diffctx 1.12.2__tar.gz → 1.12.3__tar.gz

{diffctx-1.12.2 → diffctx-1.12.3}/CHANGELOG.md

@@ -7,8 +7,74 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Added
+
+- **Scoop install on Windows.** This repository is now itself the bucket:
+  `scoop bucket add diffctx https://github.com/nikolay-e/diffctx` then
+  `scoop install diffctx/diffctx`. The manifest moved from `packaging/scoop/`
+  to `bucket/diffctx.json`, where `scoop bucket add` reads it, so CD pushing
+  the regenerated manifest to `main` is the publication step. Previously the
+  manifest was correct but unreachable — no bucket existed, so
+  `scoop search diffctx` found nothing.
+
+- `--with-raw-diff` bundles git's raw unified diff ahead of the selected
+  context in every Python-CLI format (md/yaml/json/txt) and in the Python API
+  (`build_diff_context(..., with_raw_diff=True)`). Additive only — selection is
+  byte-identical with and without it — and not charged to `--budget`; the
+  stderr token summary reports the real output size and breaks out the patch's
+  share. Lock-file (#112), ignored and secret-like sections stay omitted. Not
+  in the native binary or the MCP server yet (#150).
+- Token counting is documented explicitly: every count and `--budget` use
+  tiktoken `o200k_base` (GPT-4o family) and are approximate for other model
+  families — `docs/product/token-budget.md`, plus a `Token counting` section in
+  `--help` (#150).
+- A GitHub Action (`action.yml`) runs diffctx as a CI step and exposes the
+  context file, its exact token count and an `empty` flag as step outputs, so a
+  downstream job can feed an LLM without re-tokenizing
+  (`docs/product/github-action.md`, #145).
+- The Claude Code plugin declares its MCP server self-bootstrapping:
+  `uvx --from 'diffctx[mcp]' diffctx-mcp`. Installing the plugin no longer
+  assumes `pip install 'diffctx[mcp]'` has already happened — `uv` fetches it
+  on first use. (The entry point rather than the `diffctx mcp` subcommand: the
+  subcommand only exists from this release on, so on any older version the
+  bare command would try to map a directory named `mcp`.)
+- `diffctx mcp` starts the MCP server, alongside the existing `diffctx-mcp`
+  console script. MCP registries publish a *package* name, and clients that
+  derive the executable from it run `diffctx` — which started a tree-mapping
+  run and wrote 31 MB of the working directory into the protocol transport
+  instead of speaking MCP. Both spellings now reach the same server.
+
+### Security
+
+- **A diff range could re-enable repository-configured diff commands.** The
+  range validator allowed a leading `-`, so `--ext-diff` or `--textconv` passed
+  as a "range" landed in argv *after* the `--no-ext-diff --no-textconv` that
+  every diff invocation sets — undoing them and letting `.gitattributes` run
+  external commands. Neither side of a range may now begin with a dash, and
+  single revisions are additionally rejected for whitespace or control
+  characters (a newline split one `cat-file --batch` request into two). Refs
+  starting with a dash are unaddressable on a git command line, so nothing
+  legitimate is lost.
+- MCP tool calls had no wall-clock deadline — the CLI's watchdog covered only
+  the CLI, leaving parse, fragmentation and scoring unbounded on the MCP path.
+  All three tools now share the CLI's 300 s default.
+- `SECURITY.md` documents prompt injection via repository content honestly: it
+  is inherent to moving repository text into a model's context, diffctx does
+  not detect or neutralize it, and output must be treated as untrusted.
+
 ### Fixed
 
