Performance improvements for large projects

[JustNoon]

Performance changes:

• lastMacroCache is now invalidated using lodash after parameter comparison. This was on the order of 1000x faster in worst case scenarios (~4000 macros)
• Initial passage collection and parsing now runs asynchronously. This was on the order of 10x faster in worst case scenarios. (~500 files/ 10k passages)
• parseConfiguration now runs asynchronously. Each additional configuration file now introduces minimal performance impact

Function changes:

• changeStoryFormat return type changed from Promise<void> to Promise<vscode.TextDocument>
• yamlParse made into separate function for easier error management and future language specific handling. configuration.ts:115
• parseRawText and parseText now accept an extra optional parameter that bypasses updating workspaceState. Passages will be resolved through or stored in this parameter
• macroList and enumList are explicitely async functions now instead of TS rewriting them
• hover and hoverProvider now respect cancellation tokens
• isObjectSimpleEqual has been commented out in favor of lodash's isEqual

Bug Fixes:

• Parameter compare could crash the extension if parameters failed to load completely

Package updates:

• glob updated to 7.2.0
• lodash 4.17.21 installed

Future considerations:

• glob can be removed by switching to vscode.workspace.findFiles which is faster in almost all circumstances (it manages its own cache)
• updateDiagnostics and parseText can be optimized further
• macroCache updating could be simplified and improved to run asynchronously without so many loops