Wiedemann GmbH         Fax:    +49 621 3392239
Flo?w?rthstra?e 41          e-mail: lholzheid@bihl-wiedemann.de
D-68199 Mannheim, Germany
---------------------------------------------------------------

From Michael.Wojcik at MicroFocus.com  Thu Apr 12 10:06:55 2007
From: Michael.Wojcik at MicroFocus.com (Michael Wojcik)
Date: Thu Apr 12 10:08:22 2007
Subject: [splint-discuss] Splint and Vim Errorformat
In-Reply-To: <20070411224206.GA832@svr5.bihl-wiedemann.de>
References: <D8CE9CFA707C894180BAC8CDD154F8460CF2BC01@msiexch01.marvell.com>
	<20070411224206.GA832@svr5.bihl-wiedemann.de>
Message-ID: <11352F9641010A418AD5057945A3A6598B41F5@MTV-EXCHANGE.microfocus.com>

> From: splint-discuss-bounces@cs.virginia.edu 
> [mailto:splint-discuss-bounces@cs.virginia.edu] On Behalf Of 
> Ludolf Holzheid
> Sent: Wednesday, 11 April, 2007 18:42
> 
> On Wed, 2007-04-11 15:06:39 -0700, Keith Prickett wrote:
> > I am using Splint 3.1.1 and attempting to integrate it with gVIM on
my
> > Windows box.  I have tried multiple errorformats but can't seem to
find
> > one that parses the Splint output properly.  Has anyone had any
> > experience with this that can lend some insight?
> 
> I don't know anything about vi. However, I expect vi to be able to
> parse GNU-C-style error messages properly (see the
> "-paren-file-format" flag to splint).

-paren-file-format does indeed appear to produce error messages that vim
understands out of the box.  (Plain vi doesn't have Quickfix mode; this
is a vim extension.)

I'm no expert on setting vim errorformats (I've never looked into it
before now, actually), but the following seems to work, more or less,
for Splint without -paren-file-format:

	:set errorformat=%A%f(%l\\,%c):\ %m,%+C\ %.%#

That sets a multiline error format that begins with the line containing
the filename followed by line and column number in parentheses, and
continues with lines that begin with a space.

-- 
Michael Wojcik
Principal Software Systems Developer, Micro Focus


From keithp at marvell.com  Thu Apr 12 14:23:49 2007
From: keithp at marvell.com (Keith Prickett)
Date: Thu Apr 12 14:24:10 2007
Subject: [splint-discuss] Splint and Vim Errorformat
In-Reply-To: <11352F9641010A418AD5057945A3A6598B41F5@MTV-EXCHANGE.microfocus.com>
References: <D8CE9CFA707C894180BAC8CDD154F8460CF2BC01@msiexch01.marvell.com><20070411224206.GA832@svr5.bihl-wiedemann.de>
	<11352F9641010A418AD5057945A3A6598B41F5@MTV-EXCHANGE.microfocus.com>
Message-ID: <D8CE9CFA707C894180BAC8CDD154F8460CF97018@msiexch01.marvell.com>

:set errorformat=%A%f(%l\\,%c):\ %m,%+C\ %.%# Seems to work pretty well,
except files with windows paths are not recognized using %f.  For
example "foobar.c(line,column)" IS recognized and the error is shown
using :cn and :cp.  But the line containing
"C:\foo\bar\foobar.c(line,colum)" is NOT recognized.

Has anyone had any experience in this area?


Thanks,

--
Keith Prickett


-----Original Message-----
From: splint-discuss-bounces@cs.virginia.edu
[mailto:splint-discuss-bounces@cs.virginia.edu] On Behalf Of Michael
Wojcik
Sent: Thursday, April 12, 2007 10:07 AM
To: Discussions about the Splint annotation-assisted static
analysisproject
Subject: RE: [splint-discuss] Splint and Vim Errorformat

> From: splint-discuss-bounces@cs.virginia.edu
> [mailto:splint-discuss-bounces@cs.virginia.edu] On Behalf Of Ludolf 
> Holzheid
> Sent: Wednesday, 11 April, 2007 18:42
> 
> On Wed, 2007-04-11 15:06:39 -0700, Keith Prickett wrote:
> > I am using Splint 3.1.1 and attempting to integrate it with gVIM on
my
> > Windows box.  I have tried multiple errorformats but can't seem to
find
> > one that parses the Splint output properly.  Has anyone had any 
> > experience with this that can lend some insight?
> 
> I don't know anything about vi. However, I expect vi to be able to 
> parse GNU-C-style error messages properly (see the 
> "-paren-file-format" flag to splint).

-paren-file-format does indeed appear to produce error messages that vim
understands out of the box.  (Plain vi doesn't have Quickfix mode; this
is a vim extension.)

I'm no expert on setting vim errorformats (I've never looked into it
before now, actually), but the following seems to work, more or less,
for Splint without -paren-file-format:

	:set errorformat=%A%f(%l\\,%c):\ %m,%+C\ %.%#

That sets a multiline error format that begins with the line containing
the filename followed by line and column number in parentheses, and
continues with lines that begin with a space.

--
Michael Wojcik
Principal Software Systems Developer, Micro Focus


_______________________________________________
splint-discuss mailing list
splint-discuss@ares.cs.Virginia.EDU
http://www.cs.Virginia.EDU/mailman-2.1.5/listinfo/splint-discuss

From Michael.Wojcik at MicroFocus.com  Fri Apr 13 12:37:51 2007
From: Michael.Wojcik at MicroFocus.com (Michael Wojcik)
Date: Fri Apr 13 12:39:12 2007
Subject: [splint-discuss] Splint and Vim Errorformat
In-Reply-To: <D8CE9CFA707C894180BAC8CDD154F8460CF97018@msiexch01.marvell.com>
References: <D8CE9CFA707C894180BAC8CDD154F8460CF2BC01@msiexch01.marvell.com><20070411224206.GA832@svr5.bihl-wiedemann.de><11352F9641010A418AD5057945A3A6598B41F5@MTV-EXCHANGE.microfocus.com>
	<D8CE9CFA707C894180BAC8CDD154F8460CF97018@msiexch01.marvell.com>
Message-ID: <11352F9641010A418AD5057945A3A6598B4208@MTV-EXCHANGE.microfocus.com>

> From: splint-discuss-bounces@cs.virginia.edu 
> [mailto:splint-discuss-bounces@cs.virginia.edu] On Behalf Of 
> Keith Prickett
> Sent: Thursday, 12 April, 2007 17:24
> 
> :set errorformat=%A%f(%l\\,%c):\ %m,%+C\ %.%# Seems to work 
> pretty well,
> except files with windows paths are not recognized using %f.  For
> example "foobar.c(line,column)" IS recognized and the error is shown
> using :cn and :cp.  But the line containing
> "C:\foo\bar\foobar.c(line,colum)" is NOT recognized.

