Re: [dev] fsbm

From: Calvin Morrison <mutantturkey_AT_gmail.com>
Date: Fri, 7 Nov 2014 13:51:55 -0500

If I fflush it, it works, but that seems hacky

On 7 November 2014 13:28, Calvin Morrison <mutantturkey_AT_gmail.com> wrote:
> folks,
>
> I can't grep this program. I do not know why... Any ideas?
>
> On 7 November 2014 05:49, Dimitris Papastamos <sin_AT_2f30.org> wrote:
>> On Fri, Nov 07, 2014 at 11:40:03AM +0100, k0ga_AT_shike2.com wrote:
>>>
>>> >> > * Use strtonum() or estrtol() from sbase instead of atoi().
>>> >>
>>> >> I agree here if the string comes from the user. If the string
>>> >> is an internal string then there is no problem with atoi.
>>> >
>>> > It is generally unlikely that the string has been validated to
>>> > be an integer before getting to atoi(). With atoi() you cannot
>>> > distinguish between an invalid integer and 0.
>>>
>>> There are a lot of cases where it is validated, for example
>>> if you used first a regular expression to filter lines. There
>>> are also a lot of uses where doesn't mind the difference
>>> between 0 and no integer (for example sort -n). There
>>> are also a lot of cases where 0 is not a valid value.
>>>
>>> > Generally speaking, it should never be used.
>>>
>>> 'should never be used' is not a very general sentence.
>>>
>>> But, ok. This is a personal taste (this and the other question
>>> about strlcpy), and I don't want to begin a war.
>>
>> BTW, just to clarify, strtol() and friends has its set of problems
>> as well. It is too elaborate to check for error conditions unless
>> you already have a wrapper to do this once.
>>
>> strtonum() is I think the most appropriate interface to default to
>> and perhaps switch over to atoi() when you are aware of its limitations
>> and they do not impact the correctness of your code.
>>
Received on Fri Nov 07 2014 - 19:51:55 CET

This archive was generated by hypermail 2.3.0 : Fri Nov 07 2014 - 20:00:11 CET