RE: tmpfile crash?

From: Snyder, Arthur E. <snyder_at_slac.stanford.edu>
Date: Tue, 10 Jan 2012 08:07:02 -0800

I've been running with |gdb| attached, but so far haven't caught it in the act ...

---

From: Axel Naumann [Axel.Naumann_at_cern.ch] Sent: Tuesday, January 10, 2012 6:54 AM
To: Snyder, Arthur E.
Cc: roottalk_at_root.cern.ch
Subject: Re: [ROOT] tmpfile crash?

Hi Art,

This error handling is now much improved in the trunk and v5-32-00-patches: CINT will now re-throw the exception.

Nevertheless, as Timur points out, there is an exception caused by something in your code... gdb's "catch throw" will find it.

Cheers, Axel.

On 1/6/12 6:19 PM, Snyder, Arthur E. wrote:
> Dear roottalk:
>
> Error: Symbol #include is not defined in current scope (tmpfile):1:
> Error: Symbol exception is not defined in current scope (tmpfile):1:
> Syntax Error: #include <exception> (tmpfile):1:
> Error: Symbol G__exception is not defined in current scope (tmpfile):1:
> Error: type G__exception not defined FILE:(tmpfile) LINE:1
> (Long64_t)0
> *** Interpreter error recovered ***
> --what the hell is that?
>
>
>
> I get this while running a |selector| job that outputs a selected entries from the input TChain. It may be that it only occurs when I re-run in same root session. I'm testing that ... I do know that if I re-run the |selector| w/o restarting root after the crash occurs, it crashes again immediately rather than after some number of events have been processed.
>
>
>
> I haven't been able to find where |tmpfile| is, so far ...
>
>
>
> Does anybody know what causes this?
>
>
>
> -Art S.
>
>
Received on Tue Jan 10 2012 - 17:08:13 CET