Really?  It works for me.  This is (character-mode) Vim 6.3 on XP.  I
don't have gvim installed, so I can't test with it, but I'd be surprised
if it was different.

I note that the operation of the %f (filename) token depends on the
setting of isfname, but I don't believe I've changed mine from the
default.  Here's what it is, just in case yours is different:

  isfname=@,48-57,/,\,.,-,_,+,,,#,$,%,{,},[,],:,@-@,!,~,=

Backslash and colon are both present, so full Windows pathnames should
be (and, on my system, are) recognized by %f.

-- 
Michael Wojcik
Principal Software Systems Developer, Micro Focus

From keithp at marvell.com  Mon Apr 16 20:09:45 2007
From: keithp at marvell.com (Keith Prickett)
Date: Mon Apr 16 20:09:50 2007
Subject: [splint-discuss] Help on "Storage x may become null" error
Message-ID: <D8CE9CFA707C894180BAC8CDD154F8460D08AFBC@msiexch01.marvell.com>

I've got code that looks kind of like this:

-----------------
typedef struct
{
  int a;
  int b;
  char c;
} MY_OWN_TYPE;

static MY_OWN_TYPE gAType;

void init( void )
{

  // create a list of my own types
  MY_OWN_TYPE * listOfMyOwnType[2];


  // initialize item 0 to non-null value 
  listOfMyOwnType[0] = &gAType;
  // initialize item 1 to null value
  listOfMyOwnType[1] = NULL; 


  // initializes item 1 to non-null value
  initializeItem1( &listOfMyOwnType[1] );
  // ensure value is not null
  ASSERT( NULL != listOfMyOwnType[1] );

  // pass list to function who wants list
  useMyOwnList( listOfMyOwnType );


  return;
}
----------------------------
First, Splint says: Possibly null storage listOfMyOwnType[] derivable
from parameter
     useMyOwnList (..., listOfMyOwnType, ...)...
Why would it say that?  If it's what I think it is (complaining that a
particular item _may_ be null), then how can I say that each of the
items individually can be null.

Second, Splint is reporting that Storage listOfMyOwnType[] may become
null on the line where item 0 is assigned to the address of a static
global.  Why is it complaining about that?  It seems like that would
NEVER become null.

Thanks,

--
Keith Prickett


From Francois.Isabelle at ca.kontron.com  Tue Apr 17 08:17:59 2007
From: Francois.Isabelle at ca.kontron.com (Isabelle, Francois)
Date: Tue Apr 17 08:18:26 2007
Subject: [splint-discuss] Help on "Storage x may become null" error
In-Reply-To: <D8CE9CFA707C894180BAC8CDD154F8460D08AFBC@msiexch01.marvell.com>
Message-ID: <C2866F9FC4CB034EB51A633DF16859860194363E@ssbarcelonegb>

It's not an error. As you sais, the item _may_ be null, and it instructs you to specify /*@null@*/ annotations for items that may become null. 

This can be added to return types, parameter types and  variable declaration

/*@null@*/ void*  function(/*@null@*/ void * ptr)

/*@null@*/ void* listofvoid[2];



Francois Isabelle


>-----Original Message-----
>From: splint-discuss-bounces@cs.virginia.edu
>[mailto:splint-discuss-bounces@cs.virginia.edu]On Behalf Of Keith
>Prickett
>Sent: 16 avril, 2007 20:10
>To: splint-discuss@ares.cs.Virginia.EDU
>Subject: [splint-discuss] Help on "Storage x may become null" error
>
>
>I've got code that looks kind of like this:
>
>-----------------
>typedef struct
>{
>  int a;
>  int b;
>  char c;
>} MY_OWN_TYPE;
>
>static MY_OWN_TYPE gAType;
>
>void init( void )
>{
>
>  // create a list of my own types
>  MY_OWN_TYPE * listOfMyOwnType[2];
>
>
>  // initialize item 0 to non-null value 
>  listOfMyOwnType[0] = &gAType;
>  // initialize item 1 to null value
>  listOfMyOwnType[1] = NULL; 
>
>
>  // initializes item 1 to non-null value
>  initializeItem1( &listOfMyOwnType[1] );
>  // ensure value is not null
>  ASSERT( NULL != listOfMyOwnType[1] );
>
>  // pass list to function who wants list
>  useMyOwnList( listOfMyOwnType );
>
>
>  return;
>}
>----------------------------
>First, Splint says: Possibly null storage listOfMyOwnType[] derivable
>from parameter
>     useMyOwnList (..., listOfMyOwnType, ...)...
>Why would it say that?  If it's what I think it is (complaining that a
>particular item _may_ be null), then how can I say that each of the
>items individually can be null.
>
>Second, Splint is reporting that Storage listOfMyOwnType[] may become
>null on the line where item 0 is assigned to the address of a static
>global.  Why is it complaining about that?  It seems like that would
>NEVER become null.
>
>Thanks,
>
>--
>Keith Prickett
>
>
>_______________________________________________
>splint-discuss mailing list
>splint-discuss@ares.cs.Virginia.EDU
>http://www.cs.Virginia.EDU/mailman-2.1.5/listinfo/splint-discuss
>

From Francois.Isabelle at ca.kontron.com  Tue Apr 17 08:17:59 2007
From: Francois.Isabelle at ca.kontron.com (Isabelle, Francois)
Date: Tue Apr 17 08:19:37 2007
Subject: [splint-discuss] Help on "Storage x may become null" error
In-Reply-To: <D8CE9CFA707C894180BAC8CDD154F8460D08AFBC@msiexch01.marvell.com>
Message-ID: <C2866F9FC4CB034EB51A633DF16859860194363E@ssbarcelonegb>

It's not an error. As you sais, the item _may_ be null, and it instructs you to specify /*@null@*/ annotations for items that may become null. 

