Re: problem compiling with PROOF and stdio.h

From: Kathrin Stoerig <Kathrin.Stoerig_at_cern.ch>
Date: Thu, 4 Feb 2010 15:15:06 +0100


Hi Gerri,

I already tried this dictionary-solution and replacing the TLorentzVector-Single-Components by vector<TLorentzVector>. I was not very successful, although the code compiled, I always got segfaults when I trying to access these objects. All other branches were perfectly fine, even vector<vector<int> > etc. - this was the point when I gave up on MakeSelector and tried MakeProxy (successfully).

At the moment, using PROOF is not my highest priority. With the Proxy, I get a lot of interesting features I can already use with root. So, for me it is perfectly fine to wait for a new ROOT version. :-)

Cheers,

    Kathrin

On 02/04/2010 02:24 PM, ganis wrote:
>
> Hi Kathrin,
>
> Including the dictionary for vector<TLorentzVector> as suggested by
> Philippe should work also with
> TSelector. what errors do you get?
>
> I will try to add support for proxies in PROOF asap.
>
> Cheers,
> Gerri
>
> Kathrin Stoerig wrote:
>> Hi Gerri,
>>
>> the principle idea was to read some "not-totally flat" ntuple (means
>> vector<float> and vector<TLorentzVector> mainly) with automatically
>> generated code and PROOF. - Example data file can be found at
>> https://james.physik.uni-freiburg.de/~stoerig/testdata.root , use
>> Tree CommonContainersTree
>>
>> I already "played" around quite a while using MakeSelector, but I
>> encountered the problem that I could not access the TLorentzVectors..
>> as the structure is not recognized correctly and also manual changes
>> were not successful.
>> I read somewhere that MakeProxy offers better support for such
>> peculiar data types - and nearly out of the box, MakeProxy works fine
>> for me and I can do my analysis.
>>
>> Of course I am open for any suggestions.
>>
>> Cheers,
>> Kathrin
>>
>>
>> On 02/03/2010 07:12 PM, ganis wrote:
>>>
>>> Dear Kathrin,
>>>
>>> There no support for proxy in PROOF yet.
>>> Can you give more details about what you are trying to do so that
>>> we can try to suggest a
>>> possible solution?
>>>
>>> Gerri Ganis
>>>
>>> Kathrin Stoerig wrote:
>>>> Hello,
>>>>
>>>> I try to implement my analysis with MakeProxy and Proof.
>>>> The automatically generated code works fine with ROOT (5-26a 32 bit
>>>> from cernafs), but when I try to compile within PROOF, I get the
>>>> following error messages related to stdio.h?!
>>>>
>>>> Any help is very welcome.
>>>> Kind regards,
>>>>
>>>> Kathrin
>>>>
>>>> >>>>>>>>>>>>>>>>>>>>>>>>
>>>>
>>>> In file included from /usr/include/stdio.h:909,
>>>> from
>>>> /home/stoerig/.proof/session-pb-d-128-141-42-207-1265208987-10792/worker-0.0-pb-d-128-141-42-207-1265208988-10803/./proxyWD3PD_C_ACLiC_dict.h:11,
>>>>
>>>> from
>>>> /home/stoerig/.proof/session-pb-d-128-141-42-207-1265208987-10792/worker-0.0-pb-d-128-141-42-207-1265208988-10803/./proxyWD3PD_C_ACLiC_dict.cxx:16:
>>>>
>>>> /usr/include/bits/stdio.h: In function ‘__ssize_t getline(char**,
>>>> size_t*, FILE*)’:
>>>> /usr/include/bits/stdio.h:118: error: ‘__getdelim’ was not declared
>>>> in this scope
>>>> g++:
>>>> /home/stoerig/.proof/session-pb-d-128-141-42-207-1265208987-10792/worker-0.0-pb-d-128-141-42-207-1265208988-10803/./proxyWD3PD_C_ACLiC_dict.o:
>>>> No such file or directory
>>>> 15:56:48 10803 Wrk-0.0 | Error in <ACLiC>: Compilation failed!
>>>>
>>>> >>>>>>>>>>>>>>>>>>>>>>>>
>>>>
>
Received on Thu Feb 04 2010 - 15:17:04 CET

This archive was generated by hypermail 2.2.0 : Thu Feb 04 2010 - 17:50:01 CET