nogipx / rpc_dart / 26113930664
80%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 72
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 80.161%. Remained the same
26113930664

push

github

nogipx 
feat(rpc_dart_log): add MCP server, rename logview -> LogCollector

Add MCP server (dart_mcp) exposing logview_get_logs and
logview_list_devices tools for AI assistants.

Rename all Logview* classes to LogCollector* for consistency
with the package name. Add global executables:
  rpc_dart_log     -- CLI collector
  rpc_dart_log_mcp -- MCP server for Claude Code

refactor(rpc_dart_log): single executable with collector + MCP

Merge CLI collector and MCP server into one process. rpc_dart_log
starts WebSocket collector and MCP server simultaneously.
Logs render to stderr, MCP protocol on stdout.

Remove separate collector_mcp entry point. One executable: rpc_dart_log.

fix(rpc_dart_log): rename MCP tools to rpc_log_get_logs, rpc_log_list_devices

refactor(rpc_dart_log): switch MCP to HTTP transport

Replace dart_mcp stdio with plain HTTP JSON-RPC via shelf.
MCP server runs on separate port (default 9501) alongside
WebSocket collector (default 9500). No stdout conflicts.

Remove dart_mcp, shelf_router dependencies. Add shelf.
Single executable: rpc_dart_log (both collector + MCP).

fix(rpc_dart_log): implement OAuth 2.1 flow for HTTP MCP transport

Claude Code requires OAuth for HTTP MCP servers. Implement minimal
OAuth 2.1 endpoints: discovery, dynamic client registration, authorize
(redirect with code), token exchange. All return dummy values for
local dev use.

Fix shelf Response.ok with Future body (use async handler).

feat(rpc_dart_log): add rpc_log_sources tool, enhance rpc_log_get_logs

New MCP tool rpc_log_sources: shows connected devices, scopes with
record counts per level, time range, active traceIds. AI calls this
first to understand what data is available before querying.

Enhanced rpc_log_get_logs: add device, traceId, cursor filters.
Cursor enables incremental tail (get only new records since last call).
Records now include device labels in output.

MCP instructions guide AI investigation strategy: sources first,
then targeted query, then ... (continued)

5063 of 6316 relevant lines covered (80.16%)

9.0 hits per line

Jobs
ID Job ID Ran Files Coverage
1 26113930664.1 72
80.16
 GitHub Action Run
Source Files on build 26113930664