This can be added to return types, parameter types and  variable declaration

/*@null@*/ void*  function(/*@null@*/ void * ptr)

/*@null@*/ void* listofvoid[2];



Francois Isabelle


>-----Original Message-----
>From: splint-discuss-bounces@cs.virginia.edu
>[mailto:splint-discuss-bounces@cs.virginia.edu]On Behalf Of Keith
>Prickett
>Sent: 16 avril, 2007 20:10
>To: splint-discuss@ares.cs.Virginia.EDU
>Subject: [splint-discuss] Help on "Storage x may become null" error
>
>
>I've got code that looks kind of like this:
>
>-----------------
>typedef struct
>{
>  int a;
>  int b;
>  char c;
>} MY_OWN_TYPE;
>
>static MY_OWN_TYPE gAType;
>
>void init( void )
>{
>
>  // create a list of my own types
>  MY_OWN_TYPE * listOfMyOwnType[2];
>
>
>  // initialize item 0 to non-null value 
>  listOfMyOwnType[0] = &gAType;
>  // initialize item 1 to null value
>  listOfMyOwnType[1] = NULL; 
>
>
>  // initializes item 1 to non-null value
>  initializeItem1( &listOfMyOwnType[1] );
>  // ensure value is not null
>  ASSERT( NULL != listOfMyOwnType[1] );
>
>  // pass list to function who wants list
>  useMyOwnList( listOfMyOwnType );
>
>
>  return;
>}
>----------------------------
>First, Splint says: Possibly null storage listOfMyOwnType[] derivable
>from parameter
>     useMyOwnList (..., listOfMyOwnType, ...)...
>Why would it say that?  If it's what I think it is (complaining that a
>particular item _may_ be null), then how can I say that each of the
>items individually can be null.
>
>Second, Splint is reporting that Storage listOfMyOwnType[] may become
>null on the line where item 0 is assigned to the address of a static
>global.  Why is it complaining about that?  It seems like that would
>NEVER become null.
>
>Thanks,
>
>--
>Keith Prickett
>
>
>_______________________________________________
>splint-discuss mailing list
>splint-discuss@ares.cs.Virginia.EDU
>http://www.cs.Virginia.EDU/mailman-2.1.5/listinfo/splint-discuss
>

From keithp at marvell.com  Tue Apr 17 13:22:07 2007
From: keithp at marvell.com (Keith Prickett)
Date: Tue Apr 17 13:22:19 2007
Subject: [splint-discuss] Help on "Storage x may become null" error
In-Reply-To: <C2866F9FC4CB034EB51A633DF16859860194363E@ssbarcelonegb>
References: <D8CE9CFA707C894180BAC8CDD154F8460D08AFBC@msiexch01.marvell.com>
	<C2866F9FC4CB034EB51A633DF16859860194363E@ssbarcelonegb>
Message-ID: <D8CE9CFA707C894180BAC8CDD154F8460D0ECE6C@msiexch01.marvell.com>

Even if I add /*@null@*/ in front of the line: "MY_OWN_TYPE *
listOfMyOwnType[2];" I still see the same warnings, plus some.  This
splint warning just doesn't make sense to me.  I have declared a list of
pointers.  The list itself will never be null.  The items in the list
may be, but how does splint want me to deal with that?


--
Keith Prickett


-----Original Message-----
From: splint-discuss-bounces@cs.virginia.edu
[mailto:splint-discuss-bounces@cs.virginia.edu] On Behalf Of Isabelle,
Francois
Sent: Tuesday, April 17, 2007 5:18 AM
To: Discussions about the Splint annotation-assisted static
analysisproject; splint-discuss@ares.cs.Virginia.EDU
Subject: RE: [splint-discuss] Help on "Storage x may become null" error

It's not an error. As you sais, the item _may_ be null, and it instructs
you to specify /*@null@*/ annotations for items that may become null. 

This can be added to return types, parameter types and  variable
declaration

/*@null@*/ void*  function(/*@null@*/ void * ptr)

/*@null@*/ void* listofvoid[2];



Francois Isabelle


>-----Original Message-----
>From: splint-discuss-bounces@cs.virginia.edu
>[mailto:splint-discuss-bounces@cs.virginia.edu]On Behalf Of Keith 
>Prickett
>Sent: 16 avril, 2007 20:10
>To: splint-discuss@ares.cs.Virginia.EDU
>Subject: [splint-discuss] Help on "Storage x may become null" error
>
>
>I've got code that looks kind of like this:
>
>-----------------
>typedef struct
>{
>  int a;
>  int b;
>  char c;
>} MY_OWN_TYPE;
>
>static MY_OWN_TYPE gAType;
>
>void init( void )
>{
>
>  // create a list of my own types
>  MY_OWN_TYPE * listOfMyOwnType[2];
>
>
>  // initialize item 0 to non-null value  listOfMyOwnType[0] = &gAType;

> // initialize item 1 to null value  listOfMyOwnType[1] = NULL;
>
>
>  // initializes item 1 to non-null value  initializeItem1( 
> &listOfMyOwnType[1] );  // ensure value is not null  ASSERT( NULL != 
> listOfMyOwnType[1] );
>
>  // pass list to function who wants list  useMyOwnList( 
> listOfMyOwnType );
>
>
>  return;
>}
>----------------------------
>First, Splint says: Possibly null storage listOfMyOwnType[] derivable 
>from parameter
>     useMyOwnList (..., listOfMyOwnType, ...)...
>Why would it say that?  If it's what I think it is (complaining that a 
>particular item _may_ be null), then how can I say that each of the 
>items individually can be null.
>
>Second, Splint is reporting that Storage listOfMyOwnType[] may become 
>null on the line where item 0 is assigned to the address of a static 
>global.  Why is it complaining about that?  It seems like that would 
>NEVER become null.
>
>Thanks,
>
>--
>Keith Prickett
>
>
>_______________________________________________
>splint-discuss mailing list
>splint-discuss@ares.cs.Virginia.EDU
>http://www.cs.Virginia.EDU/mailman-2.1.5/listinfo/splint-discuss
>

_______________________________________________
splint-discuss mailing list
splint-discuss@ares.cs.Virginia.EDU
http://www.cs.Virginia.EDU/mailman-2.1.5/listinfo/splint-discuss

From keithp at marvell.com  Tue Apr 17 13:22:07 2007
From: keithp at marvell.com (Keith Prickett)
Date: Tue Apr 17 13:51:21 2007
Subject: [splint-discuss] Help on "Storage x may become null" error
In-Reply-To: <C2866F9FC4CB034EB51A633DF16859860194363E@ssbarcelonegb>
References: <D8CE9CFA707C894180BAC8CDD154F8460D08AFBC@msiexch01.marvell.com>
	<C2866F9FC4CB034EB51A633DF16859860194363E@ssbarcelonegb>
Message-ID: <D8CE9CFA707C894180BAC8CDD154F8460D0ECE6C@msiexch01.marvell.com>

Even if I add /*@null@*/ in front of the line: "MY_OWN_TYPE *
listOfMyOwnType[2];" I still see the same warnings, plus some.  This
splint warning just doesn't make sense to me.  I have declared a list of
pointers.  The list itself will never be null.  The items in the list
may be, but how does splint want me to deal with that?


--
Keith Prickett


-----Original Message-----
From: splint-discuss-bounces@cs.virginia.edu
[mailto:splint-discuss-bounces@cs.virginia.edu] On Behalf Of Isabelle,
Francois
Sent: Tuesday, April 17, 2007 5:18 AM
To: Discussions about the Splint annotation-assisted static
analysisproject; splint-discuss@ares.cs.Virginia.EDU
Subject: RE: [splint-discuss] Help on "Storage x may become null" error

It's not an error. As you sais, the item _may_ be null, and it instructs
you to specify /*@null@*/ annotations for items that may become null. 

