public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Greg Dietsche <greg@farmergreg.com>
To: Julia Lawall <julia@diku.dk>
Cc: Greg Dietsche <greg@farmergreg.com>,
	linux-kernel@vger.kernel.org, cocci@diku.dk,
	Gilles.Muller@lip6.fr
Subject: Re: [Cocci] Re: [PATCH v2] coccinelle: if (ret) return ret; return ret; semantic patch
Date: Wed, 15 Jun 2011 10:33:08 -0500	[thread overview]
Message-ID: <4DF8D0B4.7020505@gregd.org> (raw)
In-Reply-To: <Pine.LNX.4.64.1106150751390.27735@ask.diku.dk>

On 06/15/2011 12:58 AM, Julia Lawall wrote:
> On Tue, 14 Jun 2011, Greg Dietsche wrote:
>
>>
>>
>> On 06/14/2011 04:24 PM, Greg Dietsche wrote:
>>> On 06/14/2011 12:50 AM, Julia Lawall wrote:
>>>> On Mon, 13 Jun 2011, Greg Dietsche wrote:
>>>>> just curious... i see you usually just write "return ret;" here when
>>>>> posting.
>>>>> I've assumed that's because it will 1) work and 2) is close enough.
>>>>>
>>>>> You'll notice I've been doing:
>>>>> -return ret;
>>>>> +return ret;
>>>>> because it seems to help coccinelle realize that it can get rid of
>>>>> extra line
>>>>> feeds - does this make sense - or should i just be doing a "return ret"?
>>>> I wondered why you were doing that :)
>>>>
>>>> Is the problem that the removed if is being replaced by a blank line? If
>>>> so, that is not intentional. I will look into it. If it doesn't happen
>>>> always, an example where it does happen could be helpful.
>>>>
>>>
>>> Some times it gets it right, so it's not always wrong. For example:
>>>
>>> diff -u -p a/arch/m68k/math-emu/fp_log.c b/arch/m68k/math-emu/fp_log.c
>>> --- a/arch/m68k/math-emu/fp_log.c 2011-06-13 14:06:37.943954469 -0500
>>> +++ b/arch/m68k/math-emu/fp_log.c 2011-06-14 16:07:22.394954040 -0500
>>> @@ -105,9 +105,6 @@ fp_fetoxm1(struct fp_ext *dest, struct f
>>>
>>> fp_monadic_check(dest, src);
>>>
>>> - if (IS_ZERO(dest))
>>> - return dest;
>>> -
>>> return dest;
>>> }
>>>
>>>
>>>
>>> Here's an example where it got it "wrong" - notice how the blank line is
>>> missing the - :
>>>
>>> diff -u -p a/arch/frv/mm/pgalloc.c b/arch/frv/mm/pgalloc.c
>>> --- a/arch/frv/mm/pgalloc.c 2011-06-13 14:06:37.855954391 -0500
>>> +++ b/arch/frv/mm/pgalloc.c 2011-06-14 16:07:16.714954008 -0500
>>> @@ -136,8 +136,6 @@ pgd_t *pgd_alloc(struct mm_struct *mm)
>>> pgd_t *pgd;
>>>
>>> pgd = quicklist_alloc(0, GFP_KERNEL, pgd_ctor);
>>> - if (!pgd)
>>> - return pgd;
>>>
>>> return pgd;
>>> }
>
> OK, but it is going to be hard for Coccinelle to know that, although the
> programmer previously thought that return should be separated from the
> rest of the function by a blank line, that is now no longer the case.
> Perhaps it is due to the fact that there is now only one other line in the
> body of the function, but it seems like an opinion as to how to draw the
> line.
>

OK - I can see how that would be hard for Coccinelle to guess what we 
really want in this case.

> So your - return ret; + return ret; is probably the appropriate solution.
> You want to get rid of the whole pattern if (...) return ret; return ret;
> and replace it with just return ret;, which will then be inserted at the
> point of the beginning of the match to the pattern.

ok

> It would be nicer to put the - return ret; + return ret; inside the last
> line of the ( | ).  Then only those return ret's are rewritten rather than
> every return ret in the program.  It should improver performance and

except that 4 of the 5 ORs are cases where we want to do the -return 
ret; + return ret; So I suppose for performance, I should actually add 
the +/- to each of the 4 cases that we want cocci to generate a patch for?


thanks,
Greg

> reduce the risk of changing spacing.  The other ifs in the ( | ) don't
> need to be followed by return ret.  They are just ifs that you want to
> ignore completely.
>
> julia

  reply	other threads:[~2011-06-15 15:33 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-06-06  0:26 [PATCH] coccinelle: if(ret)return ret; return ret; semantic patch Greg Dietsche
2011-06-06  4:55 ` Julia Lawall
2011-06-07 15:47   ` Greg Dietsche
2011-06-07 16:56     ` Julia Lawall
2011-06-07 21:54       ` Nicolas Palix
2011-06-07 22:59         ` Greg Dietsche
2011-06-08  7:11           ` Nicolas Palix
2011-06-08  7:24             ` Julia Lawall
2011-06-11 15:37   ` Greg Dietsche
2011-06-11 15:43     ` Julia Lawall
2011-06-06  4:56 ` Julia Lawall
2011-06-13 18:23   ` [PATCH v2] coccinelle: if (ret) return " Greg Dietsche
2011-06-13 18:36     ` Joe Perches
2011-06-13 18:38     ` Julia Lawall
2011-06-13 20:55       ` Greg Dietsche
2011-06-14  5:50         ` Julia Lawall
2011-06-14 21:24           ` Greg Dietsche
2011-06-15  1:15             ` Greg Dietsche
2011-06-15  5:58               ` Julia Lawall
2011-06-15 15:33                 ` Greg Dietsche [this message]
2011-06-15 15:34                   ` [Cocci] " Julia Lawall
2011-06-15 15:35                   ` Julia Lawall
2011-06-15 16:05                     ` Greg Dietsche
2011-06-15  1:29       ` [PATCH v3] " Greg Dietsche
2011-06-15  5:50         ` Julia Lawall
2011-06-15 15:50         ` [PATCH v4] " Greg Dietsche

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4DF8D0B4.7020505@gregd.org \
    --to=greg@farmergreg.com \
    --cc=Gilles.Muller@lip6.fr \
    --cc=cocci@diku.dk \
    --cc=julia@diku.dk \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox