hermes-agent/agent/lsp/eventlog.py
Teknium 83b93898c2
feat(lsp): semantic diagnostics from real language servers in write_file/patch (#24168)
* feat(lsp): semantic diagnostics from real language servers in write_file/patch

Wire ~26 language servers (pyright, gopls, rust-analyzer, typescript-language-server,
clangd, bash-language-server, ...) into the post-write lint check used by write_file
and patch. The model now sees type errors, undefined names, missing imports, and
project-wide semantic issues introduced by its edits, not just syntax errors.

LSP is gated on git workspace detection: when the agent's cwd or the file being
edited is inside a git worktree, LSP runs against that workspace; otherwise the
existing in-process syntax checks are the only tier. This keeps users on
user-home cwds (Telegram/Discord gateway chats) from spawning daemons.

The post-write check is layered: in-process syntax check first (microseconds),
then LSP semantic diagnostics second when syntax is clean. Diagnostics are
delta-filtered against a baseline captured at write start, so the agent only
sees errors its edit introduced. A flaky/missing language server can never
break a write -- every LSP failure path falls back silently to the syntax-only
result.

New module agent/lsp/ split into:

- protocol.py: Content-Length JSON-RPC framer + envelope helpers
- client.py: async LSPClient (spawn, initialize, didOpen/didChange,
  ContentModified retry, push/pull diagnostic stores)
- workspace.py: git worktree walk-up + per-server NearestRoot resolver
- servers.py: registry of 26 language servers (extension match,
  root resolver, spawn builder per language)
- install.py: auto-install dispatch (npm install --prefix, go install
  with GOBIN, pip install --target) into HERMES_HOME/lsp/bin/
- manager.py: LSPService (per-(server_id, root) client registry, lazy
  spawn, broken-set, in-flight dedupe, sync facade for tools layer)
- reporter.py: <diagnostics> block formatter (severity-1-only, 20-per-file)
- cli.py: hermes lsp {status,list,install,install-all,restart,which}

Wired into tools/file_operations.py:

- write_file/patch_replace now call _snapshot_lsp_baseline before write
- _check_lint_delta gains a third tier: LSP semantic diagnostics when
  syntax is clean
- All LSP code paths swallow exceptions; write_file's contract unchanged

Config: 'lsp' section in DEFAULT_CONFIG with enabled (default true),
wait_mode, wait_timeout, install_strategy (default 'auto'), and per-server
overrides (disabled, command, env, initialization_options).

Tests: tests/agent/lsp/ -- 49 tests covering protocol framing (encode and
read_message round-trip, EOF/truncation/missing Content-Length), workspace
gate (git walk-up, exclude markers, fallback to file location), reporter
(severity filter, max-per-file cap, truncation), service-level delta filter,
and an in-process mock LSP server that exercises the full client lifecycle
including didChange version bumps, dedup, crash recovery, and idempotent
teardown.

Live E2E verified end-to-end through ShellFileOperations: pyright
auto-installed via npm into HERMES_HOME, baseline captured, type error
introduced, single delta diagnostic surfaced with correct line/column/code/
source, then patch fix removes the diagnostic from the output.

Docs: new website/docs/user-guide/features/lsp.md page covering supported
languages, configuration knobs, performance characteristics, and
troubleshooting; cli-commands.md updated with the 'hermes lsp' reference;
sidebar updated.

* feat(lsp): structured logging, backend gate, defensive walk caps

Cherry-picks the substantive ideas from #24155 (different scope, same
problem space) onto our PR.

agent/lsp/eventlog.py (new): dedicated structured logger
``hermes.lint.lsp`` with steady-state silence. Module-level dedup sets
keep a 1000-write session at exactly ONE INFO line ("active for
<root>") at the default INFO threshold; clean writes log at DEBUG so
they never reach agent.log under normal config. State transitions
(server starts, no project root for a file, server unavailable) fire
at INFO/WARNING once per (server_id, key); novel events (timeouts,
unexpected errors) fire WARNING per call. Grep recipe: ``rg 'lsp\\['``.

agent/lsp/manager.py: wire the eventlog into _get_or_spawn and
get_diagnostics_sync so users can answer "did LSP fire on this edit?"
with a single grep, plus surface "binary not on PATH" warnings once
instead of silently retrying every write.

tools/file_operations.py: backend-type gate. ``_lsp_local_only()``
returns False for non-local backends (Docker / Modal / SSH /
Daytona); ``_snapshot_lsp_baseline`` and ``_maybe_lsp_diagnostics``
now skip entirely on remote envs. The host-side language server
can't see files inside a sandbox, so this prevents pretending to
lint a file the host process can't open.

agent/lsp/protocol.py: 8 KiB cap on the header block in
``read_message``. A pathological server that streams headers
without ever emitting CRLF-CRLF would have looped forever consuming
bytes; now raises ``LSPProtocolError`` instead.

agent/lsp/workspace.py: 64-step cap on ``find_git_worktree`` and
``nearest_root`` upward walks, plus try/except containment around
``Path(...).resolve()`` and child ``.exists()`` calls. Defensive
against pathological inputs (symlink loops, encoding errors,
permission failures mid-walk) — the lint hook is hot-path code and
must never raise.

