Re: [tracecompass-dev] Hello

[Geneviève Bastien <gbastien+lttng@xxxxxxxxxxxx>]

Hi Florian,

Thank you for your interest in TraceCompass and welcome to the community!

When you try to push on gerrit, if it rejects your contribution, it will
usually display some [helpful] error message. Can you post the result of
the git push command? Also make sure that the e-mail address associated
with your Eclipse account is the same as the one in the signed-off and
of the author of the patch. You may need to update your git config.

As for the custom parser, you may want to look at the TmfXmlTraceStub
class in the o.e.t.tmf.core.tests plugin. We had the same issue as you:
the custom parsers parsed the events all right, but they were all the
same type and it was more difficult to develop analyses for it this way.
This class uses a custom XML parser to read the events of the trace and
rewrites them with a different event type and fields. It's probably not
useful to you, but it may give you ideas.

Cheers,
Geneviève


On 04/30/2016 08:24 PM, Florian Mösch wrote:
> Hi!
> I wanted to try to set up my git and gerrit account to be able to push
> changes - I have actually just chosen one occurence of a null type
> safety warning that seems to be valid to me.
> But gerrit does not allow me to push my commit. I have signed the CLA,
> installed the hook and put my "Signed-off-by" signature in the commit
> message. Still I must have missed something...
> Any hints?
> Thanks,
>   Florian.
>
> 2016-05-01 0:19 GMT+02:00 Florian Mösch <florian@xxxxxxxxxx>:
>> Hi Patrick,
>> thanks for your feedback. I have indeed noticed that different
>> matching lines can define different fields. That helps a bit but
>> unfortunately, the table view will grow huge if you need to extract
>> different information from many different events...
>> On the other hand, it looks like writing your own "real" parsers is
>> not that difficult and maybe a more suitable way to go.
>> Anyway: I think it would help a lot if different root lines could
>> generate different event types or different parsers could on the other
>> hand generate the same event types. I'll have a look at the code and
>> see if I can make a proposal for e change. But first, I'll continue to
>> work on my first own parser.
>> Regarding the null-pointer safety warnings, I have meanwhile read
>> about some of the limitations that lead to many false warnings: the
>> compiler should know that AnyObject.class is NonNull e.g. But I think,
>> I've also see a couple of warnings that seem to make sense.
>> I'll check if I see some that could or should be fixed.
>> Best regards,
>>   Florian.
> _______________________________________________
> tracecompass-dev mailing list
> tracecompass-dev@xxxxxxxxxxx
> To change your delivery options, retrieve your password, or unsubscribe from this list, visit
> https://dev.eclipse.org/mailman/listinfo/tracecompass-dev