chore(native): sync native release v0.4.1-1

CHANGELOG.md

@@ -1,5 +1,11 @@
 ## Unreleased
 
+* Updated the default llama.cpp native runtime pin to
+  `leehack/llamadart-native@v0.4.1-1`, keeping the `v0.4.1` llama.cpp
+  ABI/bindings while picking up wrapper-only native fixes. Refreshed
+  the `llamadart_llama_cpp_flutter` Apple SwiftPM checksum and
+  aligned current README/website native override docs.
+
 - Extend the GGUF speech-to-text validation pack with four synthetic edge
   fixtures built in-process, so no extra audio is stored: generated digital
   silence, plus a truncated RIFF, a stereo 44.1 kHz re-encode and a 33-second

README.md

@@ -152,7 +152,7 @@ Current default runtime pins:
 
 | Runtime | Pin |
 | --- | --- |
-| Native llama.cpp / GGUF | `leehack/llamadart-native@v0.4.1` |
+| Native llama.cpp / GGUF | `leehack/llamadart-native@v0.4.1-1` |
 | Native LiteRT-LM / `.litertlm` | `leehack/litert-lm-native@v0.17.0-6` |
 | Web llama.cpp / GGUF | `leehack/llama-web-bridge-assets@v0.1.44` |
 | Web LiteRT-LM / `.litertlm` | `@litert-lm/core@0.15.0` |

lib/src/backends/llama_cpp/bindings.dart

@@ -6298,6 +6298,15 @@ external int llama_dart_tts_api_version();
 @ffi.Native<ffi.Void Function(ffi.Pointer<llama_dart_tts>)>()
 external void llama_dart_tts_cancel(ffi.Pointer<llama_dart_tts> tts);
 
+@ffi.Native<
+  ffi.Bool Function(ffi.Pointer<ggml_tensor>, ffi.Bool, ffi.Pointer<ffi.Void>)
+>()
+external bool llama_dart_tts_eval_callback(
+  ffi.Pointer<ggml_tensor> tensor,
+  bool ask,
+  ffi.Pointer<ffi.Void> user_data,
+);
+
 @ffi.Native<ffi.Void Function(ffi.Pointer<llama_dart_tts>)>()
 external void llama_dart_tts_free(ffi.Pointer<llama_dart_tts> tts);
 
@@ -6403,6 +6412,23 @@ llama_dart_tts_status llama_dart_tts_reset(ffi.Pointer<llama_dart_tts> tts) {
   return llama_dart_tts_status.fromValue(_llama_dart_tts_reset(tts));
 }
 
+@ffi.Native<
+  ffi.Int Function(ffi.Pointer<llama_dart_tts>, ffi.Pointer<ffi.Int8>)
+>(symbol: 'llama_dart_tts_set_cancel_flag')
+external int _llama_dart_tts_set_cancel_flag(
+  ffi.Pointer<llama_dart_tts> tts,
+  ffi.Pointer<ffi.Int8> flag,
+);
+
+llama_dart_tts_status llama_dart_tts_set_cancel_flag(
+  ffi.Pointer<llama_dart_tts> tts,
+  ffi.Pointer<ffi.Int8> flag,
+) {
+  return llama_dart_tts_status.fromValue(
+    _llama_dart_tts_set_cancel_flag(tts, flag),
+  );
+}
+
 @ffi.Native<
   ffi.Int Function(
     ffi.Pointer<llama_dart_tts>,
@@ -8719,7 +8745,7 @@ external int mtmd_tokenize_from_parts(
   bool add_special,
 );
 
-typedef FILE = __sFILE;
+typedef FILE = _IO_FILE;
 
 const int GGML_BACKEND_META_MAX_DEVICES = 16;
 
@@ -8858,88 +8884,90 @@ final class UnnamedUnion extends ffi.Union {
   external ffi.Array<ffi.Char> val_str;
 }
 
-typedef __darwin_off_t = __int64_t;
-typedef __int64_t = ffi.LongLong;
-typedef Dart__int64_t = int;
+final class _IO_FILE extends ffi.Struct {
+  @ffi.Int()
+  external int _flags;
 
-final class __sFILE extends ffi.Struct {
-  external ffi.Pointer<ffi.UnsignedChar> _p;
+  external ffi.Pointer<ffi.Char> _IO_read_ptr;
 
-  @ffi.Int()
-  external int _r;
+  external ffi.Pointer<ffi.Char> _IO_read_end;
 
-  @ffi.Int()
-  external int _w;
+  external ffi.Pointer<ffi.Char> _IO_read_base;
 
-  @ffi.Short()
-  external int _flags;
+  external ffi.Pointer<ffi.Char> _IO_write_base;
 
