[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <e87c12f2-40ab-69b7-2f55-f1fcc980e784@web.de>
Date: Sun, 25 Aug 2019 17:30:57 +0200
From: Markus Elfring <Markus.Elfring@....de>
To: Denis Efremov <efremov@...ux.com>,
Gilles Muller <Gilles.Muller@...6.fr>,
Julia Lawall <julia.lawall@...6.fr>,
Masahiro Yamada <yamada.masahiro@...ionext.com>,
Michal Marek <michal.lkml@...kovi.net>,
Nicolas Palix <nicolas.palix@...g.fr>, cocci@...teme.lip6.fr
Cc: linux-kernel@...r.kernel.org
Subject: Re: [PATCH] scripts: coccinelle: check for !(un)?likely usage
> +(
> +* !likely(E)
> +|
> +* !unlikely(E)
> +)
Can the following code variant be nicer?
+*! \( likely \| unlikely \) (E)
> +(
> +-!likely(E)
> ++unlikely(E)
> +|
> +-!unlikely(E)
> ++likely(E)
> +)
I would find the following SmPL change specification more succinct.
+(
+-!likely
++unlikely
+|
+-!unlikely
++likely
+)(E)
> +coccilib.org.print_todo(p[0], "WARNING use unlikely instead of !likely")
…
> +msg="WARNING: Use unlikely instead of !likely"
> +coccilib.report.print_report(p[0], msg)
1. I find such a message construction nicer without the extra variable “msg”.
2. I recommend to make the provided information unique.
* How do you think about to split the SmPL disjunction in the rule “r”
for this purpose?
* Should the transformation become clearer?
Regards,
Markus
Powered by blists - more mailing lists