context-loader 1.3.4__tar.gz → 1.3.6__tar.gz

{context_loader-1.3.4 → context_loader-1.3.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: context-loader
-Version: 1.3.4
+Version: 1.3.6
 Summary: Render deterministic local Git context for any executor
 License:                                  Apache License
                                     Version 2.0, January 2004
@@ -219,14 +219,14 @@ Python standard library.
 
 ## Open-source quick start
 
-Current stable release: **1.3.4**. Local CLI for coding-agent workflows; renders deterministic Markdown or JSON.
+Current stable release: **1.3.6**. Local CLI for coding-agent workflows; renders deterministic Markdown or JSON.
 
 ```bash
-uv tool install 'context-loader==1.3.4'
+uv tool install 'context-loader==1.3.6'
 ```
 
 ```bash
-pip install 'context-loader==1.3.4'
+pip install 'context-loader==1.3.6'
 ```
 
 ```bash
@@ -239,21 +239,21 @@ project-context --repo /path/to/repo --format json
 
 ## Install
 
-Source version: `1.3.4`. Install its matching published release or an exact source commit.
+Source version: `1.3.6`. Install its matching published release or an exact source commit.
 
 Install via `uv`:
 
 ```bash
-uv tool install 'context-loader==1.3.4'
+uv tool install 'context-loader==1.3.6'
 ```
 
 Install via `pip`:
 
 ```bash
-pip install 'context-loader==1.3.4'
+pip install 'context-loader==1.3.6'
 ```
 
-For development or source-based installs tracking current repository (`1.3.4`):
+For development or source-based installs tracking current repository (`1.3.6`):
 
 ```bash
 uv tool install git+https://github.com/xuanheng-tech/context-loader.git
@@ -350,7 +350,7 @@ set differs, as described under “Compact model consumption”. The declared co
   "schema_version": 3,
   "tool": {
     "name": "context-loader",
-    "version": "1.3.4"
+    "version": "1.3.6"
   },
   "repository": {
     "requested_path": "/canonical/requested/path",
@@ -470,7 +470,7 @@ although 1.1.0 added the top-level `statuses` key. From release 1.2.0 onward eac
 names exactly one key set, and a new number is issued whenever a key set would otherwise be reused.
 
 The JSON schema version and package version are independent: this build emits `schema_version` `3`
-for `--format json` and `4` for `--format json-compact`, while `tool.version` is `1.3.4`. Version 1
+for `--format json` and `4` for `--format json-compact`, while `tool.version` is `1.3.6`. Version 1
 and 2 are the exact shapes already published in release 1.2.0 and are never reused: because
 `nested_context` changes the default document's key set, version 3 names the current full-document
 key set and version 4 names the compact projection.
@@ -512,12 +512,20 @@ heading on the first line is still recognized as a heading.
 
 The root `AGENTS.md` is split at Markdown headings outside fenced code blocks. The output always
 starts with complete early sections fitting a 4-KiB head, then adds complete relevant sections in
-source order using exact normalized focus/path tokens and their parent context. Remaining headings
+source order using normalized focus/path tokens and their parent context. Remaining headings
 appear in an explicit index whose body text is not loaded. With no selection signals, only the small
 head and index are emitted. Unsafe heading parsing falls back to a bounded head and an explicit
 manual-recovery notice. A head whose own headings would not fit the budget falls back to the same
 bounded head without a section index; a single instruction file never fails the whole collection.
 
+Matching uses NFKC normalization and case folding. Ordinary tokens require at least three characters;
+two-character Han terms such as `安全` and `部署`, and the technical abbreviations `AI`, `CD`, `CI`,
+`DB`, `IO`, `JS`, `ML`, `OS`, `PR`, `QA`, `TS`, `UI` and `UX`, are also eligible. Han terms can match
+within entirely Han words, so `安全` matches `安全规则`; Latin terms still require complete tokens,
+so `CI` does not match `special`. Existing stopwords, scoring and byte budgets still apply. Repeating
+a Han term in a heading and body does not increase its score. This is literal matching without word
+segmentation, synonyms or translation.
+
 `Omitted source characters` is measured against the whole normalized source, so a bounded source
 scan that ends before EOF still reports the characters it could not select.
 