This can be added to return types, parameter types and  variable
declaration

/*@null@*/ void*  function(/*@null@*/ void * ptr)

/*@null@*/ void* listofvoid[2];



Francois Isabelle


>-----Original Message-----
>From: splint-discuss-bounces@cs.virginia.edu
>[mailto:splint-discuss-bounces@cs.virginia.edu]On Behalf Of Keith 
>Prickett
>Sent: 16 avril, 2007 20:10
>To: splint-discuss@ares.cs.Virginia.EDU
>Subject: [splint-discuss] Help on "Storage x may become null" error
>
>
>I've got code that looks kind of like this:
>
>-----------------
>typedef struct
>{
>  int a;
>  int b;
>  char c;
>} MY_OWN_TYPE;
>
>static MY_OWN_TYPE gAType;
>
>void init( void )
>{
>
>  // create a list of my own types
>  MY_OWN_TYPE * listOfMyOwnType[2];
>
>
>  // initialize item 0 to non-null value  listOfMyOwnType[0] = &gAType;

> // initialize item 1 to null value  listOfMyOwnType[1] = NULL;
>
>
>  // initializes item 1 to non-null value  initializeItem1( 
> &listOfMyOwnType[1] );  // ensure value is not null  ASSERT( NULL != 
> listOfMyOwnType[1] );
>
>  // pass list to function who wants list  useMyOwnList( 
> listOfMyOwnType );
>
>
>  return;
>}
>----------------------------
>First, Splint says: Possibly null storage listOfMyOwnType[] derivable 
>from parameter
>     useMyOwnList (..., listOfMyOwnType, ...)...
>Why would it say that?  If it's what I think it is (complaining that a 
>particular item _may_ be null), then how can I say that each of the 
>items individually can be null.
>
>Second, Splint is reporting that Storage listOfMyOwnType[] may become 
>null on the line where item 0 is assigned to the address of a static 
>global.  Why is it complaining about that?  It seems like that would 
>NEVER become null.
>
>Thanks,
>
>--
>Keith Prickett
>
>
>_______________________________________________
>splint-discuss mailing list
>splint-discuss@ares.cs.Virginia.EDU
>http://www.cs.Virginia.EDU/mailman-2.1.5/listinfo/splint-discuss
>

_______________________________________________
splint-discuss mailing list
splint-discuss@ares.cs.Virginia.EDU
http://www.cs.Virginia.EDU/mailman-2.1.5/listinfo/splint-discuss

From cbfalconer at yahoo.com  Tue Apr 17 14:05:21 2007
From: cbfalconer at yahoo.com (CBFalconer)
Date: Tue Apr 17 16:33:19 2007
Subject: [splint-discuss] Help on "Storage x may become null" error
References: <D8CE9CFA707C894180BAC8CDD154F8460D08AFBC@msiexch01.marvell.com>
	<C2866F9FC4CB034EB51A633DF16859860194363E@ssbarcelonegb>
	<D8CE9CFA707C894180BAC8CDD154F8460D0ECE6C@msiexch01.marvell.com>
Message-ID: <46250C61.99332843@yahoo.com>

Keith Prickett wrote:
> 
> Even if I add /*@null@*/ in front of the line: "MY_OWN_TYPE *
> listOfMyOwnType[2];" I still see the same warnings, plus some. 
> This splint warning just doesn't make sense to me.  I have
> declared a list of pointers.  The list itself will never be null. 
> The items in the list may be, but how does splint want me to deal
> with that?

Please do not top-post.  Your answer belongs after (or intermixed
with) the quoted material to which you reply, after snipping all
irrelevant material.  See the following links:

-- 
  <http://www.catb.org/~esr/faqs/smart-questions.html>
  <http://www.caliburn.nl/topposting.html>
  <http://www.netmeister.org/news/learn2quote.html>
  <http://cfaj.freeshell.org/google/>  (taming google)
  <http://members.fortunecity.com/nnqweb/>  (newusers)

From Francois.Isabelle at ca.kontron.com  Tue Apr 17 17:16:04 2007
From: Francois.Isabelle at ca.kontron.com (Isabelle, Francois)
Date: Tue Apr 17 17:16:17 2007
Subject: [splint-discuss] Help on "Storage x may become null" error
In-Reply-To: <46250C61.99332843@yahoo.com>
Message-ID: <C2866F9FC4CB034EB51A633DF16859860194364C@ssbarcelonegb>

>
>Please do not top-post.  Your answer belongs after (or intermixed
>with) the quoted material to which you reply, after snipping all
>irrelevant material.  See the following links:
>
>-- 
>  <http://www.catb.org/~esr/faqs/smart-questions.html>
>  <http://www.caliburn.nl/topposting.html>
>  <http://www.netmeister.org/news/learn2quote.html>
>  <http://cfaj.freeshell.org/google/>  (taming google)
>  <http://members.fortunecity.com/nnqweb/>  (newusers)
>

So sorry, I whish you don't spend time around looking for offending users, I guess you
could find a big number of sad outlook users.

Thanks 
Francois Isabelle

From wenzel at bbr-vt.de  Wed Apr 18 02:31:14 2007
From: wenzel at bbr-vt.de (Wenzel, Bodo)
Date: Wed Apr 18 05:39:14 2007
Subject: [splint-discuss] RE: Help on "Storage x may become null" error
Message-ID: <4625D752.28160.25C714@wenzel.bbr-vt.de>

Would you mind to show us your splint commandline, the declarations 
and annotations for the called function, and so on, so that it's 
possible to reproduce your problem, please?

Mit freundlichen Gr??en,
Bodo Wenzel
- Entwicklung Software -

-- 
BBR - Baudis Bergmann R?sch
Verkehrstechnik GmbH
Pillaustra?e 1e	
D - 38126 Braunschweig

T: +49.531.27300-766
F: +49.531.27300-999
@: wenzel@bbr-vt.de
W: http://www.bbr-vt.de

Registergericht:   
AG Braunschweig  HRB 3037

Gesch?ftsf?hrer: 
Dipl.-Ing. Arne Baudis
Dipl.-Ing. Thomas Bergmann
Dipl.-Ing. Frank-Michael R?sch

USt.-ID-Nr.:
DE 114 877 881


From keithp at marvell.com  Wed Apr 18 12:10:50 2007
From: keithp at marvell.com (Keith Prickett)
Date: Wed Apr 18 12:10:55 2007
Subject: [splint-discuss] RE: Help on "Storage x may become null" error
In-Reply-To: <4625D752.28160.25C714@wenzel.bbr-vt.de>
References: <4625D752.28160.25C714@wenzel.bbr-vt.de>
Message-ID: <D8CE9CFA707C894180BAC8CDD154F8460D166F60@msiexch01.marvell.com>