+- **A changed file could vanish from the diff entirely when `.gitignore`
+  excluded its parent directory.** `check-ignore --no-index` — required for
+  `.diffctx/ignore` to apply to tracked files — also revives git's rule that a
+  file cannot be re-included once a parent directory is excluded. pandoc
+  excludes every dotted root entry with `/*.*` and re-includes `!.github/**`;
+  git keeps the tracked workflow file, diffctx dropped it, and a one-file
+  commit rendered as an empty selection with exit 4. An exclusion inherited
+  from an excluded ancestor no longer counts; a pattern matching the path
+  itself still excludes it, so `.diffctx/ignore` and per-directory
+  `.gitignore` rules are unaffected (#153).
+
 - **The container images ran as root.** The published 1.12.1 image had no
   `USER` directive, so every `docker run` executed as uid 0 while the README
   promised an unprivileged user. The image now runs as uid 10001.
@@ -20,8 +86,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - **The native binary hard-capped every run at 4096 tokens.** `--budget`
   carried a fixed clap default instead of leaving the budget unset, so the
   auto-sizing the Python CLI has always used never ran on the binary shipped
-  via crates.io, npm, the container images, Scoop, AUR and the release
-  archives — the same command returned a truncated selection there (14-file
+  via crates.io, npm, the container images and the release archives — the
+  same command returned a truncated selection there (14-file
   self-eat range: 15 fragments across 10 files, against 172 across 44). The
   binary now defaults to auto sizing, and `--budget -1` means unlimited as it
   does in Python.
@@ -58,6 +124,13 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   actionable facts about such a run — were dropped. They are now always
   written.
 
+### Removed
+
+- **AUR support.** `diffctx-bin` had never been submitted (the AUR RPC reported
+  `resultcount: 0`), the publishing job gated on a credential this repository
+  does not hold, and every release regenerated a PKGBUILD nobody could install.
+  Arch users are served by `pipx install diffctx` and `cargo install diffctx`.
+
 ### Changed
 
 - **Lock files no longer render their hunks in diff mode.** A dependency bump
@@ -98,10 +171,11 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   the release tag and smoke-tested against a real repository before the tag
   moves: `docker run --rm -v "$PWD:/repo" ghcr.io/nikolay-e/diffctx . --diff HEAD~1`.
   Mirrored to Docker Hub as `nikolajer/diffctx`.
-- Packaging manifests generated from the release checksums: Scoop
-  (`packaging/scoop/diffctx.json`), AUR (`packaging/aur/`) and an npm wrapper
-  (`packaging/npm/`) that downloads the platform binary and verifies its
-  SHA-256 against the published checksum.
+- npm wrapper (`packaging/npm/`) that downloads the platform binary and
+  verifies its SHA-256 against the published checksum. Scoop
+  (`packaging/scoop/diffctx.json`) and AUR (`packaging/aur/`) manifests are
+  generated from the same checksums but are not published to any bucket or
+  to the AUR.
 
 ### Fixed
 

{diffctx-1.12.2 → diffctx-1.12.3}/Cargo.lock

@@ -179,9 +179,9 @@ dependencies = [
 
 [[package]]
 name = "crossbeam-epoch"
-version = "0.9.18"
+version = "0.9.20"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5b82ac4a3c2ca9c3460964f020e1402edd5753411d7737aa39c3714ad1b5420e"
+checksum = "2d6914041f254d6e9176c01941b21115dcfb7089e55135a35411081bd106ef3f"
 dependencies = [
  "crossbeam-utils",
 ]
@@ -194,7 +194,7 @@ checksum = "d0a5c400df2834b80a4c3327b3aad3a4c4cd4de0629063962b03235697506a28"
 
 [[package]]
 name = "diffctx"
-version = "1.12.2"
+version = "1.12.3"
 dependencies = [
  "anyhow",
  "clap",
@@ -399,9 +399,9 @@ checksum = "09edd9e8b54e49e587e4f6295a7d29c3ea94d469cb40ab8ca70b288248a81db2"
 
 [[package]]
 name = "libc"
-version = "0.2.186"
+version = "0.2.189"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "68ab91017fe16c622486840e4c83c9a37afeff978bd239b5293d61ece587de66"
+checksum = "3eaf3ede3fee6db1a4c2ee091bf8a8b4dccdc6d17f656fb07896ee72867612f2"
 