@@ -596,6 +604,8 @@ bullet is a per-component, read or capture bound, as each one states:
   repository root the filesystem could not decode makes the machine formats exit 1 where Markdown
   still renders
 - Git subprocess output: 16 MiB (bounded while reading)
+- Git collection: one shared 30-second deadline for repository discovery, configuration/submodule
+  preflight, refs, status and recent commits
 - Git status preflight: at most 128 repositories, including initialized nested submodules
 - Candidate file validation: 16 MiB hard read bound
 
@@ -611,6 +621,12 @@ Symlinked candidate files are skipped, and directory symlinks are not traversed.
 Git state is derived only from the working tree and local refs. The command does not fetch or query
 the remote.
 
+All Git commands in one collection share the same elapsed-time budget; each command receives only
+the remaining time. Once that budget expires, the loader stops without starting subsequent commands,
+terminates and reaps a running Git child, and exits with empty stdout. This deadline does not impose
+a hard wall-clock limit on blocked filesystem/process-creation calls or operating-system child
+reaping, and does not cover the later context-file collection.
+
 - An unborn working tree reports its symbolic branch name, such as `Branch: main`, and
   `HEAD: unborn`.
 - A detached HEAD reports `Branch: detached` and the resolved commit object ID.
@@ -639,7 +655,7 @@ or candidate-file content.
 
 ## Not Included
 
-Version 1.3.4 does not provide AI summaries, project-type detection, loading or transport of
+Version 1.3.6 does not provide AI summaries, project-type detection, loading or transport of
 nested `AGENTS.md` contents (the JSON `nested_context` field reports bounded existence only),
 Memory retrieval, semantic ranking, ignore-rule parsing, plugins, profiles, caches, databases,
 network services, MCP, daemons, GUIs, CI/CD, telemetry, or automatic updates.
@@ -654,10 +670,12 @@ just check
 
 The shared check also builds a wheel and installs it into a fresh virtual environment. Package
 acceptance runs the installed CLI outside the source checkout, checking deterministic agreement
-between all three formats, linked worktrees, and safe failures. Build and installation run offline
+between all three formats, linked worktrees, long multi-module instructions with focus/path selection,
+explicit truncation and nested-instruction presence, and safe failures. Build and installation run offline
 using the backend cache populated by `uv run --frozen`; the tests use only disposable repositories.
 Git deadline regressions verify that silent, streaming, and stdout-closed child processes are
-terminated and reaped, and that a timeout leaves CLI stdout empty.
+terminated and reaped, that successful commands cannot reset the shared budget, and that process
+creation consumes it. A timeout leaves CLI stdout empty in every output format.
 
 The lockfile uses public PyPI artifact URLs; no organization mirror or local Renovate preset
 is needed. Renovate keeps scheduled, rate-limited updates and manual merges using a public

{context_loader-1.3.4 → context_loader-1.3.6}/README.md

@@ -8,14 +8,14 @@ Python standard library.
 
 ## Open-source quick start
 
-Current stable release: **1.3.4**. Local CLI for coding-agent workflows; renders deterministic Markdown or JSON.
+Current stable release: **1.3.6**. Local CLI for coding-agent workflows; renders deterministic Markdown or JSON.
 
 ```bash
-uv tool install 'context-loader==1.3.4'
+uv tool install 'context-loader==1.3.6'
 ```
 
 ```bash
-pip install 'context-loader==1.3.4'
+pip install 'context-loader==1.3.6'
 ```
 
 ```bash
@@ -28,21 +28,21 @@ project-context --repo /path/to/repo --format json
 
 ## Install
 
-Source version: `1.3.4`. Install its matching published release or an exact source commit.
+Source version: `1.3.6`. Install its matching published release or an exact source commit.
 
 Install via `uv`:
 
 ```bash
-uv tool install 'context-loader==1.3.4'
+uv tool install 'context-loader==1.3.6'
 ```
 
 Install via `pip`:
 
 ```bash
-pip install 'context-loader==1.3.4'
+pip install 'context-loader==1.3.6'
 ```
 
-For development or source-based installs tracking current repository (`1.3.4`):
+For development or source-based installs tracking current repository (`1.3.6`):
 
 ```bash
 uv tool install git+https://github.com/xuanheng-tech/context-loader.git
@@ -139,7 +139,7 @@ set differs, as described under “Compact model consumption”. The declared co
   "schema_version": 3,
   "tool": {
     "name": "context-loader",
-    "version": "1.3.4"
+    "version": "1.3.6"
   },
   "repository": {
     "requested_path": "/canonical/requested/path",
@@ -259,7 +259,7 @@ although 1.1.0 added the top-level `statuses` key. From release 1.2.0 onward eac
 names exactly one key set, and a new number is issued whenever a key set would otherwise be reused.
 