---Original Code/Message---
I've got code that looks kind of like this:

-----------------
typedef struct
{
  int a;
  int b;
  char c;
} MY_OWN_TYPE;

static MY_OWN_TYPE gAType;

void init( void )
{

  // create a list of my own types
  MY_OWN_TYPE * listOfMyOwnType[2];


  // initialize item 0 to non-null value
  listOfMyOwnType[0] = &gAType;
  // initialize item 1 to null value
  listOfMyOwnType[1] = NULL; 


  // initializes item 1 to non-null value
  initializeItem1( &listOfMyOwnType[1] );
  // ensure value is not null
  ASSERT( NULL != listOfMyOwnType[1] );

  // pass list to function who wants list
  useMyOwnList( listOfMyOwnType );


  return;
}
----------------------------

-----Wenzel, Bodo Wrote-----
From: splint-discuss-bounces@cs.virginia.edu
[mailto:splint-discuss-bounces@cs.virginia.edu] On Behalf Of Wenzel,
Bodo
Sent: Tuesday, April 17, 2007 11:31 PM
To: splint-discuss@ares.cs.Virginia.EDU
Subject: [splint-discuss] RE: Help on "Storage x may become null" error

> Would you mind to show us your splint commandline, the declarations
and annotations for the called function,
> and so on, so that it's possible to reproduce your problem, please?


I only have "splint +paren-file-format file.c" plus my include flags and
defines for my own code.  The "+paren-file-format" is because I am
running on a WIN32 system.  I think my biggest problem is that I don't
understand why splint would think my variable "listOfMyOwnType" would
EVER become NULL.  Maybe I need to review the C-language.  The above
code is a snippet of code I wrote (and didn't test/compile) to match my
scenario, but is not an exact replica of my code.

I annotated the line during testing from:
<code>
  MY_OWN_TYPE * listOfMyOwnType[2];
</code>

to:
<code>
/*@null@*/  MY_OWN_TYPE * listOfMyOwnType[2];
</code>



Before the annotation I get the warnings:
"""
Possibly null storage listOfMyOwnType[] derivable from parameter
     useMyOwnList (..., listOfMyOwnType, ...)
   A possibly null pointer is reachable from a parameter or global
variable that
   is not declared using a /*@null@*/ annotation. (Use -nullstate to
inhibit
   warning)
"""
Then:
"""
Storage listOfMyOwnType[] may become null.
"""

After I make the annotation I get the two previous warnings plus the
following:
"""
  Index of possibly null pointer listOfMyOwnType: listOfMyOwnType
   A possibly null pointer is dereferenced.  Value is either the result
of a
   function which may return null (in which case, code should check it
is not
   null), or a global, parameter or structure field declared with the
null
   qualifier. (Use -nullderef to inhibit warning)
"""

I am just not seeing where "listOfMyOwnType" may become null.  If I put
checks like:

if (NULL != listOfMyOwnType)
{
  listOfMyOwnType[0] = &gAType;
}

the warnings go away (surprise).  I have to put this test around each
use of the variable.  This doesn't seem like a reasonable solution for
the example given, that is why I sent an e-mail to the list, to figure
out a reasonable solution.



Thanks in advance,

--
Keith Prickett




From lholzheid at bihl-wiedemann.de  Wed Apr 18 16:21:28 2007
From: lholzheid at bihl-wiedemann.de (Ludolf Holzheid)
Date: Wed Apr 18 16:21:42 2007
Subject: [splint-discuss] RE: Help on "Storage x may become null" error
In-Reply-To: <D8CE9CFA707C894180BAC8CDD154F8460D166F60@msiexch01.marvell.com>
References: <4625D752.28160.25C714@wenzel.bbr-vt.de>
	<D8CE9CFA707C894180BAC8CDD154F8460D166F60@msiexch01.marvell.com>
Message-ID: <20070418202128.GB16498@svr5.bihl-wiedemann.de>

On Wed, 2007-04-18 09:10:50 -0700, Keith Prickett wrote:
> [..]
> I am just not seeing where "listOfMyOwnType" may become null.
> [..]

Keith,

I think the problem is, splint knows listOfMyOwnType[1] being NULL and
warns about this fact. However, there seems to be no code annotation
for the object a pointer points to. i.e., if you write the prototype
for initializeItem1() as follows:

 void initializeItem1 (/*@null@*/ MY_OWN_TYPE **arg);

you allow **arg to be NULL, but *arg still has to be non-NULL (unless
**art is NULL, of course).

In your case, you want initializeItem1() to _return_ *arg in **arg, so
you should explain _this_ to splint:

 void initializeItem1 (/*@out@*/ MY_OWN_TYPE **arg);

In this case, *arg may be NULL or even undefined.

HTH,

Ludolf


-- 

---------------------------------------------------------------
Ludolf Holzheid             Tel:    +49 621 339960
Bihl+Wiedemann GmbH         Fax:    +49 621 3392239
Flo?w?rthstra?e 41          e-mail: lholzheid@bihl-wiedemann.de
D-68199 Mannheim, Germany
---------------------------------------------------------------

From lholzheid at bihl-wiedemann.de  Wed Apr 18 16:21:28 2007
From: lholzheid at bihl-wiedemann.de (Ludolf Holzheid)
Date: Wed Apr 18 16:24:26 2007
Subject: [splint-discuss] RE: Help on "Storage x may become null" error
In-Reply-To: <D8CE9CFA707C894180BAC8CDD154F8460D166F60@msiexch01.marvell.com>
References: <4625D752.28160.25C714@wenzel.bbr-vt.de>
	<D8CE9CFA707C894180BAC8CDD154F8460D166F60@msiexch01.marvell.com>
Message-ID: <20070418202128.GB16498@svr5.bihl-wiedemann.de>

On Wed, 2007-04-18 09:10:50 -0700, Keith Prickett wrote:
> [..]
> I am just not seeing where "listOfMyOwnType" may become null.
> [..]

Keith,

I think the problem is, splint knows listOfMyOwnType[1] being NULL and
warns about this fact. However, there seems to be no code annotation
for the object a pointer points to. i.e., if you write the prototype
for initializeItem1() as follows:

 void initializeItem1 (/*@null@*/ MY_OWN_TYPE **arg);

you allow **arg to be NULL, but *arg still has to be non-NULL (unless
**art is NULL, of course).

In your case, you want initializeItem1() to _return_ *arg in **arg, so
you should explain _this_ to splint:

 void initializeItem1 (/*@out@*/ MY_OWN_TYPE **arg);

In this case, *arg may be NULL or even undefined.