 [[package]]
 name = "libmimalloc-sys"
@@ -1105,9 +1105,9 @@ dependencies = [
 
 [[package]]
 name = "tree-sitter-javascript"
-version = "0.23.1"
+version = "0.25.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bf40bf599e0416c16c125c3cec10ee5ddc7d1bb8b0c60fa5c4de249ad34dc1b1"
+checksum = "68204f2abc0627a90bdf06e605f5c470aa26fdcb2081ea553a04bdad756693f5"
 dependencies = [
  "cc",
  "tree-sitter-language",
@@ -1211,9 +1211,9 @@ dependencies = [
 
 [[package]]
 name = "tree-sitter-python"
-version = "0.23.6"
+version = "0.25.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3d065aaa27f3aaceaf60c1f0e0ac09e1cb9eb8ed28e7bcdaa52129cffc7f4b04"
+checksum = "6bf85fd39652e740bf60f46f4cda9492c3a9ad75880575bf14960f775cb74a1c"
 dependencies = [
  "cc",
  "tree-sitter-language",
@@ -1271,9 +1271,9 @@ dependencies = [
 
 [[package]]
 name = "tree-sitter-swift"
-version = "0.6.0"
+version = "0.7.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d65aeb41726119416567d0333ec17580ac4abfb96db1f67c4bd638c65f9992fe"
+checksum = "fe36052155b9dd69ca82b3b8f1b4ccfb2d867125ac1a4db1dd7331829242668c"
 dependencies = [
  "cc",
  "tree-sitter-language",

diffctx-1.12.3/Cargo.toml

@@ -0,0 +1,20 @@
+[workspace]
+members = ["crates/diffctx-native"]
+resolver = "2"
+
+[profile.release]
+opt-level = 3
+lto = "fat"
+codegen-units = 1
+strip = true
+panic = "abort"
+debug = "line-tables-only"
+split-debuginfo = "packed"
+
+# `cargo test --release` with panic = "abort" compiles the whole dependency
+# graph twice (abort for the cdylib lib artifact, unwind for the test
+# universe); the hashless cdylib output names then collide (cargo#6313).
+# Tests run through this profile instead: same optimizations, one universe.
+[profile.release-unwind]
+inherits = "release"
+panic = "unwind"

{diffctx-1.12.2 → diffctx-1.12.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: diffctx
-Version: 1.12.2
+Version: 1.12.3
 Classifier: Development Status :: 5 - Production/Stable
 Classifier: Environment :: Console
 Classifier: Intended Audience :: Developers
@@ -105,6 +105,9 @@ changed lines and stops once more context stops paying for itself.
 | Python API | ✗ | ✗ | ✗ | ✓ |
 | MCP server | ✗ | ✗ | ✗ | ✓ |
 
+Fuller positioning — measured results, and when a whole-repo packer or a
+persistent code-graph server fits better: [COMPARISON.md](COMPARISON.md).
+
 ## Install
 
 ```bash
@@ -122,6 +125,13 @@ npx diffctx . --diff HEAD~1             # npm wrapper over the native binary
 docker run --rm -v "$PWD:/repo" ghcr.io/nikolay-e/diffctx . --diff HEAD~1
 ```
 
+On Windows, via Scoop (this repository is the bucket):
+
+```powershell
+scoop bucket add diffctx https://github.com/nikolay-e/diffctx
+scoop install diffctx/diffctx
+```
+
 The image runs as a non-root user (uid 10001) and writes to stdout — the
 native binary has no `-o` flag, so redirect to capture: `... --diff HEAD~1 >
 context.yaml`.
@@ -158,11 +168,12 @@ the `--budget` token cap is hit.
 | Flag        | Default | Description                                                              |
 |-------------|---------|--------------------------------------------------------------------------|
 | `--scoring` | `ego`   | `ego` = bounded expansion around changed nodes (fast, predictable radius); `ppr` = Personalized PageRank (global, smoother decay, slower); `bm25` = lexical retrieval against the diff hunks (baseline for sparse graphs) |
-| `--budget`  | auto    | Hard token cap: `N` enforces a fixed cap, `-1` disables it, `0` is a strict-zero floor (empty selection; use `--full` for changed files only) |
+| `--budget`  | auto    | Hard cap in o200k_base tokens (see [Token counting](docs/product/token-budget.md)): `N` enforces a fixed cap, `-1` disables it, `0` is a strict-zero floor (empty selection; use `--full` for changed files only) |
 | `--alpha`   | 0.60    | PPR damping; higher = context clusters tighter around changes (`--scoring ppr` only) |
 | `--tau`     | 0.12    | Relevance threshold for full fragment content; lower-scoring fragments are stubbed or dropped (lower = more context) |
 | `--full`    | false   | Only the changed files, every fragment, no related-code context          |
 | `--timeout` | 300     | Wall-clock deadline in seconds; on expiry diffctx exits 124 instead of hanging |
+| `--with-raw-diff` | false | Also embed git's raw unified diff ahead of the selected fragments — additive (selection unchanged), not charged to `--budget`, lock/ignored/secret-like sections omitted. Python CLI only |
 