 The JSON schema version and package version are independent: this build emits `schema_version` `3`
-for `--format json` and `4` for `--format json-compact`, while `tool.version` is `1.3.4`. Version 1
+for `--format json` and `4` for `--format json-compact`, while `tool.version` is `1.3.6`. Version 1
 and 2 are the exact shapes already published in release 1.2.0 and are never reused: because
 `nested_context` changes the default document's key set, version 3 names the current full-document
 key set and version 4 names the compact projection.
@@ -301,12 +301,20 @@ heading on the first line is still recognized as a heading.
 
 The root `AGENTS.md` is split at Markdown headings outside fenced code blocks. The output always
 starts with complete early sections fitting a 4-KiB head, then adds complete relevant sections in
-source order using exact normalized focus/path tokens and their parent context. Remaining headings
+source order using normalized focus/path tokens and their parent context. Remaining headings
 appear in an explicit index whose body text is not loaded. With no selection signals, only the small
 head and index are emitted. Unsafe heading parsing falls back to a bounded head and an explicit
 manual-recovery notice. A head whose own headings would not fit the budget falls back to the same
 bounded head without a section index; a single instruction file never fails the whole collection.
 
+Matching uses NFKC normalization and case folding. Ordinary tokens require at least three characters;
+two-character Han terms such as `安全` and `部署`, and the technical abbreviations `AI`, `CD`, `CI`,
+`DB`, `IO`, `JS`, `ML`, `OS`, `PR`, `QA`, `TS`, `UI` and `UX`, are also eligible. Han terms can match
+within entirely Han words, so `安全` matches `安全规则`; Latin terms still require complete tokens,
+so `CI` does not match `special`. Existing stopwords, scoring and byte budgets still apply. Repeating
+a Han term in a heading and body does not increase its score. This is literal matching without word
+segmentation, synonyms or translation.
+
 `Omitted source characters` is measured against the whole normalized source, so a bounded source
 scan that ends before EOF still reports the characters it could not select.
 
@@ -385,6 +393,8 @@ bullet is a per-component, read or capture bound, as each one states:
   repository root the filesystem could not decode makes the machine formats exit 1 where Markdown
   still renders
 - Git subprocess output: 16 MiB (bounded while reading)
+- Git collection: one shared 30-second deadline for repository discovery, configuration/submodule
+  preflight, refs, status and recent commits
 - Git status preflight: at most 128 repositories, including initialized nested submodules
 - Candidate file validation: 16 MiB hard read bound
 
@@ -400,6 +410,12 @@ Symlinked candidate files are skipped, and directory symlinks are not traversed.
 Git state is derived only from the working tree and local refs. The command does not fetch or query
 the remote.
 
+All Git commands in one collection share the same elapsed-time budget; each command receives only
+the remaining time. Once that budget expires, the loader stops without starting subsequent commands,
+terminates and reaps a running Git child, and exits with empty stdout. This deadline does not impose
+a hard wall-clock limit on blocked filesystem/process-creation calls or operating-system child
+reaping, and does not cover the later context-file collection.
+
 - An unborn working tree reports its symbolic branch name, such as `Branch: main`, and
   `HEAD: unborn`.
 - A detached HEAD reports `Branch: detached` and the resolved commit object ID.
@@ -428,7 +444,7 @@ or candidate-file content.
 
 ## Not Included
 
-Version 1.3.4 does not provide AI summaries, project-type detection, loading or transport of
+Version 1.3.6 does not provide AI summaries, project-type detection, loading or transport of
 nested `AGENTS.md` contents (the JSON `nested_context` field reports bounded existence only),
 Memory retrieval, semantic ranking, ignore-rule parsing, plugins, profiles, caches, databases,
 network services, MCP, daemons, GUIs, CI/CD, telemetry, or automatic updates.
@@ -443,10 +459,12 @@ just check
 
 The shared check also builds a wheel and installs it into a fresh virtual environment. Package
 acceptance runs the installed CLI outside the source checkout, checking deterministic agreement
-between all three formats, linked worktrees, and safe failures. Build and installation run offline
+between all three formats, linked worktrees, long multi-module instructions with focus/path selection,
+explicit truncation and nested-instruction presence, and safe failures. Build and installation run offline
 using the backend cache populated by `uv run --frozen`; the tests use only disposable repositories.
 Git deadline regressions verify that silent, streaming, and stdout-closed child processes are
-terminated and reaped, and that a timeout leaves CLI stdout empty.
+terminated and reaped, that successful commands cannot reset the shared budget, and that process
+creation consumes it. A timeout leaves CLI stdout empty in every output format.
 