HTH,

Ludolf


-- 

---------------------------------------------------------------
Ludolf Holzheid             Tel:    +49 621 339960
Bihl+Wiedemann GmbH         Fax:    +49 621 3392239
Flo?w?rthstra?e 41          e-mail: lholzheid@bihl-wiedemann.de
D-68199 Mannheim, Germany
---------------------------------------------------------------

From keithp at marvell.com  Wed Apr 18 18:59:06 2007
From: keithp at marvell.com (Keith Prickett)
Date: Wed Apr 18 18:59:16 2007
Subject: [splint-discuss] RE: Help on "Storage x may become null" error
In-Reply-To: <20070418202128.GB16498@svr5.bihl-wiedemann.de>
References: <4625D752.28160.25C714@wenzel.bbr-vt.de>
	<D8CE9CFA707C894180BAC8CDD154F8460D166F60@msiexch01.marvell.com>
	<20070418202128.GB16498@svr5.bihl-wiedemann.de>
Message-ID: <D8CE9CFA707C894180BAC8CDD154F8460D1CF02D@msiexch01.marvell.com>

-----From: Ludolf Holzheid [mailto:lholzheid@bihl-wiedemann.de]-----
Sent: Wednesday, April 18, 2007 1:21 PM
To: splint-discuss@ares.cs.Virginia.EDU; Discussions about the Splint
annotation-assisted static analysis project
Cc: Keith Prickett
Subject: Re: [splint-discuss] RE: Help on "Storage x may become null"
error

> On Wed, 2007-04-18 09:10:50 -0700, Keith Prickett wrote:
> > [..]
> > I am just not seeing where "listOfMyOwnType" may become null.
> > [..]
>
> I think the problem is, splint knows listOfMyOwnType[1] being NULL
> and warns about this fact. However, there seems to be no code
annotation
> for the object a pointer points to. i.e., if you write the prototype
for
> initializeItem1() as follows:

This may be true, except splint contains about the line where an
assignment to listOfMyOwnType[0] is occuring.  The Splint warning is:
Index of possibly null pointer listOfMyOwnType: listOfMyOwnType
  A possibly null pointer is dereferenced.  Value is either the result
of a
  function which may return null (in which case, code should check it is
not
  null), or a global, parameter or structure field declared with the
null
  qualifier. (Use -nullderef to inhibit warning)

This again, is the warning message when the assignment to
listOfMyOwnType[0] is occuring, not 1.

So I am still perplexed on this warning.

However, I did realize, due to your response, that splint doesn't
realize I am checking the "out" value from the "initializeItem1"
function with an ASSERT so I added code like follows and the second set
of warnings went away:
if ( (listOfMyOwnType[0] != NULL) && (listOfMyOwnType[1] != NULL))
{
  useMyOwnList( listOfMyOwnType );
}
else
{
  ...
}

I can't get Splint to recognize my "ASSERT" (instead of using "ifs" like
above) but that's for a different topic.


Thanks again and in advance,

--
Keith

From keithp at marvell.com  Wed Apr 18 18:59:06 2007
From: keithp at marvell.com (Keith Prickett)
Date: Wed Apr 18 18:59:17 2007
Subject: [splint-discuss] RE: Help on "Storage x may become null" error
In-Reply-To: <20070418202128.GB16498@svr5.bihl-wiedemann.de>
References: <4625D752.28160.25C714@wenzel.bbr-vt.de>
	<D8CE9CFA707C894180BAC8CDD154F8460D166F60@msiexch01.marvell.com>
	<20070418202128.GB16498@svr5.bihl-wiedemann.de>
Message-ID: <D8CE9CFA707C894180BAC8CDD154F8460D1CF02D@msiexch01.marvell.com>

-----From: Ludolf Holzheid [mailto:lholzheid@bihl-wiedemann.de]-----
Sent: Wednesday, April 18, 2007 1:21 PM
To: splint-discuss@ares.cs.Virginia.EDU; Discussions about the Splint
annotation-assisted static analysis project
Cc: Keith Prickett
Subject: Re: [splint-discuss] RE: Help on "Storage x may become null"
error

> On Wed, 2007-04-18 09:10:50 -0700, Keith Prickett wrote:
> > [..]
> > I am just not seeing where "listOfMyOwnType" may become null.
> > [..]
>
> I think the problem is, splint knows listOfMyOwnType[1] being NULL
> and warns about this fact. However, there seems to be no code
annotation
> for the object a pointer points to. i.e., if you write the prototype
for
> initializeItem1() as follows:

This may be true, except splint contains about the line where an
assignment to listOfMyOwnType[0] is occuring.  The Splint warning is:
Index of possibly null pointer listOfMyOwnType: listOfMyOwnType
  A possibly null pointer is dereferenced.  Value is either the result
of a
  function which may return null (in which case, code should check it is
not
  null), or a global, parameter or structure field declared with the
null
  qualifier. (Use -nullderef to inhibit warning)

This again, is the warning message when the assignment to
listOfMyOwnType[0] is occuring, not 1.

So I am still perplexed on this warning.

However, I did realize, due to your response, that splint doesn't
realize I am checking the "out" value from the "initializeItem1"
function with an ASSERT so I added code like follows and the second set
of warnings went away:
if ( (listOfMyOwnType[0] != NULL) && (listOfMyOwnType[1] != NULL))
{
  useMyOwnList( listOfMyOwnType );
}
else
{
  ...
}

I can't get Splint to recognize my "ASSERT" (instead of using "ifs" like
above) but that's for a different topic.


Thanks again and in advance,

--
Keith

From wenzel at bbr-vt.de  Thu Apr 19 02:18:15 2007
From: wenzel at bbr-vt.de (Wenzel, Bodo)
Date: Thu Apr 19 02:52:25 2007
Subject: [splint-discuss] RE: Help on "Storage x may become null" error
Message-ID: <462725C7.12683.12C9D0@wenzel.bbr-vt.de>

Well, when I asked for your source, I ment to get a _complete_ 
example. If I use just the provided code snippet, there are several 
other warnings...

test.c: (in function init)
test.c(24,3): Unrecognized identifier: initializeItem1
  Identifier used in code has not been declared. (Use -unrecog to
  inhibit warning)
test.c(26,3): Unrecognized identifier: ASSERT
test.c(29,3): Unrecognized identifier: useMyOwnList

You don't want to waste my time letting me develop the missing stuff, 
do you?