 Calibration of `--alpha`, `--tau`, and the edge-weight priors:
 [`docs/engineering/parameter-strategy.md`](docs/engineering/parameter-strategy.md).
@@ -186,19 +197,20 @@ diffctx graph . --level file -f graphml -o g.xml # file-level graph as GraphML
 <!-- BEGIN USAGE -->
 ```bash
 # full codebase export:
-diffctx .                                # Markdown to stdout + token count
-diffctx . -f md -c                       # Markdown → clipboard
-diffctx . -f json -o tree.json           # JSON → file
-diffctx . --no-content                   # structure only, no file contents
-diffctx . --max-depth 3                  # limit depth
-diffctx . -i custom.ignore               # custom ignore patterns
+diffctx .                                 # Markdown to stdout + token count
+diffctx . -f md -c                        # Markdown → clipboard
+diffctx . -f json -o tree.json            # JSON → file
+diffctx . --no-content                    # structure only, no file contents
+diffctx . --max-depth 3                   # limit depth
+diffctx . -i custom.ignore                # custom ignore patterns
 
 # diff context mode (requires git repo):
-diffctx . --diff                         # uncommitted changes (working tree vs HEAD)
-diffctx . --diff HEAD~1                  # context for last commit
-diffctx . --diff main..feature           # context for feature branch
-diffctx . --diff HEAD~1 --budget 30000   # limit to ~30k tokens
-diffctx . --diff HEAD~1 -c               # diff context to clipboard
+diffctx . --diff                          # uncommitted changes (working tree vs HEAD)
+diffctx . --diff HEAD~1                   # context for last commit
+diffctx . --diff main..feature            # context for feature branch
+diffctx . --diff HEAD~1 --budget 30000    # limit to ~30k tokens
+diffctx . --diff HEAD~1 -c                # diff context to clipboard
+diffctx . --diff HEAD~1 --with-raw-diff   # raw patch + selected context
 ```
 <!-- END USAGE -->
 
@@ -209,6 +221,11 @@ Every run reports token count and size on stderr — `12,847 tokens
 `<binary file: N bytes>`, `<file too large: N bytes>`, or
 `<unreadable content: not utf-8>`.
 