Tests:
- tests/agent/lsp/test_eventlog.py: 18 tests covering steady-state
  silence (clean writes stay DEBUG), state-transition INFO-once
  semantics (active for, no project root), action-required
  WARNING-once (server unavailable), per-call WARNING (timeouts,
  spawn failures), and the "1000 clean writes => 1 INFO" contract.
- tests/agent/lsp/test_backend_gate.py: 5 tests verifying
  _lsp_local_only / snapshot_baseline / maybe_lsp_diagnostics skip
  the LSP layer for non-local backends and route correctly for
  LocalEnvironment.
- tests/agent/lsp/test_protocol.py: new test_read_message_rejects_runaway_header
  exercising the 8 KiB cap.

Validation:
- 73/73 LSP tests pass (49 original + 18 eventlog + 5 backend-gate + 1 framer cap)
- 198/198 pass when run alongside existing file_operations tests
- Live E2E re-run with pyright still surfaces "ERROR [2:12] Type
  ... reportReturnType (Pyright)" through the full path, then patch
  fix removes it on the next call.

* feat(lsp): atexit cleanup + separate lsp_diagnostics JSON field

Two improvements salvaged from #24414's plugin-form alternative,
keeping our core-integrated design:

1. atexit cleanup of spawned language servers
   ----------------------------------------------------------------
   ``agent/lsp/__init__.get_service`` now registers an ``atexit``
   handler on first creation that tears down the LSPService on
   Python exit.  Without this, every ``hermes chat`` exit was
   leaking pyright/gopls/etc. processes for a few seconds while
   their stdout buffers drained -- they got reaped by the kernel
   eventually but a watchful ``ps aux`` would catch them.

   The handler runs once per process (gated by
   ``_atexit_registered``); idempotent ``shutdown_service``
   ensures double-fire is a no-op.  Errors during shutdown are
   swallowed at debug level since by the time atexit fires the
   user has already seen the agent's final response.

2. Separate ``lsp_diagnostics`` field on WriteResult / PatchResult
   ----------------------------------------------------------------
   Previously the LSP layer folded its diagnostic block into the
   ``lint.output`` string, conflating the syntax-check tier with
   the semantic tier.  The agent (and any downstream parsers) now
   read syntax errors and semantic errors as independent signals:

       {
         "bytes_written": 42,
         "lint": {"status": "ok", "output": ""},
         "lsp_diagnostics": "<diagnostics file=...>\nERROR [2:12] ..."
       }

   ``_check_lint_delta`` returns to its original two-tier shape
   (syntax check + delta filter); ``write_file`` and
   ``patch_replace`` independently fetch LSP diagnostics via
   ``_maybe_lsp_diagnostics`` and pass them into the new field.
   ``patch_replace`` propagates the inner write_file's
   ``lsp_diagnostics`` so the outer PatchResult carries the patch's
   delta correctly.

Tests: 19 new
- tests/agent/lsp/test_lifecycle.py (8 tests): atexit registration
  fires once and only once across N get_service calls; the
  registered callable is our internal shutdown wrapper;
  shutdown_service is idempotent and safe when never started;
  exceptions during shutdown are swallowed; inactive service is
  cached so we don't rebuild on every check.
- tests/agent/lsp/test_diagnostics_field.py (11 tests): WriteResult
  / PatchResult dataclass shape, to_dict include/omit semantics,
  channel separation (lint and lsp_diagnostics carry independent
  signals), write_file populates the field via
  _maybe_lsp_diagnostics only when the syntax tier is clean,
  patch_replace propagates the field forward from its internal
  write_file.

Validation:
- 92/92 LSP tests pass (73 prior + 8 lifecycle + 11 diagnostics field)
- 217/217 pass with file_operations + LSP combined
- Live E2E reverified: clean writes -> both fields empty/none; type
  error introduced -> lint clean (parses), lsp_diagnostics carries
  the pyright reportReturnType block; patch fix -> both fields
  clean again.

* fix(lsp): broken-set short-circuit so a wedged server isn't paid every write

Discovered while auditing failure paths: a language server binary that
hangs (sleep forever, no LSP traffic on stdin/stdout) caused EVERY
subsequent write to re-pay the 8s snapshot_baseline timeout. Five
writes = ~64s of dead time.

