Re: Unwanted error message

From: Tom Roberts <tjrob_at_fnal.gov>
Date: Wed, 17 Dec 2008 11:05:08 -0600


Andrew:

Thank you very much. That works just fine.

Tom Roberts

Andrew T. Mastbaum wrote:
> Tom,
>
> To hide the errors, you can set ROOT's global print threshold with the
> line "gErrorIgnoreLevel = LEVEL;", where LEVEL is kVerbose, kDebug, kInfo,
> kWarning, kError, or kFatal. In this case, you'd want to set
> gErrorIgnoreLevel to kFatal. This trick worked for me for suppressing
> useless warnings about a full TSpectrum peak buffer.
>
> Since you typically want errors reported, it would be good practice to set
> this back to warning when you're done TTreeFormula'ing.
>
> Hope this works for you.
>
> Cheers,
>
> Andy Mastbaum
> P-23 Weak Interactions Team
> Los Alamos National Laboratory
> TA-53 Building 1, D115
> Phone: (505) 665-6040
> Email: mastbaum_at_lanl.gov
>
>

>> To check an expression for validity with a given TNtuple, I construct a
>> temporary TTreeFormula and check whether it succeeded. For invalid
>> expressions I get the error message: Error in <TTreeFormula::Compile>:
>> Bad numerical expression : "in_Px"
>>
>> This looks bad to my users (there can be dozens in a row). It is
>> legitimate for these expressions to be invalid for the TNtuple that is
>> currently selected -- that's why I am checking them before trying to use
>> them. So I don't want the error message. Zeroing stderr and stdout do
>> not suppress it. How can I suppress it? Alternatively, is there some
>> other way to check whether a given expression is valid for a given
>> TNtuple?
>>
>> At present I'm printing a notice to ignore the errors -- this workable,
>> but looks silly....
>>
>>
>> Tom Roberts
>>
>>

>
Received on Wed Dec 17 2008 - 18:05:36 CET

This archive was generated by hypermail 2.2.0 : Wed Dec 17 2008 - 23:50:02 CET