BRLCAD 7.42.0 NIRT Issues

[randerson503-debug]

NIRT is working as currently intended, though I need to run an old code that uses NIRT to determine if the ray hits any region from a particular point forward. Because the -b option has been deprecated and NIRT always interrogates the whole model, the script is failing to do what it needs to do. Is there an option to run NIRT the old way (into the screen from a discreet point)?

[starseeker]

I think the -c option is what you're after here - that should trigger the old "default" behavior

[randerson503-debug]

You got me all excited, but alas, the -c option didn't change anything. NIRT still backs out before producing output.

[randerson503-debug]

My center is at -7.60, 1.677, 2099.55. When I use the -c option, the origin is still backing up to -30.46, 1.68, 2099.55 which is the same without the -c.

[randerson503-debug]

I installed an old version of BRLCAD 7.24.4 just to get the script to work. It would be nice to know if there is a way to perform a nirt the "old" way

[starseeker]

Hmm. OK. From the operating system command prompt, here's what I'm seeing without and with "-c" using moss.g as an example:

Without (i.e. backout active)

./bin/nirt share/db/moss.g all.g -e "center 20 0 2;ae 0 90;s;q"
BRL-CAD Release 7.42.1  Natalie's Interactive Ray Tracer
    Fri, 21 Nov 2025 20:40:01 UTC, Compilation 0
Output format: default (specify -L option for descriptive listing)
Formats available: csv-gap csv default diff entryexit gap1 gap2 ged (specify via -f option)
Database file:  'share/db/moss.g'
Building the directory...
Get trees...
Prepping the geometry...
Object 'all.g' processed
Database title: 'Gary Moss's "World on a Platter"'
Database units: 'mm'
model_min = (-50, -70, -24)    model_max = (90, 70, 63)
Origin (x y z) = (20.00000000 0.00000000 235.76141588)  (h v d) = (0.0000 2.0000 20.0000)
Direction (x y z) = (-0.00000000 -0.00000000 -1.00000000)  (az el) = (0.00000000 90.00000000)
    Region Name               Entry (x y z)              LOS  Obliq_in Attrib
ellipse.r            (  20.0000    0.0000   42.1922)  14.2318  52.7480 
platform.r           (  20.0000    0.0000  -16.8177)   6.9508   0.0000 
Quitting...

With (i.e. backout inactive)

./bin/nirt -c share/db/moss.g all.g -e "center 20 0 2;ae 0 90;s;q"
BRL-CAD Release 7.42.1  Natalie's Interactive Ray Tracer
    Fri, 21 Nov 2025 20:40:01 UTC, Compilation 0
Output format: default (specify -L option for descriptive listing)
Formats available: csv-gap csv default diff entryexit gap1 gap2 ged (specify via -f option)
Database file:  'share/db/moss.g'
Building the directory...
Get trees...
Prepping the geometry...
Object 'all.g' processed
Database title: 'Gary Moss's "World on a Platter"'
Database units: 'mm'
model_min = (-50, -70, -24)    model_max = (90, 70, 63)
Origin (x y z) = (20.00000000 0.00000000 2.00000000)  (h v d) = (0.0000 2.0000 20.0000)
Direction (x y z) = (-0.00000000 -0.00000000 -1.00000000)  (az el) = (0.00000000 90.00000000)
    Region Name               Entry (x y z)              LOS  Obliq_in Attrib
platform.r           (  20.0000    0.0000  -16.8177)   6.9508   0.0000 
Quitting...

So at a basic command line level it seems to be working. In MGED, however, "-c" does not have a similar result:

Without -c (backout active)

./bin/mged share/db/moss.g "e all.g;center 20 0 2;ae 0 90;nirt"

Firing from view center...

Origin (x y z) = (20.00 0.00 235.76)  (h v d) = (0.00 2.00 20.00)
Direction (x y z) = (-0.0000 -0.0000 -1.0000)  (az el) = (0.00 90.00)
Region Name                   Entry (x y z)              LOS   Obliq_in  Attrib
ellipse.r            (   20.000     0.000    42.192)    14.23   52.748
platform.r           (   20.000     0.000   -16.818)     6.95    0.000

With -c (should disable backout)

./bin/mged share/db/moss.g "e all.g;center 20 0 2;ae 0 90;nirt -c"

Firing from view center...

Origin (x y z) = (20.00 0.00 235.76)  (h v d) = (0.00 2.00 20.00)
Direction (x y z) = (-0.0000 -0.0000 -1.0000)  (az el) = (0.00 90.00)
Region Name                   Entry (x y z)              LOS   Obliq_in  Attrib
ellipse.r            (   20.000     0.000    42.192)    14.23   52.748
platform.r           (   20.000     0.000   -16.818)     6.95    0.000

So it looks like "-c" isn't working as expected for MGED's nirt.

[starseeker]

OK, confirmed - it was a bug. I think I've got a fix, which will be in the next release: a788c09

[starseeker]

As a workaround, you can use -e to explicitly set backout:

./bin/mged share/db/moss.g "e all.g;center 20 0 2;ae 0 90;nirt -e \"backout 0;s\""

Origin (x y z) = (20.00 0.00 2.00)  (h v d) = (0.00 2.00 20.00)
Direction (x y z) = (-0.0000 -0.0000 -1.0000)  (az el) = (0.00 90.00)
Region Name                   Entry (x y z)              LOS   Obliq_in  Attrib
platform.r           (   20.000     0.000   -16.818)     6.95    0.000

[PrashamJain1318]

Hi @randerson503-debug! I'm interested in contributing to help resolve this NIRT issue.

Based on the discussion, I understand the problem:

1. The -b option for NIRT has been deprecated in newer versions
2. The NIRT tool always interrogates the whole model, but the user needs the ability to run NIRT in the "old way" from a discrete point
3. The -c option should trigger the old "default" behavior but currently doesn't work

I can help with:

• Investigating the NIRT command-line option handling and parameter parsing
• Tracing through the code to understand why the -c option isn't producing the expected behavior
• Implementing or fixing the legacy behavior mode for discrete point interrogation
• Writing test cases to verify the fix works correctly
• Testing the solution with your provided test data/model

I have experience with C/C++ system programming and debugging command-line tools. Looking forward to helping solve this!