The bug: ``_get_or_spawn`` adds the (server_id, root) pair to
``_broken`` inside its inner exception handler, but when the OUTER
``_loop.run`` timeout fires, it cancels the inner task before that
handler runs. The pair never makes it to broken-set, so the next
write re-enters the spawn path and re-pays the timeout.

Fix:

- New ``_mark_broken_for_file`` helper at the service layer marks
  the (server_id, workspace_root) pair broken from the OUTSIDE when
  the outer timeout fires. Called from the except branches in
  ``snapshot_baseline``, ``get_diagnostics_sync`` (asyncio.TimeoutError
  + generic Exception). Also kills any orphan client process that
  survived the cancelled future, fire-and-forget with a 1s ceiling.

- ``enabled_for`` now consults the broken-set BEFORE returning True.
  Files in already-broken (server_id, root) pairs short-circuit to
  False, so the file_operations layer skips the LSP path entirely
  with no spawn cost. Until the service is restarted (``hermes lsp
  restart``) or the process exits.

- A single eventlog WARNING is emitted on first mark-broken so the
  user knows which server gave up. Subsequent edits in the same
  project stay silent.

Tests: 7 new in tests/agent/lsp/test_broken_set.py — covers the
key shape (server_id, per_server_root), enabled_for short-circuit,
sibling-file skip in same project, project isolation (broken in
A doesn't affect B), graceful no-op for missing-server / no-workspace,
and an end-to-end test that snapshots after a failure and verifies
the next ``enabled_for`` returns False.

Validation:

- Live retest of the wedged-binary scenario: 5 sequential writes,
  first 8.88s (the one snapshot timeout), subsequent four ~0.84s
  (no LSP cost). Down from 5x12.85s = 64s before this fix.
- 99/99 LSP tests pass (92 prior + 7 broken-set)
- 224/224 pass with file_operations + LSP combined
- Happy path E2E reverified — clean write, type error introduced,
  patch fix all behave correctly with the new broken-set logic.

Note: the FIRST write to a wedged binary still pays 8s (the
snapshot_baseline timeout). We could shorten that, but pyright/
tsserver normally take 2-3s and slow CI rust-analyzer can need
5+ seconds, so 8s is the conservative ceiling. Subsequent writes
are instant.
2026-05-12 16:31:54 -07:00

