100% Test Coverage (#19)

* #8 Fleshed out unit tests, achieving 100% coverage
* Revised unit test names and added comments to define a rough spec
* Ignored difficult to reproduce failure recovery code from code coverage
* Removed unnecessary checks from internal class `BundleSource`
* Separated merge and validation logic into multiple files to improve maintainability
* Fixed error in duplicate detection logic of config validation
* #21 Fixed issue where a hang could occur due to a paused internal stream
* #22 Fixed flaw in results callback data collection where styles would replace styles within the same bundle
* Improved logging, particularly around source identification
* Removed redundant double usage of nyc test coverage wrapper
* Updated `CHANGELOG.md` and `README.md`

.travis.yml

@@ -1,3 +1,5 @@
+os:
+  - linux
 language: node_js
 node_js:
   - "8"

CHANGELOG.md

@@ -7,6 +7,16 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 
 ## [Unreleased]
 
+This release focused on internal refactoring and improving test coverage. Many discoverd bugs have been fixed.
+
+### Fixed
+- Potential deadlock in bundles processor.
+- Swallowing of exceptions raised within bundler factory and bundler factory streams.
+- Virtual path rules in a config not being correctly checked for duplicate matchers.
+- Fixed edge case where if the first file in a bundle failed to a resolve a file the resulting exception would never bubble up due to the source stream never being unpaused. #21
+- Fixed issue where bundles using both `scripts` and `styles` would result in only files emitted as part of `styles` bundling appearing in the results callback data. #22
+- Other undocumented assorted edge cases.
+
 ## [3.0.0-rc.1] - 2019-01-09
 
 This release focuses on simplifying the package for UserFrosting 4 to improve maintainability. Features unsupported by UF4 are largely removed.

README.md

@@ -7,6 +7,8 @@
 
 Orchastrates JS and CSS bundle creation in a highly efficient and configurable manner.
 
+**CAUTION** The implementation currently produces a great deal backpressure. This can result in signficiant RAM usage. Projects dealing with a significant number resources are better off not using this tool until the custom stream source is implemented in v4.
+
 ## Install
 
 ```bash