Parsing: Enforce block naming requirements consistently

[aduth]

This pull request seeks to resolve inconsistencies between block registration validation and name parsing expectations. Specifically:

• Registration requires that a block name be lowercase, but this was not enforced by the block parser
• The block parser enforces that a block name start with a letter, but this was not validated by registration

Further, these requirements were not documented in block registration documentation.

The proposed changes remedies each of the above issues.

Testing instructions:

Ensure tests pass:

npm test

Follow-up Tasks:

Gutenberg examples have names beginning with numbers and must be updated:

https://github.com/WordPress/gutenberg-examples

[codecov]

Codecov Report

Merging #3521 into master will increase coverage by 1.08%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #3521      +/-   ##
==========================================
+ Coverage   34.54%   35.62%   +1.08%     
==========================================
  Files         261      262       +1     
  Lines        6710     6896     +186     
  Branches     1225     1286      +61     
==========================================
+ Hits         2318     2457     +139     
- Misses       3704     3738      +34     
- Partials      688      701      +13

Impacted Files	Coverage Δ
blocks/api/registration.js	100% <100%> (ø)	⬆️
blocks/library/shortcode/index.js	33.33% <0%> (-6.67%)	⬇️
blocks/library/html/index.js	14.28% <0%> (-2.39%)	⬇️
blocks/library/more/index.js	20% <0%> (-2.23%)	⬇️
blocks/hooks/index.js	100% <0%> (ø)	⬆️
editor/writing-flow/index.js	0% <0%> (ø)	⬆️
editor/store-persist.js	100% <0%> (ø)	⬆️
blocks/hooks/custom-class-name.js	84.61% <0%> (ø)
blocks/api/parser.js	98.48% <0%> (+0.4%)	⬆️
blocks/library/list/index.js	7.31% <0%> (+0.42%)	⬆️
... and 8 more

---

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 e7aaa99...df43eb4. Read the comment docs.

[dmsnell]

this seems a bit verbose. what if instead we ripped out the whole ASCII nomenclature and just simplified down to something like Block_Name or Block_Name_Part - we're not using the other rules so we can get rid of them

Namespaced_Block_Name
  = $( Block_Name_Part "/" Block_Name_Part )

Core_Block_Name
  = type:Block_Name_Part { return 'core/' + type }

Block_Name_Part
  = $([a-z] [a-z0-9_-]*)

[aduth]

this seems a bit verbose. what if instead we ripped out the whole ASCII nomenclature and just simplified down to something like Block_Name or Block_Name_Part

Sounds good to me. Applied in df43eb4.

[dmsnell]

looks nice!