drydock-cli 3.0.92__tar.gz → 3.0.94__tar.gz

{drydock_cli-3.0.92/drydock_cli.egg-info → drydock_cli-3.0.94}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: drydock-cli
-Version: 3.0.92
+Version: 3.0.94
 Summary: Drydock — a local, provider-agnostic terminal coding agent for local LLMs
 Author: Frank Bobe III
 License-Expression: Apache-2.0

{drydock_cli-3.0.92 → drydock_cli-3.0.94}/drydock/tools/__init__.py

@@ -46,6 +46,20 @@ def _detect_bash() -> str | None:
 _BASH_SHELL = _detect_bash()
 
 
+# ANSI escape sequences (CSI colour/cursor, OSC title, and lone two-char escapes).
+# Some tools emit these even to a pipe (--color=always, forced-colour test runners),
+# and they're pure display control — noise that wastes the model's tokens.
+_ANSI_RE = re.compile(r"\x1b(?:\[[0-9;?]*[ -/]*[@-~]|\][^\x07\x1b]*(?:\x07|\x1b\\)|[@-Z\\-_])")
+
+
+def _sanitize_bash_output(text: str) -> str:
+    """Strip ANSI escape sequences and drop NUL bytes from command output. NUL is
+    valid UTF-8 (so the errors='replace' decode keeps it) but raw NUL in a JSON
+    tool-result trips some LLM servers' tokenizers; ANSI codes are display noise.
+    Neither carries information the model needs. Newlines/tabs/Unicode are kept."""
+    return _ANSI_RE.sub("", text).replace("\x00", "")
+
+
 def _collapse_repeated_lines(text: str, run: int = 20) -> str:
     """Collapse a run of >= `run` IDENTICAL consecutive lines into one line + a
     count. Repetitive output (`yes`, a spinning progress log) tokenizes densely —
@@ -942,7 +956,13 @@ def tool_bash(params: dict, config: dict) -> str:
             # explicitly via a pipe/redirect (echo x | cmd, cmd < file), which
             # overrides this.
             stdin=subprocess.DEVNULL,
-            text=True, cwd=config.get("cwd"), start_new_session=True,
+            # errors="replace" so BINARY / non-UTF8 output (cat a binary, a tool
+            # emitting raw bytes, gzip to stdout) decodes to replacement chars
+            # instead of raising UnicodeDecodeError inside the reader thread —
+            # which killed the thread and handed the agent "(no output)", losing
+            # even the text parts of mixed output.
+            text=True, encoding="utf-8", errors="replace",
+            cwd=config.get("cwd"), start_new_session=True,
         )
         config.setdefault("_abort", {})["proc"] = proc
         # Read output in a daemon thread with a HARD byte cap, so memory can't
@@ -1007,7 +1027,7 @@ def tool_bash(params: dict, config: dict) -> str:
         # then note if we hit the byte cap. Bounds both RAM (the cap) and context
         # tokens (the collapse). Snapshot chunks (list()) in case the reader
         # daemon is still appending for a backgrounded child.
-        output = _collapse_repeated_lines("".join(list(chunks)))
+        output = _sanitize_bash_output(_collapse_repeated_lines("".join(list(chunks))))
         if backgrounded:
             return (output.rstrip() + "\n[a process was left running in the background; "
                     "the command returned. Check it with a follow-up command.]").lstrip("\n")

{drydock_cli-3.0.92 → drydock_cli-3.0.94/drydock_cli.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: drydock-cli
-Version: 3.0.92
+Version: 3.0.94
 Summary: Drydock — a local, provider-agnostic terminal coding agent for local LLMs
 Author: Frank Bobe III
 License-Expression: Apache-2.0

{drydock_cli-3.0.92 → drydock_cli-3.0.94}/drydock_cli.egg-info/SOURCES.txt

@@ -50,9 +50,11 @@ tests/test_advisor.py
 tests/test_approval.py
 tests/test_back_command.py
 tests/test_bash_background.py
+tests/test_bash_binary_output.py
 tests/test_bash_output_bounding.py
 tests/test_bash_process_group.py
 tests/test_bash_safety.py
+tests/test_bash_sanitize.py
 tests/test_bash_shell.py
 tests/test_bash_stdin.py
 tests/test_bash_timeout_network.py

