Update README.md

pipedemo: todo

.github/workflows/go.yml

@@ -5,25 +5,45 @@ name: Go
 
 on:
   push:
-    branches: [ "main" ]
+    branches: [ "*" ]
   pull_request:
-    branches: [ "main" ]
+    branches: [ "*" ]
 
 jobs:
 
-  test-macos:
-    runs-on: macos-latest
+  test:
     strategy:
       matrix:
-        llvm: [17]
+        os: [macos-latest, ubuntu-latest]
+        # os: [macos-latest]
+        llvm: [18]
+    runs-on: ${{ matrix.os }}
     steps:
       - uses: actions/checkout@v4
 
       - name: Update Homebrew
-        if: matrix.llvm == 17 # needed as long as LLVM 17 is still fresh
+        # needed as long as LLVM 18 is still fresh
+        if: matrix.llvm == 18 && startsWith(matrix.os, 'macos')
         run: brew update
-      - name: Install LLVM ${{ matrix.llvm }}
-        run: HOMEBREW_NO_AUTO_UPDATE=1 brew install llvm@${{ matrix.llvm }}
+      - name: Install LLVM ${{ matrix.llvm }} and bdw-gc
+        if: startsWith(matrix.os, 'macos')
+        run: |
+          HOMEBREW_NO_AUTO_UPDATE=1 brew install llvm@${{ matrix.llvm }} bdw-gc
+          echo `brew --prefix llvm@${{ matrix.llvm }}`/bin >> $GITHUB_PATH
+      - name: Install LLVM ${{ matrix.llvm }} and libgc-dev
+        if: startsWith(matrix.os, 'ubuntu')
+        run: |
+          echo "deb http://apt.llvm.org/$(lsb_release -cs)/ llvm-toolchain-$(lsb_release -cs)-${{ matrix.llvm }} main" | sudo tee /etc/apt/sources.list.d/llvm.list
+          wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key | sudo apt-key add -
+          sudo apt-get update
+          sudo apt-get install -y llvm-${{ matrix.llvm }}-dev clang-${{ matrix.llvm }} lld-${{ matrix.llvm }} pkg-config libgc-dev libcjson-dev libsqlite3-dev python3.11-dev
+          echo /usr/lib/llvm-${{ matrix.llvm }}/bin >> $GITHUB_PATH
+    
+      - name: Clang information
+        run: |
+          echo $PATH
+          which clang
+          clang --version
         
       - name: Set up Go
         uses: actions/setup-go@v5
@@ -34,36 +54,40 @@ jobs:
         run: go build -v ./...
 
       - name: Test
+        if: matrix.os != 'macos-latest'
         run: go test -v ./...
 
-  test-linux:
-    runs-on: ubuntu-20.04
-    strategy:
-      matrix:
-        llvm: [17]
-    steps:
-    - uses: actions/checkout@v4
+      - name: Test with coverage
+        if: matrix.os == 'macos-latest'
+        run: go test -v -coverprofile="coverage.txt" -covermode=atomic ./...
 
-    - name: Install LLVM ${{ matrix.llvm }}
-      run: |
-        echo 'deb http://apt.llvm.org/focal/ llvm-toolchain-focal-${{ matrix.llvm }} main' | sudo tee /etc/apt/sources.list.d/llvm.list
-        wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key | sudo apt-key add -
-        sudo apt-get update
-        sudo apt-get install --no-install-recommends llvm-${{ matrix.llvm }}-dev
+      - name: Install
+        run: go install ./...
+      
+      - name: LLGO tests
+        if: matrix.os != 'ubuntu-latest'
+        run: |
+          echo "Test result on ${{ matrix.os }} with LLVM ${{ matrix.llvm }}" > result.md
+          LLGOROOT=$PWD bash .github/workflows/test_llgo.sh
+      
+      - name: Test _demo and _pydemo
+        run: |
+          set +e
+          LLGOROOT=$PWD bash .github/workflows/test_demo.sh
+          exit 0
+      
+      - name: Show test result
+        run: cat result.md
 
-    - name: Set up Go
-      uses: actions/setup-go@v5
-      with:
-        go-version: '1.20'
+      - name: PR comment with test result
+        uses: thollander/actions-comment-pull-request@v2
+        if: false
+        with:
+          filePath: result.md
+          comment_tag: test-result-on-${{ matrix.os }}-with-llvm-${{ matrix.llvm }}
 
