Re: svn commit: r35631 - in trunk/subversion: libsvn_client libsvn_fs_base libsvn_fs_base/bdb libsvn_fs_base/util libsvn_fs_fs libsvn_ra_local libsvn_ra_neon libsvn_ra_serf libsvn_ra_svn libsvn_repos libsvn_subr libsvn_wc mod_dav_svn svn svnlook svnserve tests/libsvn_subr

On Mon, 2 Feb 2009, Julian Foad wrote:

> On Mon, 2009-02-02 at 08:05 -0800, Blair Zajac wrote:
>> Julian Foad wrote:
>>> Author: julianfoad
>>> Date: Mon Feb 2 07:56:44 2009
>>> New Revision: 35631
>>>
>>> Log:
>>> Replace the expressions "? TRUE : FALSE" and "? FALSE : TRUE", which are ugly
>>> and redundant or almost redundant, with more direct ways to express the
>>> boolean test that is meant, throughout the C code base. No functional change.
>>>
>>> The changes are similar to these:
>>> boolean ? TRUE : FALSE => boolean
>>> boolean ? FALSE : TRUE => ! (boolean)
>>> x == y ? TRUE : FALSE => x != y
>>
>> I haven't reviewed all the changes, but for the third one you mean
>>
>> x == y ? FALSE : TRUE => x != y
>
> Yes. Thanks. Just a log message typo, I hope. I've edited the log
> message.

Good work!

But...

Index: trunk/subversion/libsvn_subr/sqlite.c
===================================================================
--- trunk/subversion/libsvn_subr/sqlite.c (revision 35630)
+++ trunk/subversion/libsvn_subr/sqlite.c (revision 35631)
@@ -304,8 +304,7 @@
  svn_boolean_t
  svn_sqlite__column_boolean(svn_sqlite__stmt_t *stmt, int column)
  {
- return (sqlite3_column_int64(stmt->s3stmt, column) == 0
- ? FALSE : TRUE);
+ return (sqlite3_column_int64(stmt->s3stmt, column) == 0);
  }

  int

No matter how long I stare at this one, it still looks wrong to me.
Shouldn't it be !=0 instead of ==0 ?

Martin

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1099228
Received on 2009-02-04 00:28:36 CET