With regards,
Bodo Wenzel
- Software Development -

-- 
BBR - Baudis Bergmann R?sch
Verkehrstechnik GmbH
Pillaustra?e 1e	
D - 38126 Braunschweig

T: +49.531.27300-766
F: +49.531.27300-999
@: wenzel@bbr-vt.de
W: http://www.bbr-vt.de

Registergericht:   
AG Braunschweig  HRB 3037

Gesch?ftsf?hrer: 
Dipl.-Ing. Arne Baudis
Dipl.-Ing. Thomas Bergmann
Dipl.-Ing. Frank-Michael R?sch

USt.-ID-Nr.:
DE 114 877 881


From keithp at marvell.com  Thu Apr 19 13:25:29 2007
From: keithp at marvell.com (Keith Prickett)
Date: Thu Apr 19 13:25:43 2007
Subject: [splint-discuss] RE: Help on "Storage x may become null" error
In-Reply-To: <462725C7.12683.12C9D0@wenzel.bbr-vt.de>
References: <462725C7.12683.12C9D0@wenzel.bbr-vt.de>
Message-ID: <D8CE9CFA707C894180BAC8CDD154F8460D229F3C@msiexch01.marvell.com>

-----Original Message-----
From: splint-discuss-bounces@cs.virginia.edu
[mailto:splint-discuss-bounces@cs.virginia.edu] On Behalf Of Wenzel,
Bodo
Sent: Wednesday, April 18, 2007 11:18 PM
To: splint-discuss@cs.virginia.edu
Subject: [splint-discuss] RE: Help on "Storage x may become null" error

> Well, when I asked for your source, I ment to get a _complete_
example. If I use just the provided code snippet, there are several
other warnings...
Attached is a splint-compiling (with all default options on WIN32 i.e.
"splint bla.c") example with the 1 error I am experiencing.  Please let
me know if the attachment doesn't come through as I haven't tried to
attach an item to the list before.

> You don't want to waste my time letting me develop the missing stuff,
do you?
Thanks for your help, sorry I didn't quite realize what you wanted.

--
Keith Prickett
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bla.c
Type: application/octet-stream
Size: 1228 bytes
Desc: bla.c
Url : http://www.cs.Virginia.EDU/pipermail/splint-discuss/attachments/20070419/acdc7b9a/bla.obj
From wenzel at bbr-vt.de  Fri Apr 20 03:29:54 2007
From: wenzel at bbr-vt.de (Wenzel, Bodo)
Date: Fri Apr 20 03:30:04 2007
Subject: [splint-discuss] RE: Help on "Storage x may become null" error
Message-ID: <46288812.4031.55D469@wenzel.bbr-vt.de>

Thanks! Please don't feel offended, but we better spend time to solve 
problems, not to fill gaps ;-) Your attachment was scrubbed from the 
digest I received, but there is a link to it. I'm sorry to reply so 
slowly, but my employer limits the time spend with non-profit 
stuff...

These are my results and ideas:

First, I ran splint with the contents shown below. To reduce the 
size, I have "condensed" the source a bit... As you reported, it 
complaints about
"Null storage listOfMyOwnType[1] derivable from parameter", but on
"initializeItem1 (&listOfMyOwnType[1])", not on useMyOwnList!
It shows the correct row/column where
"Storage listOfMyOwnType[1] becomes null."

Second, I added "/*@out@*/" just before the parameter of the 
prototype of "initializeItem1", AFAIK there's no need to repeat it at 
the definition of the function. Now splint reports no warnings.

Concerning the annotation for pointers to MY_OWN_TYPE; I think if you 
add "/*@null@*/" to listOfMyOwnType, it declares that symbol as 
possibly NULL. MY_OWN_TYPE*[] it is compatible to MY_OWN_TYPE**, and 
splint seems not to distinguish these. So, to annotate the elements, 
I suggest to "typedef /*@null@*/ MY_OWN_TYPE *MY_OWN_TYPE_POINTER;" 
and to declare "MY_OWN_TYPE_POINTER listOfMyOwnType[2];". Anyway, 
initializeItem1 needs the annotation at its parameter.

You don't need to "return;" at the end of a "void" typed function. 
However, your coding convention might require this. To reduce the 
size of the exmaple source even more, the called function can be 
declared as global, and their definition omitted.

BTW, my splint says "Splint 3.1.1 --- 12 April 2003" and it's running 
on Win2K.

--- test*.c ------------------------------------------

typedef struct
{
  int a;
  int b;
  char c;
} MY_OWN_TYPE;

static MY_OWN_TYPE gAType;

static void initializeItem1( MY_OWN_TYPE ** ptrToItemInList );
static void useMyOwnList( MY_OWN_TYPE ** list );

void init( void )
{
  MY_OWN_TYPE * listOfMyOwnType[2];

  listOfMyOwnType[0] = &gAType;
  listOfMyOwnType[1] = NULL; 

  initializeItem1( &listOfMyOwnType[1] );

  useMyOwnList( listOfMyOwnType );
}

static void initializeItem1( MY_OWN_TYPE ** ptrToItemInList )
{
  static MY_OWN_TYPE item1;

  *ptrToItemInList = &item1;
}

static void useMyOwnList( MY_OWN_TYPE ** list )
{
  MY_OWN_TYPE * item;

  item = list[0];
}

With regards,
Bodo Wenzel
- Software Development -

-- 
BBR - Baudis Bergmann R?sch
Verkehrstechnik GmbH
Pillaustra?e 1e	
D - 38126 Braunschweig

T: +49.531.27300-766
F: +49.531.27300-999
@: wenzel@bbr-vt.de
W: http://www.bbr-vt.de

Registergericht:   
AG Braunschweig  HRB 3037

Gesch?ftsf?hrer: 
Dipl.-Ing. Arne Baudis
Dipl.-Ing. Thomas Bergmann
Dipl.-Ing. Frank-Michael R?sch

USt.-ID-Nr.:
DE 114 877 881


From lholzheid at bihl-wiedemann.de  Fri Apr 20 14:07:47 2007
From: lholzheid at bihl-wiedemann.de (Ludolf Holzheid)
Date: Fri Apr 20 14:08:08 2007
Subject: [splint-discuss] RE: Help on "Storage x may become null" error
In-Reply-To: <D8CE9CFA707C894180BAC8CDD154F8460D1CF02D@msiexch01.marvell.com>
	<D8CE9CFA707C894180BAC8CDD154F8460D229F3C@msiexch01.marvell.com>
References: <4625D752.28160.25C714@wenzel.bbr-vt.de>
	<D8CE9CFA707C894180BAC8CDD154F8460D166F60@msiexch01.marvell.com>
	<20070418202128.GB16498@svr5.bihl-wiedemann.de>
	<D8CE9CFA707C894180BAC8CDD154F8460D1CF02D@msiexch01.marvell.com>
	<462725C7.12683.12C9D0@wenzel.bbr-vt.de>
	<D8CE9CFA707C894180BAC8CDD154F8460D229F3C@msiexch01.marvell.com>
Message-ID: <20070420180747.GA29197@svr5.bihl-wiedemann.de>

On Wed, 2007-04-18 15:59:06 -0700, Keith Prickett wrote:
> [..]
> This may be true, except splint contains about the line where an
> assignment to listOfMyOwnType[0] is occuring.

Hello Keith,

If this is for the source you sent to Bodo Wenzel: You explicitly
annotated listOfMyOwnType as being possibly NULL, so referencing it
_must_ give a warning.