-    - name: Build
-      run: go build -v ./...
-
-    - name: Test
-      run: go test -v -coverprofile="coverage.txt" -covermode=atomic ./...
-
-    - name: Upload coverage reports to Codecov
-      uses: codecov/codecov-action@v4
-      with:
-        token: ${{ secrets.CODECOV_TOKEN }}
-        slug: goplus/llgo
+      - name: Upload coverage reports to Codecov
+        uses: codecov/codecov-action@v4
+        with:
+          token: ${{ secrets.CODECOV_TOKEN }}
+          slug: goplus/llgo

.github/workflows/populate_darwin_sysroot.sh

@@ -0,0 +1,40 @@
+#!/bin/bash
+
+set -e
+
+TMPDIR="$(mktemp -d)"
+export TMPDIR
+trap 'rm -rf "${TMPDIR}"' EXIT
+
+DARWIN_AMD64_LLVM_PREFIX=.sysroot/darwin/amd64/usr/local/opt/llvm@18
+DARWIN_ARM64_LLVM_PREFIX=.sysroot/darwin/arm64/opt/homebrew/opt/llvm@18
+mkdir -p "${DARWIN_AMD64_LLVM_PREFIX}" "${DARWIN_ARM64_LLVM_PREFIX}"
+
+BREW_LLVM_FORMULA_JSON="$(mktemp)"
+curl -fsSL https://formulae.brew.sh/api/formula/llvm.json > "${BREW_LLVM_FORMULA_JSON}"
+BREW_LLVM_AMD64_BOTTLE_URL=$(jq -r '.bottle.stable.files.sonoma.url' "${BREW_LLVM_FORMULA_JSON}")
+BREW_LLVM_ARM64_BOTTLE_URL=$(jq -r '.bottle.stable.files.arm64_sonoma.url' "${BREW_LLVM_FORMULA_JSON}")
+curl -fsSL -H "Authorization: Bearer QQ==" "${BREW_LLVM_AMD64_BOTTLE_URL}" | tar -xzf - --strip-components=2 -C "${DARWIN_AMD64_LLVM_PREFIX}"
+curl -fsSL -H "Authorization: Bearer QQ==" "${BREW_LLVM_ARM64_BOTTLE_URL}" | tar -xzf - --strip-components=2 -C "${DARWIN_ARM64_LLVM_PREFIX}"
+
+patch_homebrew_lib_dir() {
+	local LIB_DIR="$1"
+	local HOMEBREW_PREFIX="$2"
+	for DYLIB_FILE in "${LIB_DIR}"/*.dylib; do
+		if [[ -f "${DYLIB_FILE}" ]]; then
+			ID=$(otool -D "${DYLIB_FILE}" | grep '@@HOMEBREW_PREFIX@@' | awk '{print $1}')
+			if [[ -n "${ID}" ]]; then
+				NEW_ID=${ID/'@@HOMEBREW_PREFIX@@'/${HOMEBREW_PREFIX}}
+				install_name_tool -id "${NEW_ID}" "${DYLIB_FILE}"
+			fi
+
+			DEPS=$(otool -L "${DYLIB_FILE}" | grep '@@HOMEBREW_PREFIX@@' | awk '{print $1}')
+			for DEP in ${DEPS}; do
+				NEW_DEP=${DEP/'@@HOMEBREW_PREFIX@@'/${HOMEBREW_PREFIX}}
+				install_name_tool -change "${DEP}" "${NEW_DEP}" "${DYLIB_FILE}"
+			done
+		fi
+	done
+}
+patch_homebrew_lib_dir "${DARWIN_AMD64_LLVM_PREFIX}/lib" /usr/lib
+patch_homebrew_lib_dir "${DARWIN_ARM64_LLVM_PREFIX}/lib" /opt/homebrew