{drydock_cli-3.0.92 → drydock_cli-3.0.94}/pyproject.toml

@@ -7,7 +7,7 @@ build-backend = "setuptools.build_meta"
 # PyPI distribution name is drydock-cli (the established install name, continued
 # from the retired v2 fork); the import package + CLI command stay `drydock`.
 name = "drydock-cli"
-version = "3.0.92"
+version = "3.0.94"
 description = "Drydock — a local, provider-agnostic terminal coding agent for local LLMs"
 readme = "README.md"
 requires-python = ">=3.11"

drydock_cli-3.0.94/tests/test_bash_binary_output.py

@@ -0,0 +1,33 @@
+"""tool_bash must not choke on binary / non-UTF8 stdout. text=True with the
+default (strict) decode raised UnicodeDecodeError INSIDE the reader thread — the
+thread died and the agent got "(no output)", losing even the text parts of mixed
+output. errors="replace" keeps it non-crashing and preserves the text. Found by
+probing the binary-output I/O path directly."""
+from __future__ import annotations
+
+from drydock.tools import tool_bash
+
+
+def _run(cmd):
+    return tool_bash({"command": cmd, "timeout": 8}, {"cwd": "/tmp"})
+
+
+def test_invalid_utf8_does_not_crash_and_keeps_text():
+    out = _run(r"printf '\xff\xfe hello'")
+    assert "hello" in out and "(no output)" not in out
+
+
+def test_mixed_text_and_binary_preserves_text():
+    # the readable lines must survive even though raw bytes sit between them
+    out = _run(r"echo START; printf '\xc0\xc1\xff'; echo END")
+    assert "START" in out and "END" in out
+
+
+def test_binary_file_head_returns_something():
+    out = _run("head -c 80 /bin/ls")
+    assert out.strip() and "(no output)" not in out and "EXCEPTION" not in out
+
+
+def test_plain_text_unaffected():
+    out = _run("echo hello && echo world")
+    assert "hello" in out and "world" in out

drydock_cli-3.0.94/tests/test_bash_sanitize.py

@@ -0,0 +1,35 @@
+"""Bash output is sanitized before the model sees it: ANSI escape sequences
+(colour/cursor control some tools emit even to a pipe) are stripped, and NUL
+bytes (which trip some LLM servers' JSON handling) are dropped. Text/tabs/
+newlines/Unicode are preserved. Found by probing the output I/O path."""
+from __future__ import annotations
+
+import pytest
+
+from drydock.tools import tool_bash, _sanitize_bash_output
+
+
+@pytest.mark.parametrize("raw,clean", [
+    ("\x1b[31mRED\x1b[0m normal", "RED normal"),   # SGR colour
+    ("\x1b[2K\x1b[1Gline", "line"),                 # cursor control
+    ("\x1b]0;window title\x07keep", "keep"),        # OSC title
+    ("a\x00b\x00c", "abc"),                          # NUL bytes
+    ("plain\ttext\nline2 🚢⚓", "plain\ttext\nline2 🚢⚓"),  # preserved
+])
+def test_sanitizer(raw, clean):
+    assert _sanitize_bash_output(raw) == clean
+
+
+def test_end_to_end_forced_color_stripped():
+    out = tool_bash({"command": r"printf '\x1b[31mRED\x1b[0m done\n'", "timeout": 8}, {"cwd": "/tmp"})
+    assert "RED done" in out and "\x1b" not in out
+
+
+def test_end_to_end_nul_dropped():
+    out = tool_bash({"command": r"printf 'a\x00b\x00c'", "timeout": 8}, {"cwd": "/tmp"})
+    assert "\x00" not in out and "abc" in out
+
+
+def test_valid_unicode_survives():
+    out = tool_bash({"command": r"printf '\xf0\x9f\x9a\xa2 ship\n'", "timeout": 8}, {"cwd": "/tmp"})
+    assert "🚢" in out