+Counts come from tiktoken's `o200k_base` encoder and are exact only for the
+GPT-4o family; Claude, Gemini and others tokenize differently, so treat
+`--budget` as an upper bound in o200k tokens and leave headroom. Details:
+[`docs/product/token-budget.md`](docs/product/token-budget.md).
+
 ## Python API
 
 ```python
@@ -224,6 +241,7 @@ ctx = build_diff_context(
     full=False,
     scoring_mode="ego",
     timeout=300,
+    with_raw_diff=False,      # True also embeds the raw unified diff (not charged to budget)
 )
 print(to_markdown(ctx))
 
@@ -264,6 +282,12 @@ anchored `/root_only.txt`). `.diffctx/whitelist` acts as an include-only filter,
 and the output file is always auto-ignored. `--no-default-ignores` disables the
 built-in patterns; `--no-ignores` disables all ignore rules (tree mode only).
 
+An excluded path never appears in the output in any role: in diff mode it is
+dropped both from `changed_files` and from the candidate universe, so it cannot
+come back as a related-context fragment either (including under `--full`). The
+same guarantee covers secret-like paths (`id_rsa`, `*.pem`, `*.key`, ...),
+which are filtered even without an ignore entry.
+
 ## Token cache
 
 Diff mode caches per-blob tokenization under
@@ -313,6 +337,12 @@ Apache 2.0
 
 - [Documentation site](https://nikolay-e.github.io/diffctx/) — the pipeline
   end to end: diff → fragments → graph → relevance → selection
+- [GitHub Action](docs/product/github-action.md) — diff context as a CI step
+  for LLM review
+- [Token counting](docs/product/token-budget.md) — which encoder, and what
+  `--budget` means for non-GPT models
+- [Comparison](COMPARISON.md) — measured results, and when a whole-repo packer
+  or a persistent code-graph server fits better
 - [Changelog](CHANGELOG.md)
 - [Security policy](SECURITY.md) — threat model and vulnerability reporting
 - [Parameter strategy](docs/engineering/parameter-strategy.md) — how `--alpha`,

{diffctx-1.12.2/crates/diffctx-native → diffctx-1.12.3}/README.md

@@ -27,6 +27,9 @@ changed lines and stops once more context stops paying for itself.
 | Python API | ✗ | ✗ | ✗ | ✓ |
 | MCP server | ✗ | ✗ | ✗ | ✓ |
 
+Fuller positioning — measured results, and when a whole-repo packer or a
+persistent code-graph server fits better: [COMPARISON.md](COMPARISON.md).
+
 ## Install
 
 ```bash
@@ -44,6 +47,13 @@ npx diffctx . --diff HEAD~1             # npm wrapper over the native binary
 docker run --rm -v "$PWD:/repo" ghcr.io/nikolay-e/diffctx . --diff HEAD~1
 ```
 
+On Windows, via Scoop (this repository is the bucket):
+
+```powershell
+scoop bucket add diffctx https://github.com/nikolay-e/diffctx
+scoop install diffctx/diffctx
+```
+
 The image runs as a non-root user (uid 10001) and writes to stdout — the
 native binary has no `-o` flag, so redirect to capture: `... --diff HEAD~1 >
 context.yaml`.
@@ -80,11 +90,12 @@ the `--budget` token cap is hit.
 | Flag        | Default | Description                                                              |
 |-------------|---------|--------------------------------------------------------------------------|
 | `--scoring` | `ego`   | `ego` = bounded expansion around changed nodes (fast, predictable radius); `ppr` = Personalized PageRank (global, smoother decay, slower); `bm25` = lexical retrieval against the diff hunks (baseline for sparse graphs) |
-| `--budget`  | auto    | Hard token cap: `N` enforces a fixed cap, `-1` disables it, `0` is a strict-zero floor (empty selection; use `--full` for changed files only) |
+| `--budget`  | auto    | Hard cap in o200k_base tokens (see [Token counting](docs/product/token-budget.md)): `N` enforces a fixed cap, `-1` disables it, `0` is a strict-zero floor (empty selection; use `--full` for changed files only) |
 | `--alpha`   | 0.60    | PPR damping; higher = context clusters tighter around changes (`--scoring ppr` only) |
 | `--tau`     | 0.12    | Relevance threshold for full fragment content; lower-scoring fragments are stubbed or dropped (lower = more context) |
 | `--full`    | false   | Only the changed files, every fragment, no related-code context          |
 | `--timeout` | 300     | Wall-clock deadline in seconds; on expiry diffctx exits 124 instead of hanging |
+| `--with-raw-diff` | false | Also embed git's raw unified diff ahead of the selected fragments — additive (selection unchanged), not charged to `--budget`, lock/ignored/secret-like sections omitted. Python CLI only |
 
 Calibration of `--alpha`, `--tau`, and the edge-weight priors:
 [`docs/engineering/parameter-strategy.md`](docs/engineering/parameter-strategy.md).
@@ -108,19 +119,20 @@ diffctx graph . --level file -f graphml -o g.xml # file-level graph as GraphML
 <!-- BEGIN USAGE -->
 ```bash
 # full codebase export:
-diffctx .                                # Markdown to stdout + token count
-diffctx . -f md -c                       # Markdown → clipboard
-diffctx . -f json -o tree.json           # JSON → file
-diffctx . --no-content                   # structure only, no file contents
-diffctx . --max-depth 3                  # limit depth
-diffctx . -i custom.ignore               # custom ignore patterns
+diffctx .                                 # Markdown to stdout + token count
+diffctx . -f md -c                        # Markdown → clipboard
+diffctx . -f json -o tree.json            # JSON → file
+diffctx . --no-content                    # structure only, no file contents
+diffctx . --max-depth 3                   # limit depth
+diffctx . -i custom.ignore                # custom ignore patterns
 
