Current coverage is 98.66%

Merging #179 into master will decrease coverage by 1.33%

1. File ...dm/_tqdm_notebook.py (not in diff) was modified. more
   • Misses +6
   • Partials 0
   • Hits -6

@@             master       #179   diff @@
==========================================
  Files             8          8          
  Lines           447        449     +2   
  Methods           0          0          
  Messages          0          0          
  Branches         80         80          
==========================================
- Hits            447        443     -4   
- Misses            0          6     +6   
  Partials          0          0          

Powered by Codecov. Last updated by 5f7c7e4...8b5a87e

Coverage decreased (-1.3%) to 98.664% when pulling 8b5a87e on misc_fixes into 5f7c7e4 on master.

Ow, that's interesting, tqdm_notebook is not at all covered now, is it because of the pragma no cover in the adapters or because the import is correctly delayed and thus coverage can't access the module?

I agree with this PR, it can be merged (after adding pragma no cover to tqdm_notebook). I would prefer to generalize the delayed/lazy import for all modules, but that will do for a quick fix, it can be enhanced later.

Coverage decreased (-1.3%) to 98.664% when pulling f492111 on misc_fixes into 5f7c7e4 on master.

I'm not a fan of adding pragma no cov to code which strictly speaking should be covered... Is there no way to test widgets? redirect html output etc? I realise this was probably asked before @lrq3000

Coverage decreased (-1.3%) to 98.664% when pulling 7b30463 on misc_fixes into 5f7c7e4 on master.

@casperdcl Sorry for the delayed answer. Yes it is possible but way too complicated at this stage for me. We need to simulate a fake JS server and IPython API mockup, like the ipywidgets project does. I tried to replicate what they did but it was way over my head...