On 12/24/10 18:32, Arjan van de Ven wrote:
> On 12/24/2010 9:20 AM, Michael Riepe wrote:
>>
>> On 12/24/10 17:47, Arjan van de Ven wrote:
>>
>>> can you tell me what your LANG (and locale) are set to ?
>>> (I kind of am wondering if it breaks when set to de_DE.utf-8 ... would
>>> help me reproduce and debug)
>> lc_ctype=de...@euro (i.e., ISO-8859-15)
>>
>> LANG and the other LC_* variables are unset. Changes of LC_CTYPE make no
>> difference.
>>
>> There simply is no 'e' in tmpbuf when fmt_prefix is passed a NaN (-nan
>> in this case). strchr() returns NULL, the call to atoi() adds +1, and
>> 0x1 is exactly the address where the SEGV happens, according to the
>> kernel message.
>>
>> By the way, this doesn't seem to be related to the "nl80211 not found"
>> messages, as it happens later (much later, on the CPU's time scale).
>
> ok I think I fixed it and pushed this out to the git tree; I also
> included that kernel header for now...
> at least it should not crash now.
> why there is a NAN there is something to investigate after the crash
> goes way...
It comes from cpudevice::power_usage(), by way of get_result_value().
But I'm not familiar enough with the code to tell you why.
> git clone git://git.kernel.org/pub/scm/status/powertop/powertop.git
>
> if you were using the tarbal version and not the git tree....
Now it segfaults in lib.cpp:282:
if (strstr(getenv("LANG"), "UTF-8"))
Of course, I should add, because LANG is unset.
With LANG=de, the program works.
--
Michael "Tired" Riepe <[email protected]>
X-Tired: Each morning I get up I die a little
_______________________________________________
Discuss mailing list
[email protected]
http://lists.lesswatts.org/listinfo/discuss
