Expand ARCHITECTURE.md: matching hash, auto-issuance, stage mapping

[mejango]

Summary

• Expand the matching hash section: explain how it is computed (incremental ABI encoding + keccak256), which fields it covers (base currency, name, ticker, salt, all per-stage economic fields, all auto-issuance entries), and why splits are excluded (chain-specific addresses). Explain how the hash is used as a CREATE2 salt component for sucker deployment to enforce cross-chain config parity.
• Expand the auto-issuance section: explain the REVAutoIssuance struct, chainId-based filtering at deploy time, deferred claiming via autoIssueFor, and stage ID assignment.
• Add a "Stage-to-Ruleset Mapping" section with a field-by-field table showing how REVStageConfig maps to JBRulesetConfig, plus the automatically-set metadata fields.
• Add two new Key Design Decisions entries (deferred auto-issuance rationale, no approval hook rationale).

Test plan

• Verify all field mappings against _makeRulesetConfiguration in REVDeployer.sol
• Verify matching hash field list against _makeRulesetConfigurations encoding logic
• Verify auto-issuance flow against autoIssueFor function

🤖 Generated with Claude Code