213 lines
7.6 KiB
Python

"""Structured logging with steady-state silence for the LSP layer.
The LSP layer fires on every write_file/patch. In a busy session
that's hundreds of events. We want users to be able to ``rg`` the
log for "did LSP fire on that edit?" without drowning in noise.
The level model:
- ``DEBUG`` for steady-state events that have no novel signal:
``clean``, ``feature off``, ``extension not mapped``, ``no project
root for already-announced file``, ``server unavailable for
already-announced binary``. These never reach ``agent.log`` at the
default INFO threshold.
- ``INFO`` for state transitions worth surfacing exactly once per
session: ``active for <root>`` the first time a (server_id,
workspace_root) client starts, ``no project root for <path>``
the first time we see that file. Plus every diagnostic event
(those are inherently rare and per-edit, exactly what users grep
for).
- ``WARNING`` for action-required failures: ``server unavailable``
(binary not on PATH) the first time per (server_id, binary),
``no server configured`` once per language. Per-call WARNING for
timeouts and unexpected bridge exceptions.
The dedup is in-process module-level sets. Each set grows at most by
the number of distinct (server_id, root) and (server_id, binary)
pairs touched in one Python process — bytes of memory in even an
aggressive monorepo session. Bounded LRU was rejected: evicting an
entry would risk re-firing the WARNING/INFO line we explicitly want
to suppress.
Grep recipe::
tail -f ~/.hermes/logs/agent.log | rg 'lsp\\['
"""
from __future__ import annotations
import logging
import os
import threading
from typing import Tuple
# Dedicated logger name so the documented grep recipe survives a
# ``logging.getLogger(__name__)`` rename of any internal module.
event_log = logging.getLogger("hermes.lint.lsp")
# ---------------------------------------------------------------------------
# Once-per-X dedup sets
# ---------------------------------------------------------------------------
_announce_lock = threading.Lock()
_announced_active: set = set() # keys: (server_id, workspace_root)
_announced_unavailable: set = set() # keys: (server_id, binary_path_or_name)
_announced_no_root: set = set() # keys: (server_id, file_path)
_announced_no_server: set = set() # keys: (server_id,)
def _short_path(file_path: str) -> str:
"""Render *file_path* relative to the cwd when sensible, else absolute.
Keeps log lines readable for the common case (the user is inside
the project they're editing) without emitting brittle ``../../..``
chains for the cross-tree case.
"""
if not file_path:
return file_path
try:
rel = os.path.relpath(file_path)
except ValueError:
return file_path
if rel.startswith(".." + os.sep) or rel == "..":
return file_path
return rel
def _emit(server_id: str, level: int, message: str) -> None:
event_log.log(level, "lsp[%s] %s", server_id, message)
def _announce_once(bucket: set, key: Tuple) -> bool:
"""Return True if *key* has not been announced for *bucket* yet.
Atomically marks the key as announced so concurrent callers
cannot both win the race and double-log.
"""
with _announce_lock:
if key in bucket:
return False
bucket.add(key)
return True
# ---------------------------------------------------------------------------
# Public event helpers — call these from the LSP layer.
# ---------------------------------------------------------------------------
def log_clean(server_id: str, file_path: str) -> None:
"""No diagnostics emitted for *file_path*. DEBUG (silent at default)."""
_emit(server_id, logging.DEBUG, f"clean ({_short_path(file_path)})")
def log_disabled(server_id: str, file_path: str, reason: str) -> None:
"""LSP intentionally skipped for this file (feature off, ext unmapped,
backend not local, etc.). DEBUG."""
_emit(server_id, logging.DEBUG, f"skipped: {reason} ({_short_path(file_path)})")
def log_active(server_id: str, workspace_root: str) -> None:
"""A new LSP client started for (server_id, workspace_root).
INFO once per (server_id, workspace_root); DEBUG thereafter.
Lets users verify "is LSP actually running?" with a single grep.
"""
key = (server_id, workspace_root)
if _announce_once(_announced_active, key):
_emit(server_id, logging.INFO, f"active for {workspace_root}")
else:
_emit(server_id, logging.DEBUG, f"reused client for {workspace_root}")
def log_diagnostics(server_id: str, file_path: str, count: int) -> None:
"""Diagnostics arrived for a file. INFO every time — these are the
failure signals users actually want to grep for, and they are
inherently rare per edit."""
_emit(server_id, logging.INFO, f"{count} diags ({_short_path(file_path)})")
def log_no_project_root(server_id: str, file_path: str) -> None:
"""File had no recognised project marker. INFO once per file,
DEBUG thereafter."""
key = (server_id, file_path)
if _announce_once(_announced_no_root, key):
_emit(server_id, logging.INFO, f"no project root for {_short_path(file_path)}")
else:
_emit(server_id, logging.DEBUG, f"no project root for {_short_path(file_path)}")
def log_server_unavailable(server_id: str, binary_or_pkg: str) -> None:
"""The server binary couldn't be resolved. WARNING once per
(server_id, binary), DEBUG thereafter so a hundred subsequent
.py edits don't spam the log."""
key = (server_id, binary_or_pkg)
if _announce_once(_announced_unavailable, key):
_emit(
server_id,
logging.WARNING,
f"server unavailable: {binary_or_pkg} not found "
"(install via `hermes lsp install <id>` or set lsp.servers.<id>.command)",
)
else:
_emit(server_id, logging.DEBUG, f"server still unavailable: {binary_or_pkg}")
def log_no_server_configured(server_id: str) -> None:
"""No spawn recipe for this language. WARNING once."""
if _announce_once(_announced_no_server, (server_id,)):
_emit(server_id, logging.WARNING, "no server configured")
def log_timeout(server_id: str, file_path: str, kind: str = "diagnostics") -> None:
"""A request to the server timed out. WARNING every time — these are
inherently novel events worth surfacing on each occurrence."""
_emit(
server_id,
logging.WARNING,
f"{kind} timed out for {_short_path(file_path)}",
)
def log_server_error(server_id: str, file_path: str, exc: BaseException) -> None:
"""An unexpected exception bubbled out of the LSP layer. WARNING."""
_emit(
server_id,
logging.WARNING,
f"unexpected error for {_short_path(file_path)}: {type(exc).__name__}: {exc}",
)
def log_spawn_failed(server_id: str, workspace_root: str, exc: BaseException) -> None:
"""The LSP server failed to spawn or initialize. WARNING."""
_emit(
server_id,
logging.WARNING,
f"spawn/initialize failed for {workspace_root}: {type(exc).__name__}: {exc}",
)
def reset_announce_caches() -> None:
"""Test-only: clear the dedup caches. Production code never calls this."""
with _announce_lock:
_announced_active.clear()
_announced_unavailable.clear()
_announced_no_root.clear()
_announced_no_server.clear()
__all__ = [
"event_log",
"log_clean",
"log_disabled",
"log_active",
"log_diagnostics",
"log_no_project_root",
"log_server_unavailable",
"log_no_server_configured",
"log_timeout",
"log_server_error",
"log_spawn_failed",
"reset_announce_caches",
]