 # diff context mode (requires git repo):
-diffctx . --diff                         # uncommitted changes (working tree vs HEAD)
-diffctx . --diff HEAD~1                  # context for last commit
-diffctx . --diff main..feature           # context for feature branch
-diffctx . --diff HEAD~1 --budget 30000   # limit to ~30k tokens
-diffctx . --diff HEAD~1 -c               # diff context to clipboard
+diffctx . --diff                          # uncommitted changes (working tree vs HEAD)
+diffctx . --diff HEAD~1                   # context for last commit
+diffctx . --diff main..feature            # context for feature branch
+diffctx . --diff HEAD~1 --budget 30000    # limit to ~30k tokens
+diffctx . --diff HEAD~1 -c                # diff context to clipboard
+diffctx . --diff HEAD~1 --with-raw-diff   # raw patch + selected context
 ```
 <!-- END USAGE -->
 
@@ -131,6 +143,11 @@ Every run reports token count and size on stderr — `12,847 tokens
 `<binary file: N bytes>`, `<file too large: N bytes>`, or
 `<unreadable content: not utf-8>`.
 
+Counts come from tiktoken's `o200k_base` encoder and are exact only for the
+GPT-4o family; Claude, Gemini and others tokenize differently, so treat
+`--budget` as an upper bound in o200k tokens and leave headroom. Details:
+[`docs/product/token-budget.md`](docs/product/token-budget.md).
+
 ## Python API
 
 ```python
@@ -146,6 +163,7 @@ ctx = build_diff_context(
     full=False,
     scoring_mode="ego",
     timeout=300,
+    with_raw_diff=False,      # True also embeds the raw unified diff (not charged to budget)
 )
 print(to_markdown(ctx))
 
@@ -186,6 +204,12 @@ anchored `/root_only.txt`). `.diffctx/whitelist` acts as an include-only filter,
 and the output file is always auto-ignored. `--no-default-ignores` disables the
 built-in patterns; `--no-ignores` disables all ignore rules (tree mode only).
 
+An excluded path never appears in the output in any role: in diff mode it is
+dropped both from `changed_files` and from the candidate universe, so it cannot
+come back as a related-context fragment either (including under `--full`). The
+same guarantee covers secret-like paths (`id_rsa`, `*.pem`, `*.key`, ...),
+which are filtered even without an ignore entry.
+
 ## Token cache
 
 Diff mode caches per-blob tokenization under
@@ -235,6 +259,12 @@ Apache 2.0
 
 - [Documentation site](https://nikolay-e.github.io/diffctx/) — the pipeline
   end to end: diff → fragments → graph → relevance → selection
+- [GitHub Action](docs/product/github-action.md) — diff context as a CI step
+  for LLM review
+- [Token counting](docs/product/token-budget.md) — which encoder, and what
+  `--budget` means for non-GPT models
+- [Comparison](COMPARISON.md) — measured results, and when a whole-repo packer
+  or a persistent code-graph server fits better
 - [Changelog](CHANGELOG.md)
 - [Security policy](SECURITY.md) — threat model and vulnerability reporting
 - [Parameter strategy](docs/engineering/parameter-strategy.md) — how `--alpha`,