 The lockfile uses public PyPI artifact URLs; no organization mirror or local Renovate preset
 is needed. Renovate keeps scheduled, rate-limited updates and manual merges using a public

{context_loader-1.3.4 → context_loader-1.3.6}/context_loader/__init__.py

@@ -1,3 +1,3 @@
 """Project Context Loader."""
 
-__version__ = "1.3.4"
+__version__ = "1.3.6"

{context_loader-1.3.4 → context_loader-1.3.6}/context_loader/collect.py

@@ -68,6 +68,9 @@ _MAKE_TARGET_RE = re.compile(
 )
 _ATX_HEADING_RE = re.compile(r"^ {0,3}(#{1,6})(?:[ \t]+(.*)|[ \t]*)$")
 _TOKEN_RE = re.compile(r"[^\W_]+", re.UNICODE)
+_SHORT_TECHNICAL_TOKENS = frozenset(
+    {"ai", "cd", "ci", "db", "io", "js", "ml", "os", "pr", "qa", "ts", "ui", "ux"}
+)
 _MATCH_TOKEN_STOPWORDS = frozenset(
     {
         "and",
@@ -138,11 +141,35 @@ def _normalized_text(value: str) -> str:
     return unicodedata.normalize("NFKC", value).casefold()
 
 
+def _is_han_token(value: str) -> bool:
+    return bool(value) and all(
+        unicodedata.name(character, "").startswith("CJK UNIFIED IDEOGRAPH-") for character in value
+    )
+
+
 def _matching_tokens(value: str, stopwords: frozenset[str]) -> frozenset[str]:
     return frozenset(
         token
         for token in _TOKEN_RE.findall(_normalized_text(value))
-        if len(token) >= 3 and token not in stopwords
+        if token not in stopwords
+        and (
+            len(token) >= 3
+            or token in _SHORT_TECHNICAL_TOKENS
+            or (len(token) == 2 and _is_han_token(token))
+        )
+    )
+
+
+def _matched_terms(terms: frozenset[str], tokens: frozenset[str]) -> frozenset[str]:
+    matches = terms & tokens
+    han_terms = tuple(term for term in terms - matches if _is_han_token(term))
+    if not han_terms:
+        return matches
+    # Han text need not delimit words with spaces. Count each query term once,
+    # without introducing substring matching for Latin words or abbreviations.
+    han_tokens = tuple(token for token in tokens if _is_han_token(token))
+    return matches | frozenset(
+        term for term in han_terms if any(term in token for token in han_tokens)
     )
 
 
@@ -334,24 +361,20 @@ def _small_head_end(content: str, sections: tuple[MarkdownSection, ...]) -> int:
 def _section_relevance(
     section: MarkdownSection, signals: _SelectionSignals
 ) -> tuple[int, frozenset[str]]:
-    heading_tokens = frozenset(
-        token
-        for token in section.normalized_heading.split()
-        if len(token) >= 3 and token not in _MATCH_TOKEN_STOPWORDS
-    )
+    heading_tokens = _matching_tokens(section.normalized_heading, _MATCH_TOKEN_STOPWORDS)
     section_tokens = _matching_tokens(section.text, _MATCH_TOKEN_STOPWORDS)
     body_tokens = section_tokens - heading_tokens
     score = 0
     reasons: set[str] = set()
 
-    focus_heading = signals.focus_tokens & heading_tokens
-    focus_body = signals.focus_tokens & body_tokens
+    focus_heading = _matched_terms(signals.focus_tokens, heading_tokens)
+    focus_body = _matched_terms(signals.focus_tokens, body_tokens) - focus_heading
     if focus_heading or focus_body:
         reasons.add("focus_match")
         score += 6 * len(focus_heading) + len(focus_body)
 
-    path_heading = signals.path_tokens & heading_tokens
-    path_body = signals.path_tokens & body_tokens
+    path_heading = _matched_terms(signals.path_tokens, heading_tokens)
+    path_body = _matched_terms(signals.path_tokens, body_tokens) - path_heading
     exact_path = bool(
         signals.normalized_path
         and _normalized_text(signals.normalized_path) in _normalized_text(section.text)

