[rhelv6-list] abrt

[Larry Martell]

On Wed, Apr 6, 2016 at 2:45 PM, Larry Martell <larry.martell at gmail.com> wrote:
> On Wed, Apr 6, 2016 at 2:40 PM, Eli Heady <eli.heady at gmail.com> wrote:
>> Hi Larry,
>>
>> I might start by configuring abrt to process these events by enabling
>> exception handling in unpackaged software. According to the latest docs [1],
>> this is done in /etc/abrt/abrt-action-save-package-data.conf by setting
>> 'ProcessUnpackaged = yes'. Disclaimer: I'm not familiar with how abrt
>> handles such events, what information can be saved, or even if this is safe
>> to set system-wide. I would consult the docs or the source before doing this
>> on a production system myself - this is a lmgtfy answer in which I did the
>> search for you, sorry for the lack of authoritative info :)
>>
>> This isn't what you asked and may be overly pedantic (apologies if so), but
>> I would personally not rely only on abrt to catch and report on failures of
>> custom Python applications. Python has very good built-in exception handling
>> capabilities. Custom exception classes are a good way to capture useful data
>> like tracebacks (see traceback and cgitb modules). When combined with the
>> logging module, one can easily relay such failures through syslog or another
>> logging vehicle. Abrt has its place, and if the Python interpreter has died
>> without executing exception handling then this paragraph is moot. A good set
>> of examples are at [2].
>
> Thanks. I do have exception handlers and I am capturing stdout and
> stderr, and there's nothing.

So from abrtd I got info on the unhanded exceptions, and they are
typical exceptions, e.g. TypeError or ValueError. So what I am
wondering is why these were not just dumped out to stderr as other
unhanded exceptions are.