Improve processor.go behaviour

[Vladyslav-Kuksiuk]

This PR:

• removes code duplication for check and embed modes in the processor.go;
• extracts ProcessingError as typed error;
• Divides SetEmbedding to StartEmbedding and FinishEmbedding methods.

[Copilot]

Pull request overview

This PR refactors the embedding processor flow to reduce duplication between embed and check modes, improves error typing for downstream handling, and clarifies parser context state transitions around embedding lifecycle.

Changes:

• Deduplicates per-doc processing loops via a shared processRequiredDocs helper used by both EmbedAll and up-to-date checking.
• Introduces a typed embedding.ProcessingError that attaches doc path + line and preserves underlying parse/source-resolution error types via Unwrap.
• Splits parsing context embedding lifecycle from SetEmbedding into explicit StartEmbedding / FinishEmbedding.

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
embedding/processor.go	Removes duplicated loops by introducing processRequiredDocs; wraps parsing errors with ProcessingError.
embedding/parsing/instruction_token.go	Switches to Context.StartEmbedding when an instruction is successfully parsed.
embedding/parsing/context.go	Replaces SetEmbedding with explicit StartEmbedding/FinishEmbedding embedding lifecycle methods.
embedding/parsing/code_fence_end.go	Uses FinishEmbedding when closing an embedding at the end of a code fence.
embedding/error.go	Adds ProcessingError typed error with user-facing formatting and Unwrap.
embedding/embedding_test.go	Adds assertions for typed error preservation (ProcessingError, parse errors, and pattern errors).

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.