Hi,
we have seen this in the past, caused by a string buffer overflow (a few
are remaining in CINT; replacing them would cause a long tail of
changes). You can try to reduce PATH, INCLUDE, ROOTSYS etc - already
looking at the output of env might show you a relevant "long" variable.
Or just post it so we can have a look together.
Cheers, Axel.
On 10/22/10 7:05 PM, Alexander Mann wrote:
>
> Hi Rene,
>
> yes, that's why I am saying that there is some trucated path and
> wondering where ROOT gets this from (apparently it writes it to some cxx
> file it creates).
>
> cu,
> Alexander
>
>
>
>
> Am 22.10.2010 19:00, schrieb Rene Brun:
>> Alexander,
>>
>> Did you see this error message: error: /i686-slc5-gcc43-opt/root/src: No
>> such file or directory ?
>>
>> Rene Brun
>>
>>
>> On 22/10/2010 14:17, Alexander Mann wrote:
>>>
>>> Hi,
>>>
>>> I have again a similar problem with rootcint. I builds correctly on
>>> one site, but fails on another, although I am explicitly passing the
>>> root path (which seemed to have fixed the earlier problem). You can
>>> find the log files here:
>>>
>>> http://voatlas20.cern.ch:25980/monitor/logs/034cc81e-7dc7-4406-a9bd-53c368e55a17/tarball_PandaJob_1133599319_ANALY_QMUL/prun_stdout.txt
>>>
>>>
>>>
>>> It says
>>>
>>> /scratch/tmp/8138846.1.lcg_long/5tgaCG_cint.cxx:1:41: error:
>>> /i686-slc5-gcc43-opt/root/src: No such file or directory
>>>
>>> It looks like some include path in truncated, but I do not know where
>>> rootcint gets this path from.
>>>
>>> The job monitoring page is here:
>>>
>>> http://panda.cern.ch:25980/server/pandamon/query?job=*&jobsetID=723&user=Alexander%20Mann&days=7
>>>
>>>
>>>
>>> Maybe it gives you an idea what is wrong here.
>>>
>>> cu,
>>> Alexander
>>>
>>>
>>>
>>> Am 29.09.2010 20:41, schrieb Philippe Canal:
>>>>
>>>> Hi Alexander,
>>>>
>>>> The -v is misplaced. I see
>>>>
>>>> ... rootcint ... -v -f src/SFrameCore_Dict.cxx -c -p -
>>>>
>>>> whereas you need
>>>>
>>>> ... rootcint ... -f src/SFrameCore_Dict.cxx -c -v -p -
>>>>
>>>> Cheers,
>>>> Philippe.
>>>>
>>>> On 9/29/10 1:36 PM, Alexander Mann wrote:
>>>>>
>>>>> Hi Philippe,
>>>>>
>>>>> same result, no additional output. Please find the full log file here:
>>>>>
>>>>> http://voatlas21.cern.ch:25980/monitor/logs/21a4504c-38de-4da1-bd64-2575ea9c91eb/tarball_PandaJob_1121142989_ANALY_QMUL/prun_stdout.txt
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> (Search for "unknown option")
>>>>>
>>>>> cu,
>>>>> Alexander
>>>>>
>>>>>
>>>>>
>>>>> Am 29.09.2010 20:10, schrieb Philippe Canal:
>>>>>>
>>>>>> Hi Alexander,
>>>>>>
>>>>>> If placed correctly the -v option should have produced just enough
>>>>>> extra
>>>>>> output.
>>>>>>
>>>>>> For example:
>>>>>>
>>>>>> $ rootcint -f EventDict.cxx -c -v -p Event.h EventLinkDef.h
>>>>>> g++ -E -C "-DG__LINUX=1" "-DG__GNUC=4" "-DG__GNUC_MINOR=4"
>>>>>> "-DG__GNUC_VER=4004" "-DG__GLIBC=2" "-DG__GLIBC_MINOR=12"
>>>>>> "-DG__x86_64=1" "-DG__amd64=1" "-D__attribute__(X)=" "-D__asm(X)="
>>>>>> "-DG__ROOT=1" "-DG__NATIVELONGLONG=1" "-DTRUE=1" "-DFALSE=0"
>>>>>> "-Dexternalref=extern" "-DSYSV" "-D__MAKECINT__"
>>>>>> "-DG__CINTVERSION=50180000" -DG__EXTERNAL_CPP -I.
>>>>>> -I/local2/pcanal/root_working/code/root-untouched/include
>>>>>> -I/local2/pcanal/root_working/code/root-untouched/src -D__CINT__
>>>>>> -I/local2/pcanal/root_working/code/root-untouched/cint/cint/include
>>>>>> -I/local2/pcanal/root_working/code/root-untouched/cint/cint/stl
>>>>>> -I/local2/pcanal/root_working/code/root-untouched/cint/cint/lib
>>>>>> /tmp/VfydZw_cint.cxx > /tmp/JpUEgB_cint
>>>>>> /local2/pcanal/root_working/code/root-untouched/cint/cint/include/stdfunc.dll:DLLREV=30051515
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> This cint accepts DLLREV=30051501~30051599 and creates 30051515
>>>>>>
>>>>>> Cheers,
>>>>>> Philippe.
>>>>>>
>>>>>> On 9/29/10 10:03 AM, Alexander Mann wrote:
>>>>>>>
>>>>>>> PS: The -v option produced surprisingly little additional output. Is
>>>>>>> there an option for even more verbosity?
>>>>>
>>>
>>
>
>
Received on Fri Oct 22 2010 - 21:42:48 CEST