LE4 strange indexing bug

[StevenClontz]

Having 0=0 as the first equation is non-ideal to begin with, but notice that also the x_4 terms appear as x_1 for some reason:

[StevenClontz]

(side note, the only time I saw this, the student simplified the equations and answered it correctly, so hurray)

[siwelwerd]

Also affects V9

[siwelwerd]

I am guessing it is a problem with latex_system_from_matrix() -- is that something we wrote? Google doesn't turn up anything for it.

checkit-tbil-la/outcomes/V9.sage

Line 17 in 8dcd044

"system": latex_system_from_matrix(A),

[StevenClontz]

we wrote it https://github.com/StevenClontz/checkit/blob/main/wrappers/sage_wrapper.sage#L53 I'll inspect this later

[siwelwerd]

So far, I have only seen this bug when there are 4 or 5 variables. Here's an example with 5

[StevenClontz]

I believe this is fixed now (somehow) - I have a build in Canvas I made in September where the bug is gone, but the build I uploaded in August (and still have been using it seems) has the bug. I'll close this issue once we publish the fall improvements to the public bank.

[siwelwerd]

The build I made for this semester contains this bug.

[StevenClontz]

Having lived through it myself, my inclination is to ignore it and hope it comes out in the wash this summer of upgrading to CheckIt 0.2 and a reorganization of sections. But my arm can be twisted to track this down if it's causing you enough trouble.

[siwelwerd]

It's low priority, but I wanted to document that it still exists and needs fixing eventually.

Do we have any particular reason to believe this is going to fix itself? Are you planning to rewrite latex_system_from_matrix ?

[StevenClontz]

Thanks for pushing on this. I was mistaken, I'd forgotten that this function was added as a CheckIt feature. (In 0.2 this will be explicit as you'll need to call CheckIt.latex_system_from_matrix(): https://github.com/StevenClontz/checkit/blob/998ec54efd4f3af441fc689299609b45d0a5a63d/platform/src/checkit/wrapper/wrapper.sage#L3 https://github.com/StevenClontz/checkit/blob/998ec54efd4f3af441fc689299609b45d0a5a63d/platform/src/checkit/wrapper/wrapper.sage#L62 )

I can't sniff out the cause of this bug for the life of me (LE4 just passes a matrix), but I'll open StevenClontz/checkit#62 to make sure it's looked at.

[siwelwerd]

I had some time this morning to try and track this down, but now I can't replicate. It definitely exists in the build I created on 1/7, but it does not appear to exist in the build Dave created on 2/1 🤷‍♂️

[StevenClontz]

It's my hope that it's an issue caused by me hacking the version of CheckIt running in the TBIL CoCalc, but I have no idea when/how that could have happened.