i.atcorr: Fix out-of-bounds array access in trunca computations

[ShubhamDesai]

This pull request fixes several instances of out-of-bounds array access in the computations.cpp file within the i.atcorr module. Specifically, it addresses issues where negative indices were being used to access arrays, which could lead to undefined behavior and potential crashes.

Changes Made

• Added a condition to ensure kk is not negative before using it as an array index.
  if (kk < 0) kk = 0;

Issues Resolved

1. Error : Array sixs_trunc.pha[83] accessed at index -1, which is out of bounds.
   Location : imagery/i.atcorr/computations.cpp:116:39
   Details : (double)((log10(sixs_trunc.pha[kk]) - log10(sixs_trunc.pha[k])) /

2. Error : Assuming condition is false
   Location : imagery/i.atcorr/computations.cpp:110:20
   Details : if (rmu[i] > 0.94)

3. Error : Assignment 'kk=i-1', assigned value is -1
   Location : imagery/i.atcorr/computations.cpp:112:16
   Details : kk = i - 1;

4. Error : Negative array index
   Location : imagery/i.atcorr/computations.cpp:116:39
   Details : (double)((log10(sixs_trunc.pha[kk]) - log10(sixs_trunc.pha[k])) /

5. Error : Array rmu[83] accessed at index -1, which is out of bounds.
   Location : imagery/i.atcorr/computations.cpp:117:27
   Details : (acos(rmu[kk]) - acos(rmu[k])));

6. Error : Assuming condition is false
   Location : imagery/i.atcorr/computations.cpp:110:20
   Details : if (rmu[i] > 0.94)

7. Error : Assignment 'kk=i-1', assigned value is -1
   Location : imagery/i.atcorr/computations.cpp:112:16
   Details : kk = i - 1;

8. Error : Negative array index
   Location : imagery/i.atcorr/computations.cpp:117:27
   Details : (acos(rmu[kk]) - acos(rmu[k])));

9. Error : Array sixs_trunc.pha[83] accessed at index -1, which is out of bounds.
   Location : imagery/i.atcorr/computations.cpp:118:46
   Details : double x1 = (double)(log10(sixs_trunc.pha[kk]));

10. Error : Assuming condition is false
    Location : imagery/i.atcorr/computations.cpp:110:20
    Details : if (rmu[i] > 0.94)

11. Error : Assignment 'kk=i-1', assigned value is -1
    Location : imagery/i.atcorr/computations.cpp:112:16
    Details : kk = i - 1;

12. Error : Negative array index
    Location : imagery/i.atcorr/computations.cpp:118:46
    Details : double x1 = (double)(log10(sixs_trunc.pha[kk]));

13. Error : Array rmu[83] accessed at index -1, which is out of bounds.
    Location : imagery/i.atcorr/computations.cpp:119:33
    Details : double x2 = (double)acos(rmu[kk]);

14. Error : Assuming condition is false
    Location : imagery/i.atcorr/computations.cpp:110:20
    Details : if (rmu[i] > 0.94)

15. Error : Assignment 'kk=i-1', assigned value is -1
    Location : imagery/i.atcorr/computations.cpp:112:16
    Details : kk = i - 1;

16. Error : Negative array index
    Location : imagery/i.atcorr/computations.cpp:119:33
    Details : double x2 = (double)acos(rmu[kk]);

[nilason]

@YannChemin Would this minor change to i.atcorr lead to any unintended consequences?

[wenzeslaus]

@ShubhamDesai Check this suggestion from ChatGPT: https://chatgpt.com/share/66f18923-0348-8009-bf2a-e415c5ae9a89

With or without that suggestion, this PS seems like it will require a regression test to make sure we are not breaking this. Can you write one in a separate PR? Either pytest style you used before for g.version (but with use of, e.g., r.univar to get raster values) or it might be easier to follow the unittest style with existing dataset like e.g., https://github.com/OSGeo/grass/blob/main/vector/v.surf.rst/testsuite/test_vsurfrst.py (check the assertVectorFitsUnivar calls). We can discuss in more detail when we meet.

[ShubhamDesai]

@ShubhamDesai Check this suggestion from ChatGPT: https://chatgpt.com/share/66f18923-0348-8009-bf2a-e415c5ae9a89

With or without that suggestion, this PS seems like it will require a regression test to make sure we are not breaking this. Can you write one in a separate PR? Either pytest style you used before for g.version (but with use of, e.g., r.univar to get raster values) or it might be easier to follow the unittest style with existing dataset like e.g., https://github.com/OSGeo/grass/blob/main/vector/v.surf.rst/testsuite/test_vsurfrst.py (check the assertVectorFitsUnivar calls). We can discuss in more detail when we meet.

Sure I will write the regression test for it.
I will refer the documentation first and other closely related modules testcases