Trick Logging methods cause different values to be logged

[obei10]

Not sure what is happening here, but different methods of logging in trick lead to different results:
Old Method:

global DR_GROUP_ID
global drg
try:
    if DR_GROUP_ID >= 0:
        DR_GROUP_ID += 1
except NameError:
    DR_GROUP_ID = 0
    drg = []

drg.append(trick.DRBinary("state_unit"))
drg[DR_GROUP_ID].set_freq(trick.DR_Always)
drg[DR_GROUP_ID].set_cycle(GROK_LOG_CYCLE)
drg[DR_GROUP_ID].set_single_prec_only(False)
drg[DR_GROUP_ID].add_variable(   "mslUnitTest.MslAlt_args.geodetic_altitude")
drg[DR_GROUP_ID].add_variable(   "mslUnitTest.MslAlt_args.geodetic_latitude")
drg[DR_GROUP_ID].add_variable(   "mslUnitTest.MslAlt_args.longitude")
drg[DR_GROUP_ID].add_variable(   "mslUnitTest.msl_alt.mslAltitude")

trick.add_data_record_group(drg[DR_GROUP_ID], trick.DR_Buffer)
drg[DR_GROUP_ID].enable()

New Method:

recording_group_name = "test_data"
dr_group = trick.DRAscii(recording_group_name)
dr_group.set_cycle(GROK_LOG_CYCLE)
dr_group.freq = trick.DR_Always
trick.add_data_record_group(dr_group, trick.DR_Buffer)

dr_group.add_variable("mslUnitTest.MslAlt_args.geodetic_altitude")
dr_group.add_variable("mslUnitTest.MslAlt_args.geodetic_latitude")
dr_group.add_variable("mslUnitTest.MslAlt_args.longitude")
dr_group.add_variable("mslUnitTest.msl_alt.mslAltitude")

Same simulation, different logging methods delta plotted:

[obei10]

The above doesn't seem large, but it is. Below shows a MUCH larger difference as it's in meters instead of a radian compare against latitude.

[dbankieris]

I don't know what's up with the Java plots. I logged SIM_Ball++L1 in ASCII and binary, and the Java diff shows differences that don't match the tabular data. Koviz showed the correct diff. I think you said you tried Koviz, right?

[obei10]

I didn't do a diff in Koviz, just a co plot. However, the co-plot showed the values were different as well. It wasn't just the diff showing the issues, you could see the co-plots didn't line up either.

[dbankieris]

Are you sure your sim is deterministic? 😅

[obei10]

It's a look up table with an interpolator. It's deterministic. Very simple. The data I plotted here is just the data we piped into the model for our unit test. So, this data isn't being operated on at all, just read in from a file.

[ddj116]

Potentially related to #862...?