Re: Build failed in Jenkins: 3.HEAD-i386-FreeBSD-6.4 #1041

O_O

I was trying something different (using another ifstream), but the
tune is the same.
To me this sounds a bug in the library :\

Thanks for looking into this, please commit at will ;)

On Wed, Aug 10, 2011 at 4:42 PM, Amos Jeffries <squid3_at_treenet.co.nz> wrote:
> On 11/08/11 02:10, Alex Rousskov wrote:
>>
>> On 08/09/2011 06:34 PM, Amos Jeffries wrote:
>>>
>>> On Tue, 09 Aug 2011 12:56:09 -0600, Alex Rousskov wrote:
>>>>
>>>> On 08/09/2011 02:13 AM, Kinkie wrote:
>>>>>>
>>>>>> ./cf_gen cf.data ../../src/cf.data.depend
>>>>>> error while opening input file 'cf.data': Unknown error: 0
>>>>>
>>>>> Now this puzzles me.
>>>>> ifstream fp.bad() reports badbits set, but errno is OK. On FreeBSD
>>>>> 6.4 only.
>>>>
>>>> Perhaps the errno is reset by subsequent system calls? How close is your
>>>> bad() detection to the errno use?
>>>>
>>>
>>> The code in question is:
>>>
>>>     fp.open(type_depend, std::ifstream::in);
>>>     if (fp.fail()) {
>>>         std::cerr<<  "error while opening type dependencies file '"<<
>>>                   input_filename<<  "': "<<  strerror(errno)<<
>>>  std::endl;
>>>         exit(1);
>>>     }
>>
>> Does rewriting the above to preserve errno (as in the sketch below) help?
>>
>>     fp.open(...);
>>     if (!fp) {
>>         const int savedErrno = errno;
>>         std::cerr<<  "error: "<<  strerror(errno)<<  std::endl;
>>     }
>>
>> Cheers,
>>
>> Alex.
>
> Neither of those worked in any combination.
>
> But....
>  looking at the snippet I see type_depends being opened and input_filename
> being reported. So fix that typo and look for input_filename being opened.
>  ... which shows that fp is re-used between the two files.
>
> Adding fp.clear() to reset the flags (including EOF?) after the fp.close()
> solves it.
>
> Amos
> --
> Please be using
>  Current Stable Squid 2.7.STABLE9 or 3.1.14
>  Beta testers wanted for 3.2.0.10
>

-- 
    /kinkie