As listOfMyOwnType[2] is of storage class "auto", it will be located
on the stack and thus its address never will be NULL. I think you
wanted to annotate the _elements_ of listOfMyOwnType. Bodo Wenzel
provided a way to accomplish this by means of an intermediate type
definition.

> [..]
> I can't get Splint to recognize my "ASSERT" (instead of using "ifs" like
> above) but that's for a different topic.

I don't know how your "ASSERT" is defined, so I can't say
anything. The C99-macro "assert" (from <assert.h>) seems to work,
however.

HTH,

Ludolf

-- 

---------------------------------------------------------------
Ludolf Holzheid             Tel:    +49 621 339960
Bihl+Wiedemann GmbH         Fax:    +49 621 3392239
Flo?w?rthstra?e 41          e-mail: lholzheid@bihl-wiedemann.de
D-68199 Mannheim, Germany
---------------------------------------------------------------

From mark at glines.org  Mon Apr 23 23:54:10 2007
From: mark at glines.org (Mark Glines)
Date: Mon Apr 23 23:56:57 2007
Subject: [splint-discuss] Preprocessor output?
Message-ID: <20070423205410.2b9a6fa4@chirp>

Hi,

I am trying to run splint on the Parrot codebase (www.parrotcode.org),
and having a difficult time of it.  I have worked through several
issues involving non-ANSI code and system header problems and so forth,
but now I'm stuck.

It would be very helpful to get access to the preprocessed source that
splint is actually trying to parse.  Is there a command line option
which convinces splint to output its preprocessor output to stdout, or
stderr or a log file or something?

In other words, a splint equivalent of the gcc options '-E' and maybe
'-dD' would be very helpful.  Does such a thing exist?

Thanks,

Mark
From mark at glines.org  Mon Apr 23 23:54:10 2007
From: mark at glines.org (Mark Glines)
Date: Tue Apr 24 00:17:25 2007
Subject: [splint-discuss] Preprocessor output?
Message-ID: <20070423205410.2b9a6fa4@chirp>

Hi,

I am trying to run splint on the Parrot codebase (www.parrotcode.org),
and having a difficult time of it.  I have worked through several
issues involving non-ANSI code and system header problems and so forth,
but now I'm stuck.

It would be very helpful to get access to the preprocessed source that
splint is actually trying to parse.  Is there a command line option
which convinces splint to output its preprocessor output to stdout, or
stderr or a log file or something?

In other words, a splint equivalent of the gcc options '-E' and maybe
'-dD' would be very helpful.  Does such a thing exist?

Thanks,

Mark
From brian.quinlan at iolfree.ie  Tue Apr 24 05:08:17 2007
From: brian.quinlan at iolfree.ie (Brian Quinlan)
Date: Tue Apr 24 05:27:50 2007
Subject: [splint-discuss] Preprocessor output?
Message-ID: <462dc901.5c2.0@iolfree.ie>

>It would be very helpful to get access to the preprocessed source that
>splint is actually trying to parse.  Is there a command line option
>which convinces splint to output its preprocessor output to stdout, or
>stderr or a log file or something?
>
>In other words, a splint equivalent of the gcc options '-E' and maybe
>'-dD' would be very helpful.  Does such a thing exist?
>
Hi Mark,
Adding +keep to the splint command-line tells splint not to delete
temporary files. The temporary files contain the preprocessor output.

Bye,
Brian
*************************** ADVERTISEMENT ******************************
Get BT Broadband from only EUR20 per month! Always-on internet
& wireless option - find out more at http://www.btireland.ie/broadband
From evans at cs.virginia.edu  Tue Apr 24 08:28:38 2007
From: evans at cs.virginia.edu (David Evans)
Date: Tue Apr 24 08:29:43 2007
Subject: [splint-discuss] Preprocessor output?
In-Reply-To: <20070423205410.2b9a6fa4@chirp>
References: <20070423205410.2b9a6fa4@chirp>
Message-ID: <462DF7F6.7010607@cs.virginia.edu>

Hi Mark,

Mark Glines wrote:
 > It would be very helpful to get access to the preprocessed source that
 > splint is actually trying to parse.  Is there a command line option
 > which convinces splint to output its preprocessor output to stdout, or
 > stderr or a log file or something?

If you use the -keep option, splint will keep the temporary 
(post-pre-processor) files and print out a list of their file names at 
the end of the run.

--- Dave

From mark at glines.org  Tue Apr 24 11:42:52 2007
From: mark at glines.org (Mark Glines)
Date: Tue Apr 24 11:43:05 2007
Subject: [splint-discuss] Information cached between source files?
In-Reply-To: <462DF7F6.7010607@cs.virginia.edu>
References: <20070423205410.2b9a6fa4@chirp> <462DF7F6.7010607@cs.virginia.edu>
Message-ID: <20070424084252.52a423df@chirp>

David, Brian, thanks for your help.  That's exactly what I needed.

On to the next problem... :)

I am running into a problem where specifying multiple source files on
the command line affects the preprocessor parsing those files.  Hmm,
that sentence didn't make a lot of sense.  Let me try an example:

File 1 (compilers/imcc/cfg.c) parses fine, when I run splint on it
alone.

File 2 (compilers/imcc/debug.c) parses fine, when I run splint on it
alone.

But if I put both filenames on the same splint command-line, the second
file gets a parse error.  Upon further investigation of the preprocessor
output, it seems that an important header-file didn't get included for
the second file, causing a parse error when using a typedef
(__sighandler_t) which never actually got typedeffed.

Is this expected behavior?  Is it a bug?

If I add a "+single-includes" to the command-line, the above problem
goes away.  But this flag breaks another source file down the line,
which expects different headerfile behavior due to having some extra
macro defined.  So I'm trying to understand exactly what
+single-includes does.

Here's my current splint command-line:

splint +keep -I./include  -DHAS_JIT -DI386 "-Isrc/pmc"
"-Icompilers/ast" -weak +showscan +showsummary -warnposixheaders
-abstract +aliasunique +ansireserved +ansireservedinternal
+assignexpose +bitwisesigned +boolcompare +booltype BOOLVAL -castfcnptr
-charint +continuecomment +controlnestdepth 15 +cppnames +declundef
+distinctinternalnames +evalorder +fielduse +forcehints -globs +imptype
+includenest 20 +incondefs +incondefslib +linelen 9999 +longintegral
+macroassign +macroempty +macroredef +matchanyintegral +nestedextern
+noeffect +readonlystrings +realcompare +shadow -type -isoreserved
-isoreservedinternal +single-includes

... followed by a couple hundred source filenames.

I guess my problem is that this source tree will not work with
+single-includes, without significant refactoring of the header
files...  But it seems to be required for specifying multiple C sources
on the same command-line.  Any other ideas for things I can try?  Is
there any alternative to breaking things up so splint runs once for each
source file?

Thanks,

Mark