On Thu, Mar 21, 2013 at 2:26 AM, Raistmer the Sorcerer <[email protected]> wrote:
>  I see crash after app's computations  completion and before " called 
> boinc_finish" line appears in stderr. Crash is hard to reproduce but I get 
> reports of such crash time to time (Windows, AstroPulse app)
>
By default, the 'safer' functions (<xxx>_s) don't return the error_t.
They call abort if there after failing validation on Windows
platforms. Did you install a handler via
_set_invalid_parameter_handler? See "Security Enhancements in the CRT
" (http://msdn.microsoft.com/en-us/library/8ef0s5kh(v=vs.80).aspx) and
_set_invalid_parameter_handler
(http://msdn.microsoft.com/en-us/library/a9yf33zb(v=vs.80).aspx)?

Also, you might be able to get a better dump if you st some flags in
_CrtSetReportMode
(http://msdn.microsoft.com/en-US/library/1y71x448(v=vs.80).aspx).
Dumps and crash reports are a source of data leakage, so be sure to
get consent before generating and sending.

Also, the Windows versions (perhaps others) do not check for success,
which is defined as 0. They check for failure with a particular value.

Jeff

> Среда, 20 марта 2013, 23:21 -07:00 от David Anderson <[email protected]>:
>>That commit avoids using strftime() to format the date/time
>>when the BOINC API (i.e. applications) writes messages to stderr
>>(e.g., error messages, "running standalone", or "called boinc_finish").
>>
>>Does this fix a bug?
>>Has anyone seen apps crash in strftime()?
>>
>>-- David
>>
>>On 19-Mar-2013 12:59 AM, Joachim Fritzsch wrote:
>>> I guess you are referring to this commit?!
>>>
>>>  
>>> https://github.com/matszpk/native-boinc-for-android/commit/8c8ccabee7b6b3f34c2bae0f2d72a49f2d8107da
>>_______________________________________________
_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.

Reply via email to