Re: [Qemu-devel] [PATCH 2/3] block: Add support to "open" /dev/fd/X filenames

[Corey Bryant <coreyb@linux.vnet.ibm.com>]

On 06/04/2012 12:36 PM, Eric Blake wrote:
> On 06/04/2012 10:28 AM, Corey Bryant wrote:
>
>>> But at least strtol lets you detect errors:
>>>
>>> char *tmp;
>>> errno = 0;
>>> fd = strtol(p,&tmp, 10);
>>> if (errno || tmp == p) {
>>>       /* raise your error here */
>>> }
>>
>> I don't think this is legitimate.  errno can be set under the covers of
>> library calls even if the strtol() call is successful.
>
> Wrong.  POSIX _specifically_ requires that strtol() leave errno
> unchanged unless strtol() is reporting a failure, no matter what other
> library calls (if any) are made under the covers by strtol().
>
> In other words, pre-setting errno to 0, then calling strtol(), then
> checking errno, _is_ the documented way to check for strtol() failures,
> and a correct usage of strtol() MUST use this method.  See also commit
> 6b0e33be88bbccc3bcb987026089aa09f9622de9.  atoi() does not have this
> same guarantee, which makes atoi() worthless at detecting errors in
> relation to strtol().
>

Great!  I see it now (excerpt below is from the opengroup 
specification).  This is definitely an exception from normal behavior of 
C APIs.

"The strtol() function will not change the setting of errno if successful."

>>
>> I was thinking if strtol returns 0 and errno is 0, perhaps we could
>> assume success, but I don't think this is guaranteed either.
>
> Actually, it _is_ guaranteed - if you pre-set errno to 0, then call
> strtol(), then errno is still 0, then the result did not encounter an
> error, so a result of 0 at that point means that you indeed parsed a 0.
>
>>
>> Maybe a combination of isdigit() then strtol() will give a better idea
>> of success.
>
> Not necessary.
>

-- 
Regards,
Corey