From mboxrd@z Thu Jan  1 00:00:00 1970
From: francois.berenger@inria.fr (Francois Berenger)
Date: Thu, 04 Dec 2014 15:31:59 +0100
Subject: [Cocci] What is the spatch file for this change and the
 corresponding spatch command line invocation
In-Reply-To: <alpine.DEB.2.10.1412041354440.2118@hadrien>
References: <547F4892.6080405@inria.fr> <20141203173449.GX25677@wotan.suse.de>
 <5480584A.30107@inria.fr> <alpine.DEB.2.10.1412041354440.2118@hadrien>
Message-ID: <5480705F.6000608@inria.fr>
To: cocci@systeme.lip6.fr
List-Id: cocci@systeme.lip6.fr

On 12/04/2014 01:56 PM, Julia Lawall wrote:
>
>
> On Thu, 4 Dec 2014, Francois Berenger wrote:
>
>> On 12/03/2014 06:34 PM, Luis R. Rodriguez wrote:
>>> On Wed, Dec 03, 2014 at 06:29:54PM +0100, Francois Berenger wrote:
>>>> Hello,
>>>>
>>>> I give up trying to write the spatch myself:
>>>>
>>>> Before file:
>>>> ---
>>>> int i;
>>>>
>>>> for ( ; ; ++i ) {
>>>>     if (1) {
>>>>       continue;
>>>>     }
>>>> }
>>>> ---
>>>>
>>>> After file:
>>>> ---
>>>> int i;
>>>>
>>>> for ( ; ; ) {
>>>>     if (1) {
>>>>       ++i;
>>>>       continue;
>>>>     }
>>>>     ++i;
>>>> }
>>>> ---
>>>>
>>>> All my trials don't work, not any change in the file appears
>>>> if I try --in-place and -o fails because
>>>> "-o can not be applied because there are no modified files"
>>>
>>> Can you take the hands on tutorial before expecting folks to
>>> write rules for you?
>>>
>>> https://www.youtube.com/watch?v=buZrNd6XkEw
>>
>> If I really need to invest two hours before being able to use coccinelle
>> productively, that would be a serious entry barrier.
>>
>> And probably not just for me: for any potential new user of
>> coccinelle.
>
> I think that you may have started with an unfortunate case.  I'm not sure
> what is the strategy for matching the empty space in a for header.  It's
> not an expression, so what is it.

It is an optional expression: either there is an expression, either 
there is none.
I could not find such thing in the SmPL grammar manual (v1.0.0-rc21).

>  It may be possible only to match it
> exactly.
>
> On the other hand, it is helpful if you provide a semantic patch that
> you have tried, rather than just the before and after code.  From one
> example of before and after code, it is not always possible to guess the
> full generality of the rule that you are trying to implement.
>
> julia
>

-- 
Regards,
Francois.