GBL track has no valid member variables
I was running the following analysis:
DESY 2019-07: run 868 TLU+Mimosa26+Timepix3+ATLASpix with the config file and geometry file here:
10_timing_analysis_apx_w23s11_run868.conf
Alignment_July2019_w23s11_1.geo
In the AnalysisEfficiency
module, I get a fatal internal error
because I'm trying to access a vector outside its scope, namely prev_hits_ts
I then traced back that this due to the fact that we give this vector some non-sense arguments (9223372036854775808).
These in turn are derived from feeding getColumn()
and getRow()
with nan
as you see in the debug output below.
|17:24:16.735| (ERROR) [R:AnalysisEfficiency:ATLASPix_0] track chi2ndof = nan
|17:24:16.735| (ERROR) [R:AnalysisEfficiency:ATLASPix_0] globalIntercept (-nan, -nan, -nan)
|17:24:16.735| (ERROR) [R:AnalysisEfficiency:ATLASPix_0] localIntercept = (-nan, -nan, -nan)
|17:24:16.735| (ERROR) [R:AnalysisEfficiency:ATLASPix_0] intercept_col/row = 9223372036854775808, 9223372036854775808
|17:24:16.735| (FATAL) [R:AnalysisEfficiency:ATLASPix_0] Fatal internal error
vector::_M_range_check: __n (which is 9223372036854775808) >= this->size() (which is 25)
When I'm running the same code without GBL, the fatal error does not show up. I also haven't seen it for other runs.
@lhuth Could you have a look at this?
Update:
I ran the above analysis without having set the correct material budget so the tracking was non-sense.
However, this situation should be caught properly.