{context_loader-1.3.4 → context_loader-1.3.6}/context_loader/git.py

@@ -118,8 +118,11 @@ def _read_bounded_process_stdout(
     proc: subprocess.Popen[bytes],
     max_bytes: int,
     timeout_seconds: float,
+    *,
+    deadline: float | None = None,
 ) -> bytes:
-    deadline = time.monotonic() + timeout_seconds
+    if deadline is None:
+        deadline = time.monotonic() + timeout_seconds
     buffer = bytearray()
     assert proc.stdout is not None
     os.set_blocking(proc.stdout.fileno(), False)
@@ -168,9 +171,23 @@ def _read_bounded_process_stdout(
     return bytes(buffer)
 
 
-def _run_git(repo: Path, arguments: tuple[str, ...], *, check: bool = True) -> GitResult:
+def _check_git_deadline(deadline: float) -> None:
+    if time.monotonic() >= deadline:
+        raise ContextLoaderError("unable to read repository Git state")
+
+
+def _run_git(
+    repo: Path,
+    arguments: tuple[str, ...],
+    *,
+    check: bool = True,
+    deadline: float | None = None,
+) -> GitResult:
+    command_deadline = time.monotonic() + GIT_TIMEOUT_SECONDS
+    deadline = command_deadline if deadline is None else min(deadline, command_deadline)
     argv = [GIT_EXECUTABLE, *_GIT_PREFIX, "-C", os.fspath(repo), *arguments]
     try:
+        _check_git_deadline(deadline)
         with subprocess.Popen(  # noqa: S603 - executable and every command are fixed here.
             argv,
             cwd="/",
@@ -179,8 +196,11 @@ def _run_git(repo: Path, arguments: tuple[str, ...], *, check: bool = True) -> G
             stdout=subprocess.PIPE,
             stderr=subprocess.DEVNULL,
         ) as proc:
-            stdout = _read_bounded_process_stdout(proc, MAX_GIT_OUTPUT_BYTES, GIT_TIMEOUT_SECONDS)
+            stdout = _read_bounded_process_stdout(
+                proc, MAX_GIT_OUTPUT_BYTES, GIT_TIMEOUT_SECONDS, deadline=deadline
+            )
             returncode = proc.returncode
+        _check_git_deadline(deadline)
     except ContextLoaderError:
         raise
     except (OSError, subprocess.TimeoutExpired) as exc:
@@ -209,6 +229,16 @@ def discover_repository(
     raw_path: str | os.PathLike[str], *, require_canonical_root: bool = False
 ) -> RepositoryLocation:
     """Resolve an absolute directory and discover its non-bare Git worktree root."""
+    return _discover_repository(
+        raw_path,
+        require_canonical_root=require_canonical_root,
+        deadline=time.monotonic() + GIT_TIMEOUT_SECONDS,
+    )
+
+
+def _discover_repository(
+    raw_path: str | os.PathLike[str], *, require_canonical_root: bool, deadline: float
+) -> RepositoryLocation:
     candidate = Path(raw_path)
     if not candidate.is_absolute():
         raise ContextLoaderError("--repo must be an absolute path", exit_code=2)
@@ -219,11 +249,14 @@ def discover_repository(
     if not canonical.is_dir():
         raise ContextLoaderError("--repo must name an existing directory", exit_code=2)
 
-    bare = _run_git(canonical, ("rev-parse", "--is-bare-repository"), check=False)
+    bare = _run_git(
+        canonical, ("rev-parse", "--is-bare-repository"), check=False, deadline=deadline
+    )
     root = _run_git(
         canonical,
         ("rev-parse", "--path-format=absolute", "--show-toplevel"),
         check=False,
+        deadline=deadline,
     )
     if bare.returncode != 0 or root.returncode != 0:
         raise ContextLoaderError(
@@ -243,6 +276,7 @@ def discover_repository(
         raise ContextLoaderError(
             "--repo must be the canonical root of a non-bare Git worktree", exit_code=2
         )
+    _check_git_deadline(deadline)
     return RepositoryLocation(requested_path=canonical, canonical_root=discovered)
 
 
@@ -251,14 +285,18 @@ def validate_repository(raw_path: str | os.PathLike[str]) -> Path:
     return discover_repository(raw_path, require_canonical_root=True).canonical_root
 
