nav() now displays some debugging information on KeyError

[linsomniac]

There's a user reporting a KeyError in nav(), so I've changed it to print out the key and dict for future debugging. Also, a bare Exception was used to capture a dictionary failure, removed that since my change checks for the existence of the key in the dictionary. If this is objectionable at all, you probably should change it to catching a KeyError rather than catching Exception.

This is to gather information for #539

[sigma67]

Did you test this change?

nav has more than 250 usages across the project, it's not something to change as a drive-by

Simply running the get_album tests fails instantly - you didn't consider that items may contain integer indices

[linsomniac]

I couldn't originally get the tests working, I believe this change addresses that problem.

[sigma67]

Sorry, it still doesn't work.

TypeError: 'NoneType' object is not subscriptable

[linsomniac]

Sorry for the delay in getting back to this, busy end of the week. I missed that exception, thanks for pointing it out. I think this modification resolves that issue.

One issue I'm running into that perhaps you can help with is that I'm getting a bunch of config-related test failures, and I didn't see that None error among those other ones. The test instructions say to run: cp tests/test.example.cfg tests/test.cfg, but the repo doesn't seem to have a "test.example.cfg" file in it, or any "cfg" file AFAICT.

[sigma67]

Sorry about the test.cfg issue, I accidentally deleted the file instead of renaming it due to the presence of a gitignore on **.cfg

[sigma67]

What's the advantage of moving the try-except inside the loop? Isn't that even more overhead?

[linsomniac]

Having it in the loop means it has access to the value of "k" to be able to log. It could be moved outside the loop but would lose that data, but might still provide enough information to track down, given more effort.

[sigma67]

But it's bad: https://docs.astral.sh/ruff/rules/try-except-in-loop/

And with nav it matters, because it's used a lot.

[linsomniac]

[I] sean@seans-laptop ~/p/realgoansible (master)> python3 -m timeit 'for x in range(10):' '   y = x + 1'
2000000 loops, best of 5: 199 nsec per loop
[I] sean@seans-laptop ~/p/realgoansible (master)> python3 -m timeit 'for x in range(10):' '   try:' '      y = x + 1' '
except:' '      pass'
1000000 loops, best of 5: 202 nsec per loop

LSP is telling me that "k is unbound" if I move the try/except out of the loop, but if I run a test I am getting the loop variable past the loop:

try:
    for x in [1,2,3,4]:
        if x == 3:
            1 / 0
except Exception as e:
    print(f"Failed, {x}")

Results in: "Failed, 3".

So, I've moved it out of the loop.