Lexery - Branch Divergence
Why This Matters
Одна з найважливіших truths про сучасний Lexery: проект існує одночасно в двох активних гілках розвитку.
Exact Observed Snapshot
Local branch vs remote brain branch
- Local branch:
legal-agent-brain-dev - Relative to
origin/legal-agent-brain-dev:0remote-only commits,10local-only commits
Local branch vs remote default branch
- Relative to
origin/dev:21remote-only commits,26local-only commits
What Lives On origin/dev
- shared contracts package
- auth refactor
- auth screens
- richer profile/user metadata shape
- subscription plan UI
- workspace/sidebar/product-shell refinements
- open auth registration infra PR
What Lives On legal-agent-brain-dev
- bounded legal orchestrator
- clarification pause/resume hardening
- deterministic ORCH cost cuts
- DocList recovery
- brain-admin proposal queue wiring
- retry reuse and tail-latency tuning
Best Interpretation
Observed:origin/devis the product integration branch.Observed:legal-agent-brain-devis the runtime/brain innovation branch.Inferred: these are not trivial parallel features; they are partially different product realities.
Consequences
Positive
- The team can move fast on product shell and brain quality independently.
Negative
- Root README can drift badly.
- Contracts can diverge.
- Plan codes and feature assumptions can split.
- Integration debt compounds silently.
Documentary Drift Caused By This Split
- Root
README.mdstill saysapps/portalis a placeholder. - Current code proves
apps/portalis already real. - Some portal docs lag behind actual auth/chat/product progress.
- Azure docs describe a deployment target that current live code only partially proves.
Strategic Reading
Lexery’s current challenge is not lack of architecture. It is reconciling:
- a fast-moving legal brain branch
- a fast-moving product shell branch
- and the contracts that must sit between them