 
-def _head_and_branch(repo: Path) -> tuple[str, str, str | None]:
-    symbolic = _run_git(repo, ("symbolic-ref", "--quiet", "--short", "HEAD"), check=False)
+def _head_and_branch(repo: Path, deadline: float) -> tuple[str, str, str | None]:
+    symbolic = _run_git(
+        repo, ("symbolic-ref", "--quiet", "--short", "HEAD"), check=False, deadline=deadline
+    )
     branch_name = (
         _decode_path(_without_one_line_ending(symbolic.stdout))
         if symbolic.returncode == 0
         else None
     )
-    resolved_head = _run_git(repo, ("rev-parse", "--verify", "HEAD^{commit}"), check=False)
+    resolved_head = _run_git(
+        repo, ("rev-parse", "--verify", "HEAD^{commit}"), check=False, deadline=deadline
+    )
     if resolved_head.returncode != 0:
         if branch_name is None:
             raise ContextLoaderError("unable to read repository Git state")
@@ -270,17 +308,17 @@ def _head_and_branch(repo: Path) -> tuple[str, str, str | None]:
     return branch_name or "detached", head, branch_name
 
 
-def _config_value(repo: Path, key: str) -> str | None:
-    result = _run_git(repo, ("config", "--get", key), check=False)
+def _config_value(repo: Path, key: str, deadline: float) -> str | None:
+    result = _run_git(repo, ("config", "--get", key), check=False, deadline=deadline)
     if result.returncode != 0:
         return None
     value = _decode_text(_without_one_line_ending(result.stdout))
     return value or None
 
 
-def _fallback_upstream(repo: Path, branch_name: str) -> str | None:
-    remote = _config_value(repo, f"branch.{branch_name}.remote")
-    merge = _config_value(repo, f"branch.{branch_name}.merge")
+def _fallback_upstream(repo: Path, branch_name: str, deadline: float) -> str | None:
+    remote = _config_value(repo, f"branch.{branch_name}.remote", deadline)
+    merge = _config_value(repo, f"branch.{branch_name}.merge", deadline)
     if remote is None or merge is None:
         return None
     merge_name = merge.removeprefix("refs/heads/")
@@ -289,7 +327,7 @@ def _fallback_upstream(repo: Path, branch_name: str) -> str | None:
     return f"{remote}/{merge_name}"
 
 
-def _upstream_state(repo: Path, branch_name: str | None) -> tuple[str, str]:
+def _upstream_state(repo: Path, branch_name: str | None, deadline: float) -> tuple[str, str]:
     if branch_name is None:
         return "not configured", "not available"
 
@@ -297,9 +335,10 @@ def _upstream_state(repo: Path, branch_name: str | None) -> tuple[str, str]:
         repo,
         ("rev-parse", "--abbrev-ref", "--symbolic-full-name", "@{upstream}"),
         check=False,
+        deadline=deadline,
     )
     if resolved.returncode != 0:
-        configured = _fallback_upstream(repo, branch_name)
+        configured = _fallback_upstream(repo, branch_name, deadline)
         return configured or "not configured", "not available"
 
     upstream = _decode_text(_without_one_line_ending(resolved.stdout))
@@ -309,6 +348,7 @@ def _upstream_state(repo: Path, branch_name: str | None) -> tuple[str, str]:
         repo,
         ("rev-list", "--left-right", "--count", "HEAD...@{upstream}"),
         check=False,
+        deadline=deadline,
     )
     if divergence.returncode != 0:
         return upstream, "not available"
@@ -318,7 +358,7 @@ def _upstream_state(repo: Path, branch_name: str | None) -> tuple[str, str]:
     return upstream, f"{counts[0].decode('ascii')} / {counts[1].decode('ascii')}"
 
 
-def _ensure_safe_status(repo: Path) -> None:
+def _ensure_safe_status(repo: Path, deadline: float) -> None:
     """Reject executable conversions before status can refresh any index entry.
 
