Fix decode_ignored_any

[DrewMcArthur]

this beefs up the implementation for ignoring values, mostly lifted from serde's implementation. added a bunch of tests, which pass and these changes also fix the broken tests that led me to these changes in another repository of mine that's making use of this library.

• would appreciate a closer look / double checking my implementations of some of the more base level helper functions here (e.g. peek_or_null, eat_char, etc.)
• definitely take a look at the tests to see if there's anything i missed there that you'd like to cover
• i think there's still some cleanup i can do; i'll take another pass at this later today.

[DrewMcArthur]

closes #2

[haydnv]

Thanks Drew! Do you think you could bring some of the parsing improvements into the parse_* methods? For example parse_number could call parse_decimal, parse_exponent, etc. This could be in a separate PR if you prefer.

[DrewMcArthur]

@haydnv thanks for the review -

1. i opened refactor parse_* methods #5 for refactoring the parse methods. i'd be open to it, but first would like to implement some benchmark tests to make sure any changes only speed things up!
2. I think I addressed all the comments. would appreciate another glance to make sure everything is covered.
3. only things missing are some "expected failures" tests. i can add those here if you'd like, or we can open an issue to make sure they eventually get added.

[haydnv]

It would be great if you can add the expected failure tests in this PR

[DrewMcArthur]

okay! @haydnv sorry for the delay. i've added a bunch more test cases, and ran cargo llvm-cov --features all --output-path lcov.info --lcov to verify that all the code paths in these ignore_... functions are covered.

[haydnv]

Thanks Drew, great work!