On 06/05/2011 11:55 PM, Julia Lawall wrote:
Thanks.  I tried this too, but I wasn't sure about the results.  The
question is why stop here.  For example, there are IS_ERR calls that one
could consider as well.  Or ret<  0.  Or why not just:

@@
expression ret;
@@

- if (...) return ret;
   return ret;

I've been thinking a bit more about this. Is there a community preference towards patches that are highly reliable v.s. ones that find things that might not be a problem? I'm leaning towards updating my patch to do the above and then later coming back when I have more time to fix it up to find only things that are really problems.

Greg
Although there might be function calls that one doesn't want to touch, so:

@@
identifier f != IS_ERR;
expression ret;
statement S;
@@

(
if (<+...f(...)...+>) S
|
- if (...) return ret;
   return ret;
)

julia


On Sun, 5 Jun 2011, Greg Dietsche wrote:

This semantic patch finds code matching this pattern:
        if(ret)
                return ret;
        return ret;

I will be submitting patches shortly against the mainline to cleanup all
code matching this pattern.

Signed-off-by: Greg Dietsche<[email protected]>
---
  scripts/coccinelle/misc/doublereturn.cocci |   20 ++++++++++++++++++++
  1 files changed, 20 insertions(+), 0 deletions(-)
  create mode 100644 scripts/coccinelle/misc/doublereturn.cocci

diff --git a/scripts/coccinelle/misc/doublereturn.cocci 
b/scripts/coccinelle/misc/doublereturn.cocci
new file mode 100644
index 0000000..656a118
--- /dev/null
+++ b/scripts/coccinelle/misc/doublereturn.cocci
@@ -0,0 +1,20 @@
+/// Remove unecessary if/return in code that follows this pattern:
+///    if(retval)
+///            return retval;
+///    return retval;
+//
+// Confidence: High
+// Copyright: (C) 2011 Greg Dietsche GPLv2.
+// URL: http://www.gregd.org
+// Comments:
+// Options: -no_includes
+
+virtual patch
+
+@@
+identifier retval;
+@@
+-if (retval)
+-      return retval;
+-return retval;
++return retval;
--
1.7.2.5


_______________________________________________
Cocci mailing list
[email protected]
http://lists.diku.dk/mailman/listinfo/cocci
(Web access from inside DIKUs LAN only)

Reply via email to