Other weight generators still work around correction matrices

[BorisVarbanov]

In the sig_test_manhattan branch Gardener was changed to transition from being passed correction matrices to being instead passed code layouts, both of which can determine if an error chain requires a logical correction. However the decoder templates and other weight generators are still based around weight matrices.

[obriente]

I don't think I completely understand the issue - the weight matrices are separate to the code layout and both need to be passed around.

It might make sense to start storing weight matrices in the code layout though, if you think this is a good idea?

[BorisVarbanov]

I am really sorry for the confusion, I meant to refer to the correction matrices in the last sentence, not the weight matrices. To be clear, the issue is that the correction matrices haven't consistently been removed across all files and that I am aware of it.

[obriente]

Not a problem - the confusion actually raised a reasonable idea.

Let me know if you want me to take a look at anything.