.github/workflows/populate_linux_sysroot.sh

@@ -0,0 +1,143 @@
+#!/bin/bash
+
+set -e
+
+TMPDIR="$(mktemp -d)"
+export TMPDIR
+trap 'rm -rf "${TMPDIR}"' EXIT
+
+LINUX_AMD64_PREFIX=.sysroot/linux/amd64
+LINUX_ARM64_PREFIX=.sysroot/linux/arm64
+mkdir -p "${LINUX_AMD64_PREFIX}" "${LINUX_ARM64_PREFIX}"
+
+POPULATE_LINUX_SYSROOT_SCRIPT="$(mktemp)"
+cat > "${POPULATE_LINUX_SYSROOT_SCRIPT}" << EOF
+#!/bin/bash
+
+export DEBIAN_FRONTEND=noninteractive
+
+apt-get update
+apt-get install -y lsb-release gnupg2 wget rsync
+
+echo "deb http://apt.llvm.org/\$(lsb_release -cs)/ llvm-toolchain-\$(lsb_release -cs)-18 main" | tee /etc/apt/sources.list.d/llvm.list
+wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key | apt-key add -
+apt-get update
+apt-get install -y llvm-18-dev
+
+error() {
+	echo -e "\$1" >&2
+	exit 1
+}
+
+exclude_list=()
+include_list=()
+
+exclude_list+=(--exclude "/bin")
+exclude_list+=(--exclude "/boot")
+exclude_list+=(--exclude "/boot*")
+exclude_list+=(--exclude "/dev")
+exclude_list+=(--exclude "/etc")
+exclude_list+=(--exclude "/home")
+exclude_list+=(--exclude "/lib/dhcpd")
+exclude_list+=(--exclude "/lib/firmware")
+exclude_list+=(--exclude "/lib/hdparm")
+exclude_list+=(--exclude "/lib/ifupdown")
+exclude_list+=(--exclude "/lib/modules")
+exclude_list+=(--exclude "/lib/modprobe.d")
+exclude_list+=(--exclude "/lib/modules-load.d")
+exclude_list+=(--exclude "/lib/resolvconf")
+exclude_list+=(--exclude "/lib/startpar")
+exclude_list+=(--exclude "/lib/systemd")
+exclude_list+=(--exclude "/lib/terminfo")
+exclude_list+=(--exclude "/lib/udev")
+exclude_list+=(--exclude "/lib/xtables")
+exclude_list+=(--exclude "/lib/ssl/private")
+exclude_list+=(--exclude "/lost+found")
+exclude_list+=(--exclude "/media")
+exclude_list+=(--exclude "/mnt")
+exclude_list+=(--exclude "/proc")
+exclude_list+=(--exclude "/root")
+exclude_list+=(--exclude "/run")
+exclude_list+=(--exclude "/sbin")
+exclude_list+=(--exclude "/srv")
+exclude_list+=(--exclude "/sys")
+exclude_list+=(--exclude "/tmp")
+exclude_list+=(--exclude "/usr/bin")
+exclude_list+=(--exclude "/usr/games")
+exclude_list+=(--exclude "/usr/sbin")
+exclude_list+=(--exclude "/usr/share")
+exclude_list+=(--exclude "/usr/src")
+exclude_list+=(--exclude "/usr/local/bin")
+exclude_list+=(--exclude "/usr/local/etc")
+exclude_list+=(--exclude "/usr/local/games")
+exclude_list+=(--exclude "/usr/local/man")
+exclude_list+=(--exclude "/usr/local/sbin")
+exclude_list+=(--exclude "/usr/local/share")
+exclude_list+=(--exclude "/usr/local/src")
+exclude_list+=(--exclude "/usr/lib/ssl/private")
+exclude_list+=(--exclude "/var")
+exclude_list+=(--exclude "/snap")
+exclude_list+=(--exclude "*python*")
+
+include_list+=(--include "*.a")
+include_list+=(--include "*.so")
+include_list+=(--include "*.so.*")
+include_list+=(--include "*.h")
+include_list+=(--include "*.hh")
+include_list+=(--include "*.hpp")
+include_list+=(--include "*.hxx")
+include_list+=(--include "*.pc")
+include_list+=(--include "*.def")
+include_list+=(--include "*.inc")
+include_list+=(--include "/lib")
+include_list+=(--include "/lib32")
+include_list+=(--include "/lib64")
+include_list+=(--include "/libx32")
+include_list+=(--include "*/")
+
+do-sync() {
+	from=\$1
+	to=\$2
+
+	args=()
+	args+=(-a)
+	args+=(-z)
+	args+=(-m)
+	args+=(-d)
+	args+=(-h)
+	args+=(--keep-dirlinks)
+	args+=("--info=progress2")
+	args+=(--delete)
+	args+=(--prune-empty-dirs)
+	args+=(--sparse)
+	args+=(--links)
+	args+=(--copy-unsafe-links)
+	args+=("\${exclude_list[@]}")
+	args+=("\${include_list[@]}")
+	args+=(--exclude "*")
+	args+=("\${from}")
+	args+=("\${to}")
+
+	echo "\${args[@]}"
+	rsync "\${args[@]}"
+
+	exit \$?
+}
+
+do-sync / /sysroot/
+EOF
+chmod +x "${POPULATE_LINUX_SYSROOT_SCRIPT}"
+
+populate_linux_sysroot() {
+	local ARCH="$1"
+	local PREFIX="$2"
+	docker run \
+		--rm \
+		--platform "linux/${ARCH}" \
+		-v "$(pwd)/${PREFIX}":/sysroot \
+		-v "${POPULATE_LINUX_SYSROOT_SCRIPT}":/populate_linux_sysroot.sh \
+		debian:bullseye \
+		/populate_linux_sysroot.sh
+}
+populate_linux_sysroot amd64 "${LINUX_AMD64_PREFIX}"
+populate_linux_sysroot arm64 "${LINUX_ARM64_PREFIX}"