{diffctx-1.12.2 → diffctx-1.12.3}/crates/diffctx-native/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "diffctx"
-version = "1.12.2"
+version = "1.12.3"
 edition = "2024"
 rust-version = "1.85"
 description = "Selects the minimum code an LLM needs to review a git diff: walks the dependency graph outward from changed lines and stops when extra context stops paying for itself"
@@ -72,8 +72,8 @@ lang-extra = [
 mimalloc = { version = "0.1", default-features = false }
 clap = { version = "4", features = ["derive"] }
 tree-sitter = "0.25"
-tree-sitter-python = { version = "0.23", optional = true }
-tree-sitter-javascript = { version = "0.23", optional = true }
+tree-sitter-python = { version = "0.25", optional = true }
+tree-sitter-javascript = { version = "0.25", optional = true }
 tree-sitter-typescript = { version = "0.23", optional = true }
 tree-sitter-go = { version = "0.25", optional = true }
 tree-sitter-rust = { version = "0.24", optional = true }
@@ -84,7 +84,7 @@ tree-sitter-c-sharp = { version = "0.23", optional = true }
 tree-sitter-ruby = { version = "0.23", optional = true }
 tree-sitter-php = { version = "0.24", optional = true }
 tree-sitter-scala = { version = "0.26", optional = true }
-tree-sitter-swift = { version = "0.6", optional = true }
+tree-sitter-swift = { version = "0.7", optional = true }
 tree-sitter-html = { version = "0.23", optional = true }
 tree-sitter-bash = { version = "0.25", optional = true }
 tree-sitter-css = { version = "0.25", optional = true }

{diffctx-1.12.2 → diffctx-1.12.3/crates/diffctx-native}/README.md

@@ -27,6 +27,9 @@ changed lines and stops once more context stops paying for itself.
 | Python API | ✗ | ✗ | ✗ | ✓ |
 | MCP server | ✗ | ✗ | ✗ | ✓ |
 
+Fuller positioning — measured results, and when a whole-repo packer or a
+persistent code-graph server fits better: [COMPARISON.md](COMPARISON.md).
+
 ## Install
 
 ```bash
@@ -44,6 +47,13 @@ npx diffctx . --diff HEAD~1             # npm wrapper over the native binary
 docker run --rm -v "$PWD:/repo" ghcr.io/nikolay-e/diffctx . --diff HEAD~1
 ```
 
+On Windows, via Scoop (this repository is the bucket):
+
+```powershell
+scoop bucket add diffctx https://github.com/nikolay-e/diffctx
+scoop install diffctx/diffctx
+```
+
 The image runs as a non-root user (uid 10001) and writes to stdout — the
 native binary has no `-o` flag, so redirect to capture: `... --diff HEAD~1 >
 context.yaml`.
@@ -80,11 +90,12 @@ the `--budget` token cap is hit.
 | Flag        | Default | Description                                                              |
 |-------------|---------|--------------------------------------------------------------------------|
 | `--scoring` | `ego`   | `ego` = bounded expansion around changed nodes (fast, predictable radius); `ppr` = Personalized PageRank (global, smoother decay, slower); `bm25` = lexical retrieval against the diff hunks (baseline for sparse graphs) |
-| `--budget`  | auto    | Hard token cap: `N` enforces a fixed cap, `-1` disables it, `0` is a strict-zero floor (empty selection; use `--full` for changed files only) |
+| `--budget`  | auto    | Hard cap in o200k_base tokens (see [Token counting](docs/product/token-budget.md)): `N` enforces a fixed cap, `-1` disables it, `0` is a strict-zero floor (empty selection; use `--full` for changed files only) |
 | `--alpha`   | 0.60    | PPR damping; higher = context clusters tighter around changes (`--scoring ppr` only) |
 | `--tau`     | 0.12    | Relevance threshold for full fragment content; lower-scoring fragments are stubbed or dropped (lower = more context) |
 | `--full`    | false   | Only the changed files, every fragment, no related-code context          |
 | `--timeout` | 300     | Wall-clock deadline in seconds; on expiry diffctx exits 124 instead of hanging |
+| `--with-raw-diff` | false | Also embed git's raw unified diff ahead of the selected fragments — additive (selection unchanged), not charged to `--budget`, lock/ignored/secret-like sections omitted. Python CLI only |
 
