ZacharyZhang-NY
e74dafcd36
Three independently ugly bugs that all fed the same log flood:
* Electron's `WebFrameMain.send` writes every "Render frame was disposed"
to stderr via `console.error`. When the worker fans a second
agent-event past a reloading renderer, we produced thousands of
identical lines and stalled the main loop on stderr I/O. Added a
`mainFrame.detached` guard so the send is short-circuited before
Electron's own catch block fires, and installed a targeted filter in
`suppress-warnings.ts` (now actually imported from the entry point)
so the specific `console.error("Error sending from webFrameMain:"…)`
noise never reaches the terminal. Any other payload still surfaces.
* `dispatchAgentMessage` accepted the human-reader sentinel as a wake
target when an agent replied to `__board__` — `wakeAgentIfPossible`
then called `db.getAgent("__board__")`, threw, and the dispatcher
kept repeating on every send. Strip `HUMAN_SENDER_ID` at
`resolveMessageWakeTargets`, and keep a defense-in-depth guard in the
heartbeat so a stray caller (future automation fan-out) cannot
reintroduce the crash.
* Pi's skills loader rejects any SKILL.md whose frontmatter `name`
diverges from its parent directory. The four bundled `my-writing-style`
copies declared `name: writing-style-skill`, so every install emitted
a warning and dropped the skill. They also lived as orphan gitlinks
(mode 160000 without .gitmodules), and the global `*.md` gitignore
hid every other SKILL.md from version control. Fixed the frontmatter
to match the dir, converted the dangling submodules to plain trees,
and allowlisted `builtin-skills/**/*.md` so a fresh clone carries the
catalogue instead of silently losing it.
3.0 KiB
3.0 KiB
title, impact, impactDescription, tags
| title | impact | impactDescription | tags |
|---|---|---|---|
| Use Partial Boundaries When Full Separation is Premature | MEDIUM-HIGH | prepares for future without over-engineering now | bound, partial-boundaries, pragmatism, yagni |
Use Partial Boundaries When Full Separation is Premature
When you anticipate needing a boundary but the cost of full separation is too high, implement a partial boundary. This prepares for future separation without premature over-engineering.
Incorrect (no boundary when one might be needed):
# Tightly coupled - if we ever need to split, massive refactor
class ReportService:
def generate_sales_report(self, start_date, end_date):
# Direct database access
conn = psycopg2.connect(DATABASE_URL)
cursor = conn.cursor()
cursor.execute("""
SELECT * FROM sales
WHERE date BETWEEN %s AND %s
""", (start_date, end_date))
# Direct PDF generation
pdf = FPDF()
for row in cursor.fetchall():
pdf.cell(200, 10, txt=str(row))
return pdf.output()
# Cannot split report generation from data access without rewriting
Correct (partial boundary - same component, prepared for split):
# Strategy 1: Facade pattern (simplest partial boundary)
class ReportingFacade:
"""Single entry point, internal implementation can be refactored later."""
def generate_sales_report(self, request: ReportRequest) -> Report:
data = self._fetch_data(request)
return self._format_report(data, request.format)
def _fetch_data(self, request):
# Could become a separate DataAccess component later
return self.repository.get_sales(request.start, request.end)
def _format_report(self, data, format):
# Could become a separate ReportRenderer component later
if format == 'pdf':
return PdfRenderer().render(data)
return CsvRenderer().render(data)
# Strategy 2: Interface with single implementation
class SalesDataProvider(Protocol):
def get_sales(self, start: date, end: date) -> list[Sale]: ...
class PostgresSalesProvider:
def get_sales(self, start: date, end: date) -> list[Sale]:
# Implementation here
# Both live in same component, but interface enables future split
# When split needed: move interface to one component, impl to another
# Strategy 3: One-dimensional boundary
# Skip reciprocal interface - simpler but weaker protection
class ReportGenerator:
def __init__(self, data_provider: SalesDataProvider):
self.data_provider = data_provider # Depends on abstraction
# SalesDataProvider implementation knows about ReportGenerator
# Not fully decoupled, but good enough for now
When to use partial boundaries:
- You suspect a boundary will be needed, but not yet
- Team is small and deployment is unified
- Cost of full boundary outweighs current benefits
Reference: Clean Architecture - Partial Boundaries