.github/workflows/release-build.yml

@@ -0,0 +1,58 @@
+name: Release Build
+
+on:
+  push:
+    tags:
+      - "*"
+
+jobs:
+  populate-darwin-sysroot:
+    runs-on: macos-latest
+    steps:
+      - name: Check out code
+        uses: actions/checkout@v4
+      - name: Populate Darwin sysroot
+        run: bash .github/workflows/populate_darwin_sysroot.sh
+      - name: Create Darwin sysroot tarball
+        run: tar -czvf .sysroot/darwin.tar.gz -C .sysroot darwin
+      - name: Upload Darwin sysroot tarball
+        uses: actions/upload-artifact@v4
+        with:
+          name: darwin-sysroot-tarball
+          path: .sysroot/darwin.tar.gz
+          compression-level: 0
+  build:
+    runs-on: ubuntu-latest
+    needs: populate-darwin-sysroot
+    steps:
+      - name: Check out code
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+      - name: Set up Go
+        uses: actions/setup-go@v5
+        with:
+          go-version: 1.20.x
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+      - name: Download Darwin sysroot tarball
+        uses: actions/download-artifact@v4
+        with:
+          name: darwin-sysroot-tarball
+          path: .sysroot
+      - name: Populate Darwin sysroot
+        run: tar -xzvf .sysroot/darwin.tar.gz -C .sysroot
+      - name: Populate Linux sysroot
+        run: bash .github/workflows/populate_linux_sysroot.sh
+      - name: Run GoReleaser
+        env:
+          GITHUB_TOKEN: ${{github.token}}
+        run: |
+          docker run \
+            --rm \
+            -e GITHUB_TOKEN=${GITHUB_TOKEN} \
+            -v /var/run/docker.sock:/var/run/docker.sock \
+            -v $(pwd):/go/src/llgo \
+            -w /go/src/llgo \
+            ghcr.io/goreleaser/goreleaser-cross:v1.22 \
+            release --clean --skip nfpm,snapcraft

.github/workflows/test_demo.sh

@@ -0,0 +1,29 @@
+#!/bin/bash
+
+# llgo run subdirectories under _demo and _pydemo
+total=0
+failed=0
+failed_cases=""
+for d in ./_demo/* ./_pydemo/*; do
+  total=$((total+1))
+  if [ -d "$d" ]; then
+    echo "Testing $d"
+    if ! llgo run "$d"; then
+      echo "FAIL"
+      failed=$((failed+1))
+      failed_cases="$failed_cases\n* :x: $d"
+    else
+      echo "PASS"
+    fi
+  fi
+done
+echo "=== Done"
+echo "$((total-failed))/$total tests passed"
+
+if [ "$failed" -ne 0 ]; then
+  echo ":bangbang: Failed demo cases:" | tee -a result.md
+  echo -e "$failed_cases" | tee -a result.md
+  exit 1
+else
+  echo ":white_check_mark: All demo tests passed" | tee -a result.md
+fi

.github/workflows/test_llgo.sh

@@ -0,0 +1,38 @@
+#!/bin/bash
+set -e
+
+export LLGOROOT=$PWD
+
+testcmd=/tmp/test
+llgo build -o $testcmd ./c/bdwgc/_test
+cases=$($testcmd)
+total=$(echo "$cases" | wc -l | tr -d ' ')
+failed=0
+failed_cases=""
+
+for idx in $(seq 1 $((total))); do
+  case=$(echo "$cases" | sed -n "${idx}p")
+  case_name=$(echo "$case" | cut -d',' -f2)
+  echo "=== Test case: $case_name"
+  set +e
+  out=$("$testcmd" "$((idx-1))" 2>&1)
+  exit_code=$?
+  set -e
+  if [ "${exit_code:-0}" -ne 0 ]; then
+    echo "failed: $out"
+    failed=$((failed+1))
+    failed_cases="$failed_cases\n* :x: $case_name"
+  else
+    echo "passed"
+  fi
+done
+echo "=== Done"
+echo "$((total-failed))/$total tests passed"
+
+if [ "$failed" -ne 0 ]; then
+  echo ":bangbang: Failed llgo cases:" | tee -a result.md
+  echo -e "$failed_cases" | tee -a result.md
+  exit 1
+else
+  echo ":white_check_mark: All llgo tests passed" | tee -a result.md
+fi

