search

SuperDARN / rst

Radar Software Toolkit (RST)
https://superdarn.github.io/rst/
GNU General Public License v3.0
22 stars 18 forks source link

Missing acfd array causing segfault in IDL #344

Closed ecbland closed 4 years ago

ecbland commented 4 years ago

This issue is a follow up to #291, which fixed a segfault caused by a missing acfd array in a rawacf file. The problem was fixed in the C code (thanks @egthomas!), but it's still present in the IDL code.

Using the same example as #291 for the DLM code:

IDL> unit=rawopen('20080103.2201.00.inv.rawacf')      
% Loaded DLM: RAWDLM.
IDL> s=rawread(unit,prm,raw)                          
IDL> while s ne -1 do s=rawread(unit,prm,raw)
Segmentation fault (core dumped)

And similarly with the native IDL code

IDL> unit=rawopen('20080103.2201.00.inv.rawacf',/read)
IDL> s=rawread(unit,prm,raw)
% Compiled module: REVERSE.
% Conflicting data structures: structure tag,<INT       Array[23, 2]>.
% Execution halted at: RADARDECODERADARPRM  275 /rst/idl/lib/superdarn/main.1.25/rprm.pro
%                      RAWREAD           118 /rst/idl/lib/superdarn/main.1.25/raw.pro
egthomas commented 4 years ago

The DLM code was a pretty easy fix (https://github.com/SuperDARN/rst/commit/c56382751be390ce96b89bd1122de4bb70a34360) but it's not quite as obvious why the native IDL code is failing. It may be something to do with the first record of that particular file, since when I tried reading a trimmed version of that RAWACF file with a different start time the native code worked fine.

ecbland commented 4 years ago

Thanks @egthomas ! That commit fixes the problem for all of the files I looked at with the DLM code (20161031.*.*.sto.rawacf)

I think you're right that the native IDL code issue is more specific to that particular file from Inuvik. I can't reproduce the problem with the sto files (I hadn't checked that earlier).

pasha-ponomarenko commented 4 years ago

Sorry for barging in: may it be that the lag array in one or more records have a dimension which is different from the default one, 23x2?

egthomas commented 4 years ago

I think @pasha-ponomarenko is correct - the ltab array in that particular rawacf file (20080103.2201.00.inv.rawacf) has dimensions [23][2] according to dmapdump, when it should be [mplgs+1][2] (where in this case mplgs = 23), or [24][2].

ecbland commented 4 years ago

Closing this now that #346 has been merged