Add More tests

[programmeroftheeve]

Adds tests for the additional load functions.
Contains commits from #100

[codecov]

Codecov Report

Merging #101 (e58bcc1) into master (52ef020) will increase coverage by 1.42%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #101      +/-   ##
==========================================
+ Coverage   81.32%   82.75%   +1.42%     
==========================================
  Files          12       12              
  Lines        1462     1461       -1     
==========================================
+ Hits         1189     1209      +20     
+ Misses        273      252      -21     

Impacted Files	Coverage Δ
src/scanner.jl	83.40% <0.00%> (+0.40%)	⬆️
src/YAML.jl	91.17% <0.00%> (+52.94%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 52ef020...e58bcc1. Read the comment docs.

[programmeroftheeve]

Ok, so I am glad it wasn't my multi-constructor code from earlier. But I am confused on why running load_all fails on empty_scaler.yml reading from the read string on CI only.
It would have to something weird with IOBuffer.

[kescobo]

Thanks for digging into it! Definitely weird - here's the first bit of the stacktrace (I'm sure you've seen it, just pulling it out for others)

Load All from String: Error During Test at D:\a\YAML.jl\YAML.jl\test\runtests.jl:168
  Test threw exception
  Expression: equivalent(first(allStringData), expected)
  AssertionError: typeof(forward!(composer.input)) == DocumentStartEvent
  Stacktrace:
   [1] compose_document(::YAML.Composer) at D:\a\YAML.jl\YAML.jl\src\composer.jl:49
   [2] compose(::YAML.EventStream) at D:\a\YAML.jl\YAML.jl\src\composer.jl:38

Have you pulled out and looked at what first(allStringData) is and compared it with expected? I understand it's tough to debug given that it's not happening locally, but even throwing in some print statements might help

[programmeroftheeve]

@kescobo So after putting in debug statements and trying some data massaging, the empty_scalar doesn't like a new line at the end of the string. I am not sure why this is, but it is a place to start.

[kescobo]

Windows often uses different line endings (\r I think) - not sure why it would only happen on the ci machine, but I'd be content to chomp the line endings and just compare the line contents if that's possible for this test (on phone so can't easily check). I'm pretty sure julia's rstrip can do this.

[programmeroftheeve]

Chomped the strings. Everything passes. Yay! 😁

[kescobo]

Thank you! I'll try to review and merge tomorrow

[kescobo]

Looks great to me - I'll open an issue about this line ending problem - should be possible to do platform-dependent tests so that we don't have to chomp line endings, but this is good for now I think