-  @ffi.Short()
-  external int _file;
+  external ffi.Pointer<ffi.Char> _IO_write_ptr;
 
-  external __sbuf _bf;
+  external ffi.Pointer<ffi.Char> _IO_write_end;
 
-  @ffi.Int()
-  external int _lbfsize;
+  external ffi.Pointer<ffi.Char> _IO_buf_base;
 
-  external ffi.Pointer<ffi.Void> _cookie;
+  external ffi.Pointer<ffi.Char> _IO_buf_end;
 
-  external ffi.Pointer<
-    ffi.NativeFunction<ffi.Int Function(ffi.Pointer<ffi.Void>)>
-  >
-  _close;
+  external ffi.Pointer<ffi.Char> _IO_save_base;
 
-  external ffi.Pointer<
-    ffi.NativeFunction<
-      ffi.Int Function(ffi.Pointer<ffi.Void>, ffi.Pointer<ffi.Char>, ffi.Int)
-    >
-  >
-  _read;
+  external ffi.Pointer<ffi.Char> _IO_backup_base;
 
-  external ffi.Pointer<
-    ffi.NativeFunction<fpos_t Function(ffi.Pointer<ffi.Void>, fpos_t, ffi.Int)>
-  >
-  _seek;
+  external ffi.Pointer<ffi.Char> _IO_save_end;
 
-  external ffi.Pointer<
-    ffi.NativeFunction<
-      ffi.Int Function(ffi.Pointer<ffi.Void>, ffi.Pointer<ffi.Char>, ffi.Int)
-    >
-  >
-  _write;
+  external ffi.Pointer<_IO_marker> _markers;
 
-  external __sbuf _ub;
+  external ffi.Pointer<_IO_FILE> _chain;
 
-  external ffi.Pointer<__sFILEX> _extra;
+  @ffi.Int()
+  external int _fileno;
 
   @ffi.Int()
-  external int _ur;
+  external int _flags2;
 
-  @ffi.Array.multi([3])
-  external ffi.Array<ffi.UnsignedChar> _ubuf;
+  @__off_t()
+  external int _old_offset;
 
-  @ffi.Array.multi([1])
-  external ffi.Array<ffi.UnsignedChar> _nbuf;
+  @ffi.UnsignedShort()
+  external int _cur_column;
 
-  external __sbuf _lb;
+  @ffi.SignedChar()
+  external int _vtable_offset;
 
-  @ffi.Int()
-  external int _blksize;
+  @ffi.Array.multi([1])
+  external ffi.Array<ffi.Char> _shortbuf;
+
+  external ffi.Pointer<_IO_lock_t> _lock;
 
-  @fpos_t()
+  @__off64_t()
   external int _offset;
-}
 
-final class __sFILEX extends ffi.Opaque {}
+  external ffi.Pointer<_IO_codecvt> _codecvt;
+
+  external ffi.Pointer<_IO_wide_data> _wide_data;
+
+  external ffi.Pointer<_IO_FILE> _freeres_list;
+
+  external ffi.Pointer<ffi.Void> _freeres_buf;
 
-final class __sbuf extends ffi.Struct {
-  external ffi.Pointer<ffi.UnsignedChar> _base;
+  @ffi.Size()
+  external int __pad5;
 
   @ffi.Int()
-  external int _size;
+  external int _mode;
+
+  @ffi.Array.multi([20])
+  external ffi.Array<ffi.Char> _unused2;
 }
 
-typedef fpos_t = __darwin_off_t;
+final class _IO_codecvt extends ffi.Opaque {}
+
+typedef _IO_lock_t = ffi.Void;
+typedef Dart_IO_lock_t = void;
+
+final class _IO_marker extends ffi.Opaque {}
+
+final class _IO_wide_data extends ffi.Opaque {}
+
+typedef __off64_t = ffi.Long;
+typedef Dart__off64_t = int;
+typedef __off_t = ffi.Long;
+typedef Dart__off_t = int;
 typedef ggml_abort_callback =
     ffi.Pointer<ffi.NativeFunction<ggml_abort_callbackFunction>>;
 typedef ggml_abort_callbackFunction =

lib/src/hook/native_release_pins.dart

@@ -5,7 +5,7 @@
 library;
 
 /// `leehack/llamadart-native` release tag downloaded for the llama.cpp runtime.
-const llamaCppTag = 'v0.4.1';
+const llamaCppTag = 'v0.4.1-1';
 
 /// `leehack/litert-lm-native` release tag downloaded for the LiteRT-LM runtime.
 const liteRtLmReleaseTag = 'v0.17.0-6';

packages/llamadart_llama_cpp_flutter/CHANGELOG.md

@@ -1,3 +1,7 @@
+## Unreleased
+
+* Updated Apple SwiftPM native pin to `leehack/llamadart-native@v0.4.1-1`.
+
 ## 0.0.19
 
 * Updated Apple SwiftPM native pin to `leehack/llamadart-native@v0.4.1`.

packages/llamadart_llama_cpp_flutter/README.md

@@ -20,7 +20,7 @@ dependencies:
 This package has no runtime Dart API of its own. Import `package:llamadart`
 normally from the core package and use `LlamaBackend()` / `LlamaEngine` there.
 