 Calibration of `--alpha`, `--tau`, and the edge-weight priors:
 [`docs/engineering/parameter-strategy.md`](docs/engineering/parameter-strategy.md).
@@ -108,19 +119,20 @@ diffctx graph . --level file -f graphml -o g.xml # file-level graph as GraphML
 <!-- BEGIN USAGE -->
 ```bash
 # full codebase export:
-diffctx .                                # Markdown to stdout + token count
-diffctx . -f md -c                       # Markdown → clipboard
-diffctx . -f json -o tree.json           # JSON → file
-diffctx . --no-content                   # structure only, no file contents
-diffctx . --max-depth 3                  # limit depth
-diffctx . -i custom.ignore               # custom ignore patterns
+diffctx .                                 # Markdown to stdout + token count
+diffctx . -f md -c                        # Markdown → clipboard
+diffctx . -f json -o tree.json            # JSON → file
+diffctx . --no-content                    # structure only, no file contents
+diffctx . --max-depth 3                   # limit depth
+diffctx . -i custom.ignore                # custom ignore patterns
 
 # diff context mode (requires git repo):
-diffctx . --diff                         # uncommitted changes (working tree vs HEAD)
-diffctx . --diff HEAD~1                  # context for last commit
-diffctx . --diff main..feature           # context for feature branch
-diffctx . --diff HEAD~1 --budget 30000   # limit to ~30k tokens
-diffctx . --diff HEAD~1 -c               # diff context to clipboard
+diffctx . --diff                          # uncommitted changes (working tree vs HEAD)
+diffctx . --diff HEAD~1                   # context for last commit
+diffctx . --diff main..feature            # context for feature branch
+diffctx . --diff HEAD~1 --budget 30000    # limit to ~30k tokens
+diffctx . --diff HEAD~1 -c                # diff context to clipboard
+diffctx . --diff HEAD~1 --with-raw-diff   # raw patch + selected context
 ```
 <!-- END USAGE -->
 
@@ -131,6 +143,11 @@ Every run reports token count and size on stderr — `12,847 tokens
 `<binary file: N bytes>`, `<file too large: N bytes>`, or
 `<unreadable content: not utf-8>`.
 
+Counts come from tiktoken's `o200k_base` encoder and are exact only for the
+GPT-4o family; Claude, Gemini and others tokenize differently, so treat
+`--budget` as an upper bound in o200k tokens and leave headroom. Details:
+[`docs/product/token-budget.md`](docs/product/token-budget.md).
+
 ## Python API
 
 ```python
@@ -146,6 +163,7 @@ ctx = build_diff_context(
     full=False,
     scoring_mode="ego",
     timeout=300,
+    with_raw_diff=False,      # True also embeds the raw unified diff (not charged to budget)
 )
 print(to_markdown(ctx))
 
@@ -186,6 +204,12 @@ anchored `/root_only.txt`). `.diffctx/whitelist` acts as an include-only filter,
 and the output file is always auto-ignored. `--no-default-ignores` disables the
 built-in patterns; `--no-ignores` disables all ignore rules (tree mode only).
 
+An excluded path never appears in the output in any role: in diff mode it is
+dropped both from `changed_files` and from the candidate universe, so it cannot
+come back as a related-context fragment either (including under `--full`). The
+same guarantee covers secret-like paths (`id_rsa`, `*.pem`, `*.key`, ...),
+which are filtered even without an ignore entry.
+
 ## Token cache
 
 Diff mode caches per-blob tokenization under
@@ -235,6 +259,12 @@ Apache 2.0
 
 - [Documentation site](https://nikolay-e.github.io/diffctx/) — the pipeline
   end to end: diff → fragments → graph → relevance → selection
+- [GitHub Action](docs/product/github-action.md) — diff context as a CI step
+  for LLM review
+- [Token counting](docs/product/token-budget.md) — which encoder, and what
+  `--budget` means for non-GPT models
+- [Comparison](COMPARISON.md) — measured results, and when a whole-repo packer
+  or a persistent code-graph server fits better
 - [Changelog](CHANGELOG.md)
 - [Security policy](SECURITY.md) — threat model and vulnerability reporting
 - [Parameter strategy](docs/engineering/parameter-strategy.md) — how `--alpha`,