First Validation performance improvements

[markwpearce]

In v1, the first run still went through incremental-style dependency analysis (figuring out changed symbols, affected files, and impacted scopes).
In this branch, first validation explicitly assumes everything should be validated, which removes a lot of expensive bookkeeping on cold start.

What is different from v1 for first validation

Before (v1)

• First validation used changed-symbol and cross-scope dependency flows similar to revalidation.
• Scope invalidation and file selection were driven by dependency discovery.
• Deferred component symbol work for Brs files could run during the initial pass.
• Some validation state reset happened earlier in the pipeline.

Now (this branch)

• First validation directly includes all files in scope-context validation.
• First validation invalidates all user scopes in one shot, instead of deriving/invalidation-by-affected-file.
• Deferred component creation for Brsfiles is skipped on first validation (still done for XML and subsequent validations).
• First-validation completion now resets validation bookkeeping in the success path (after a successful run).
• Scope iteration/invalidation paths were simplified to use all user scopes directly.
• Scope lookup results per file are cached and cache is cleared when files are added/removed, reducing repeated scope-resolution cost.

Why this improves performance

The first validation run has no meaningful incremental baseline, so trying to compute precise impact sets is mostly overhead.
By switching initial validation to a full-pass strategy, we avoid dependency graph churn and repeated scope/file impact calculations, while preserving incremental behavior for subsequent validations.

Additional branch changes (non-first-validation)

• Lazy function-name computation in call validation to avoid expensive name building unless a diagnostic actually needs it.
• Type compatibility fix allowing associative-array/array-like compatibility in interface checks where appropriate.
• Added test coverage for XML component assocArray usage passed into callfunc interface parameters.

Behavior and risk notes

• Intended semantic behavior remains the same; this is primarily a performance-path change for cold validation.
• Main risk area is any workflow depending on first-pass ordering/diagnostic timing rather than final diagnostic set.
• Revalidation logic remains incremental after first successful validation.

Metrics (Example App)

Before

[04:25:25:781 PM] Validating project finished. (6s524.620ms)
[04:25:25:789 PM] Program.validate metrics:
    wait for previous run------------------------------------------------------- 000.005ms,   1 calls
    before and on programValidate----------------------------------------------- 000.145ms,   1 calls
    get files to be validated--------------------------------------------------- 001.275ms,   1 calls
    add component reference types----------------------------------------------- 000.669ms,   1 calls
    validateFile------------------------------------------------------------- 02s027.987ms, 839 calls
    do deferred component creation---------------------------------------------- 057.636ms, 839 calls
    build component types for any component that changes------------------------ 056.637ms,   1 calls
    track and update type-time and runtime symbol dependencies and changes------ 008.704ms,   1 calls
    tracks changed symbols and prepares files and scopes for validation--------- 317.518ms,   1 calls
    invalidate affected scopes-------------------------------------------------- 026.028ms, 839 calls
    checking scopes to validate------------------------------------------------- 000.214ms,   1 calls
    beforeScopeValidate--------------------------------------------------------- 000.105ms, 294 calls
    validate scope----------------------------------------------------------- 03s955.053ms, 294 calls
    afterValidateScope---------------------------------------------------------- 068.071ms, 294 calls
    detect duplicate component names-------------------------------------------- 001.095ms,   1 calls
    Total-------------------------------------------------------------------- 06s521.140ms,

After

[09:35:21:067 PM] Validating project finished. (5s298.532ms)
[09:35:21:074 PM] Program.validate metrics:
    wait for previous run------------------------------------------------------- 000.005ms,   1 calls
    before and on programValidate----------------------------------------------- 000.148ms,   1 calls
    get files to be validated--------------------------------------------------- 001.263ms,   1 calls
    add component reference types----------------------------------------------- 000.610ms,   1 calls
    validateFile------------------------------------------------------------- 01s653.655ms, 839 calls
    do deferred component creation---------------------------------------------- 000.626ms, 839 calls
    build component types for any component that changes------------------------ 053.809ms,   1 calls
    track and update type-time and runtime symbol dependencies and changes------ 007.588ms,   1 calls
    tracks changed symbols and prepares files and scopes for validation--------- 242.128ms,   1 calls
    invalidating file segments-------------------------------------------------- 001.090ms, 839 calls
    invalidate affected scopes-------------------------------------------------- 000.141ms,   1 calls
    checking scopes to validate------------------------------------------------- 000.197ms,   1 calls
    beforeScopeValidate--------------------------------------------------------- 000.098ms, 294 calls
    validate scope----------------------------------------------------------- 03s282.035ms, 294 calls
    afterValidateScope---------------------------------------------------------- 052.066ms, 294 calls
    detect duplicate component names-------------------------------------------- 000.792ms,   1 calls
    Total-------------------------------------------------------------------- 05s296.251ms,

Roughly, a 15-20% improvement on first validation time.