-The Apple SwiftPM manifest pins `leehack/llamadart-native@v0.4.1`.
+The Apple SwiftPM manifest pins `leehack/llamadart-native@v0.4.1-1`.
 
 Source for this package lives in
 `packages/llamadart_llama_cpp_flutter` in the

packages/llamadart_llama_cpp_flutter/darwin/llamadart_llama_cpp_flutter/Package.swift

@@ -4,7 +4,7 @@ import PackageDescription
 
 let packageRoot = URL(fileURLWithPath: #filePath).deletingLastPathComponent()
 let artifactsRoot = packageRoot.appendingPathComponent("Artifacts")
-let llamaCppTag = "v0.4.1"
+let llamaCppTag = "v0.4.1-1"
 
 func localArtifactPath(_ name: String) -> String? {
     let path = artifactsRoot.appendingPathComponent(name).path
@@ -47,7 +47,7 @@ let package = Package(
             repository: "leehack/llamadart-native",
             artifactName: "llamadart-native-apple-xcframework-\(llamaCppTag).zip",
             tag: llamaCppTag,
-            checksum: "16175a4084a14af63cc66e78b9b2a8173d2e14db86775af280c859438bd8616a"
+            checksum: "bab353ce17dfbac9afe65ade73751d67f110b1807a1fb365b4d1df2db25a5418"
         ),
         .target(
             name: "llamadart_llama_cpp_flutter",

website/docs/getting-started/installation.md

@@ -88,7 +88,7 @@ hooks:
     llamadart:
       # Optional. Defaults to llamadart's tested native runtime pin.
       # Use a leehack/llamadart-native release tag when testing another build.
-      llamadart_native_tag: v0.4.1
+      llamadart_native_tag: v0.4.1-1
 
       # Optional. GitHub repository slug or github.com URL.
       llamadart_native_repository: leehack/llamadart-native
@@ -115,7 +115,7 @@ per-target module list.
 Native source overrides are for compatibility testing. They do not regenerate
 Dart FFI bindings or symbol lookups, so the selected binary still must be ABI-
 and symbol-compatible with the default
-`leehack/llamadart-native@v0.4.1` runtime.
+`leehack/llamadart-native@v0.4.1-1` runtime.
 
 Available native tags are published on the
 [`leehack/llamadart-native` releases page](https://github.com/leehack/llamadart-native/releases).
@@ -143,7 +143,7 @@ gh release list --repo leehack/llamadart-native --limit 20
 
 Before overriding, confirm the release includes the asset for your target. The
 hook downloads files named `llamadart-native-<bundle>-<tag>.tar.gz`, for example
-`llamadart-native-windows-x64-v0.4.1.tar.gz`.
+`llamadart-native-windows-x64-v0.4.1-1.tar.gz`.
 For local testing, `llamadart_native_path` may point directly at a bundle
 archive, at an extracted bundle directory, or at a directory containing
 `<tag>/<bundle>/`, `<bundle>/`, or the expected archive file.

website/docs/platforms/support-matrix.md

@@ -8,7 +8,7 @@ backend-module configuration for
 `llamadart`.
 
 The native-assets hook currently pins `llamadart-native` tag
-`v0.4.1` and
+`v0.4.1-1` and
 `litert-lm-native` release `v0.17.0-6` (`lib/src/hook/native_release_pins.dart`). Apps can
 override the llama.cpp native GitHub source with
 `hooks.user_defines.llamadart.llamadart_native_tag` and
@@ -306,7 +306,7 @@ device/model bundle, use `cpu` or `gpu` for that artifact.
   a web load failure as a package bug. The [WebGPU Bridge](./webgpu-bridge)
   page has the browser-console probe and Flutter Web smoke-test path.
 
-## Current llama.cpp module availability by bundle (`v0.4.1`)
+## Current llama.cpp module availability by bundle (`v0.4.1-1`)
 
 | Bundle key | Available backend modules in bundle |
 | --- | --- |
@@ -354,7 +354,7 @@ hooks:
   user_defines:
     llamadart:
       # Optional. Defaults to llamadart's tested native runtime pin.
-      llamadart_native_tag: v0.4.1
+      llamadart_native_tag: v0.4.1-1
 
       # Optional. GitHub repository slug or github.com URL.
       llamadart_native_repository: leehack/llamadart-native