GNU bug report logs - #17667
[PATCH] df: Initialize a variable to squash a compiler warning

Previous Next

Full log

View this message in rfc822 format

From: Ben Walton <bdwalton <at> gmail.com>
To: Pádraig Brady <P <at> draigbrady.com>
Cc: 17667-done <at> debbugs.gnu.org
Subject: bug#17667: [PATCH] df: Initialize a variable to squash a compiler warning
Date: Mon, 2 Jun 2014 20:29:27 +0100

On Mon, Jun 2, 2014 at 8:22 PM, Pádraig Brady <P <at> draigbrady.com> wrote:
> On 06/02/2014 09:09 AM, Ben Walton wrote:
>> * src/df.c: get_dev - With strict error checking, gcc complained that
>>             v may have been used prior to initialization. To avoid
>>             this, initialize to NULL.
>>
>> Signed-off-by: Ben Walton <bdwalton <at> gmail.com>
>> ---
>>  src/df.c | 4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/src/df.c b/src/df.c
>> index 01ecca6..059c958 100644
>> --- a/src/df.c
>> +++ b/src/df.c
>> @@ -924,7 +924,7 @@ get_dev (char const *disk, char const *mount_point, char const* file,
>>        char buf[LONGEST_HUMAN_READABLE + 2];
>>        char *cell;
>>
>> -      struct field_values_t *v;
>> +      struct field_values_t *v = NULL;
>>        switch (columns[col]->field_type)
>>          {
>>          case BLOCK_FLD:
>> @@ -934,7 +934,7 @@ get_dev (char const *disk, char const *mount_point, char const* file,
>>            v = &inode_values;
>>            break;
>>          case OTHER_FLD:
>> -          v = NULL;
>> +          /* Rely on NULL initialization. */
>>            break;
>>          default:
>>            assert (!"bad field_type");
>
> This is because assert() is not declared __noreturn__ on Solaris 10.
> That can be an important admonition for a compiler
> so I'm wondering should be detect this and provide a __noreturn__ wrapper.

Ah. Ok. That makes sense. I think providing this wrapper is likely a
good thing although this is a corner case we're catching here.

>
> Anyway what I don't want to do is change the current logic
> to avoid such bogus warnings. What we could do here
> is to tweak the assert path only to avoid the warning as follows.
> OK to push the following instead in your name?

Ack, that's fine with me.

Thanks
-Ben

>
> thanks,
> Pádraig.
>
> diff --git a/src/df.c b/src/df.c
> index 82b0c5f..c08ad97 100644
> --- a/src/df.c
> +++ b/src/df.c
> @@ -953,6 +953,7 @@ get_dev (char const *disk, char const *mount_point, char con
>            v = NULL;
>            break;
>          default:
> +          v = NULL; /* avoid warnings where assert() is not __noreturn__.  */
>            assert (!"bad field_type");
>          }
>
>



-- 
---------------------------------------------------------------------------------------------------------------------------
Take the risk of thinking for yourself.  Much more happiness,
truth, beauty and wisdom will come to you that way.

-Christopher Hitchens
---------------------------------------------------------------------------------------------------------------------------

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.