> From: Justin Erenkrantz [mailto:[EMAIL PROTECTED]]
> Sent: 30 September 2001 06:30
>> On Sun, Sep 30, 2001 at 12:24:08AM -0400, Cliff Woolley wrote:
>>
>> A few warnings that have shown up on Solaris 2.6 sometime
>> within the last few weeks:
>>
>> thread.c: In function `apr_thread_once_init':
>> thread.c:261: warning: missing braces around initializer
>> thread.c:261: warning: (near initialization for
>> `once_init.__pthread_once_pad')
>
> I've seen this on Solaris 8 as well. I was tempted to commit the {}s,
> and I know that it compiles fine on Linux 2.4 with the {}. Anyone
> else care to comment?
>
> Linux defines PTHREAD_ONCE as 0, but Solaris defines it as a
> structure and hence wants the {}. Is there a side effect if we
> place too many {}s? I forget if they are fluff or not.
That should work. Doing this:
int i = { 0 };
works just fine, because the initializer list matches the 'number
of fields and the depth'* to initialize. In this case, it should
be ok. Note that it isn't simply fluff, this will give you a warning:
int i = {{ 0 }};
*) This isn't a quote, this is me trying to find words to describe
it.
>> thread_rwlock.c:224: warning: no previous prototype for
>> `apr_thread_rwlock_lock'
>>
>> This should be easy to fix.
The proto simply isn't in the header file, but I can't tell if
this is intentional and apr_thread_rwlock_lock was just a copy
'n paste too many in the c file, or that it should actually be
in the header. Aaron?
> I haven't seen this, but I may have missed it.
>
>> proxy_ftp.c: In function `ap_proxy_ftp_handler':
>> proxy_ftp.c:800: warning: subscript has type `char'
>>
>> I looked at this... I must be missing something, but I swear I don't see
>> what the problem is. Anyone?
550: char buffer[MAX_STRING_LEN];
555: int i = 0, j, len, rc;
800: for (i = 0; buffer[i] && !isdigit(buffer[i]); i++);
Hmmm, this is very weird...
> I'm actually compiling proxy for the first time. If I see anything,
> I'll comment. -- justin
Sander