     Attributes can select clean/process drivers even during a read-only status.
@@ -340,6 +380,7 @@ def _ensure_safe_status(repo: Path) -> None:
             current,
             ("config", "--null", "--get-regexp", r"^filter\..*\.(clean|process)$"),
             check=False,
+            deadline=deadline,
         )
         if filters.returncode not in (0, 1):
             raise ContextLoaderError("unable to safely inspect repository Git filters")
@@ -356,7 +397,7 @@ def _ensure_safe_status(repo: Path) -> None:
                 "repository or submodule config enables Git clean/process filters; "
                 "unable to safely inspect working-tree state"
             )
-        index = _run_git(current, ("ls-files", "--stage", "-z"))
+        index = _run_git(current, ("ls-files", "--stage", "-z"), deadline=deadline)
         children: set[Path] = set()
         for record in index.stdout.split(b"\0"):
             metadata, separator, raw_path = record.partition(b"\t")
@@ -375,8 +416,8 @@ def _ensure_safe_status(repo: Path) -> None:
             raise ContextLoaderError("repository submodule inspection exceeded the safety limit")
 
 
-def _working_tree_changes(repo: Path) -> tuple[WorkingTreeChange, ...]:
-    _ensure_safe_status(repo)
+def _working_tree_changes(repo: Path, deadline: float) -> tuple[WorkingTreeChange, ...]:
+    _ensure_safe_status(repo, deadline)
     result = _run_git(
         repo,
         (
@@ -387,6 +428,7 @@ def _working_tree_changes(repo: Path) -> tuple[WorkingTreeChange, ...]:
             "--no-renames",
             "--ignore-submodules=none",
         ),
+        deadline=deadline,
     )
     if not result.stdout:
         return ()
@@ -408,7 +450,7 @@ def _working_tree_changes(repo: Path) -> tuple[WorkingTreeChange, ...]:
     return tuple(sorted(changes, key=lambda change: change.sort_key))
 
 
-def _recent_commits(repo: Path, head: str) -> tuple[RecentCommit, ...]:
+def _recent_commits(repo: Path, head: str, deadline: float) -> tuple[RecentCommit, ...]:
     if head == "unborn":
         return ()
     result = _run_git(
@@ -421,6 +463,7 @@ def _recent_commits(repo: Path, head: str) -> tuple[RecentCommit, ...]:
             "HEAD",
             "--",
         ),
+        deadline=deadline,
     )
     commits: list[RecentCommit] = []
     for line in result.stdout.splitlines():
@@ -443,13 +486,16 @@ def _recent_commits(repo: Path, head: str) -> tuple[RecentCommit, ...]:
 def collect_repository(
     raw_path: str | os.PathLike[str], *, require_canonical_root: bool = True
 ) -> tuple[RepositoryLocation, RepositoryState]:
-    """Discover one repository and collect only the Phase 01 Git fields."""
-    location = discover_repository(raw_path, require_canonical_root=require_canonical_root)
+    """Discover and collect local Git state within one shared elapsed-time budget."""
+    deadline = time.monotonic() + GIT_TIMEOUT_SECONDS
+    location = _discover_repository(
+        raw_path, require_canonical_root=require_canonical_root, deadline=deadline
+    )
     repository = location.canonical_root
-    branch, head, branch_name = _head_and_branch(repository)
-    upstream, ahead_behind = _upstream_state(repository, branch_name)
-    changes = _working_tree_changes(repository)
-    commits = _recent_commits(repository, head)
+    branch, head, branch_name = _head_and_branch(repository, deadline)
+    upstream, ahead_behind = _upstream_state(repository, branch_name, deadline)
+    changes = _working_tree_changes(repository, deadline)
+    commits = _recent_commits(repository, head, deadline)
     state = RepositoryState(
         repository=repository,
         branch=branch,
@@ -459,6 +505,7 @@ def collect_repository(
         changes=changes,
         commits=commits,
     )
+    _check_git_deadline(deadline)
     return location, state
 
 

{context_loader-1.3.4 → context_loader-1.3.6}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "uv_build"
 
 [project]
 name = "context-loader"
-version = "1.3.4"
+version = "1.3.6"
 description = "Render deterministic local Git context for any executor"
 readme = "README.md"
 requires-python = ">=3.12,<3.15"

{context_loader-1.3.4 → context_loader-1.3.6}/pyproject.toml.orig

@@ -4,7 +4,7 @@ build-backend = "uv_build"
 
 [project]
 name = "context-loader"
-version = "1.3.4"
+version = "1.3.6"
 description = "Render deterministic local Git context for any executor"
 license = { file = "LICENSE" }
 readme = "README.md"

{context_loader-1.3.4 → context_loader-1.3.6}/tool_cli_contract.json

@@ -46,5 +46,5 @@
   ],
   "schema_version": 1,
   "tool_name": "context-loader",
-  "tool_version": "1.3.4"
+  "tool_version": "1.3.6"
 }