[gentoo-user] Re: udevil question for Helmut Jarausch

[walt]

On 06/16/2015 02:32 AM, Helmut Jarausch wrote:
> On 06/15/2015 11:56:41 PM, walt wrote:
>> Hi Helmut.  sys-apps/udevil failed to compile this morning and I
>> noticed you
>> had already submitted a proposed fix for it (your patch worked, thank
>> you).
>>
>> The failing code was already in the udevil package when I emerged it
>> on Jan
>> 21 with no problems, so some other package must have changed since Jan
>> 21,
>> but which one(s)?
>>
>> I ask the question because I want to learn to think like a developer
>> (hold
>> all snarky comments until the professionals get here) so could you
>> describe
>> how you arrived at your fix so quickly?
>>
>> Have you used the "stat" function so often that you didn't even need
>> to think
>> about how to fix it, or was it more complicated?
>>
>> Thanks for any wisdom you care to share with us.
>>
>
> The build log says
>
>  device-info.c:943:33: error: implicit declaration of function 'stat' [-Werror=implicit-function-declaration]
>                                  stat( mount_source, &statbuf ) == 0 &&
>                                  ^
> device-info.c:944:33: error: implicit declaration of function 'S_ISBLK' [-Werror=implicit-function-declaration]
>                                  S_ISBLK( statbuf.st_mode ) )
>
> which means the function stat isn't declared. Together with S_ISBLK it is clear that the C-function stat (to determine
> file status) is meant.
>
> man lstat    (man stat doesn't work here)

Strange, it works here.  I wonder if the difference is caused by your l10n, or
maybe a bug.  Are man pages always in English?  I've been pondering the switch
to man-db but I'm not motivated to deal with any fallout and disk space is cheap.

> shows which header files have to be included.
>
> Furthermore
>
> grep -rF /usr/include  S_ISBLK
> shows
> /usr/include/sys/stat.h:#define S_ISBLK(mode)    ....
>
> Therefore <sys/stat.h> has to be included somewhere

Thanks, good hints.

> No wisdom at all,

Hm.