.gitignore

@@ -9,7 +9,8 @@
 *.dylib
 
 test.db
-llgo_autogen.ll
+demo.ll
+llgo_autogen*.ll
 stories*.bin
 .DS_Store
 err.log
@@ -18,6 +19,7 @@ numpy.txt
 _go/
 _runtime/
 _tinygo/
+_output/
 build.dir/
 .vscode/
 
@@ -33,3 +35,7 @@ build.dir/
 
 # Go workspace file
 go.work*
+
+# GoReleaser
+.dist/
+.sysroot/

.goreleaser.yaml

@@ -0,0 +1,145 @@
+version: 2
+
+dist: .dist
+
+env:
+  - SYSROOT_DARWIN_AMD64={{.Env.PWD}}/.sysroot/darwin/amd64
+  - SYSROOT_DARWIN_ARM64={{.Env.PWD}}/.sysroot/darwin/arm64
+  - SYSROOT_LINUX_AMD64={{.Env.PWD}}/.sysroot/linux/amd64
+  - SYSROOT_LINUX_ARM64={{.Env.PWD}}/.sysroot/linux/arm64
+  - CGO_ENABLED=1
+  - CGO_CXXFLAGS=-std=c++17
+
+before:
+  hooks:
+    - go mod download
+
+builds:
+  - id: llgo-darwin-amd64
+    main: ./cmd/llgo
+    flags:
+      - -tags=darwin,amd64,byollvm
+    ldflags:
+      - -X github.com/goplus/llgo/xtool/env.buildVersion=v{{.Version}}
+      - -X github.com/goplus/llgo/xtool/env.buildDate={{.Date}}
+      - -X github.com/goplus/llgo/xtool/env/llvm.ldLLVMConfigBin=/usr/local/opt/llvm@18/bin/llvm-config
+    env:
+      - CC=o64-clang
+      - CXX=o64-clang++
+      - CGO_CPPFLAGS=-I{{.Env.SYSROOT_DARWIN_AMD64}}/usr/local/opt/llvm@18/include -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS
+      - CGO_LDFLAGS=-L{{.Env.SYSROOT_DARWIN_AMD64}}/usr/local/opt/llvm@18/lib -Wl,-search_paths_first -Wl,-headerpad_max_install_names -lLLVM-18 -lz -lm
+    targets:
+      - darwin_amd64
+    mod_timestamp: "{{.CommitTimestamp}}"
+  - id: llgo-darwin-arm64
+    main: ./cmd/llgo
+    flags:
+      - -tags=darwin,arm64,byollvm
+    ldflags:
+      - -X github.com/goplus/llgo/xtool/env.buildVersion=v{{.Version}}
+      - -X github.com/goplus/llgo/xtool/env.buildDate={{.Date}}
+      - -X github.com/goplus/llgo/xtool/env/llvm.ldLLVMConfigBin=/opt/homebrew/opt/llvm@18/bin/llvm-config
+    env:
+      - CC=oa64-clang
+      - CXX=oa64-clang++
+      - CGO_CPPFLAGS=-I{{.Env.SYSROOT_DARWIN_ARM64}}/opt/homebrew/opt/llvm@18/include -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS
+      - CGO_LDFLAGS=-L{{.Env.SYSROOT_DARWIN_ARM64}}/opt/homebrew/opt/llvm@18/lib -Wl,-search_paths_first -Wl,-headerpad_max_install_names -lLLVM-18 -lz -lm
+    targets:
+      - darwin_arm64
+    mod_timestamp: "{{.CommitTimestamp}}"
+  - id: llgo-linux-amd64
+    main: ./cmd/llgo
+    flags:
+      - -tags=linux,amd64,byollvm
+    ldflags:
+      - -X github.com/goplus/llgo/xtool/env.buildVersion=v{{.Version}}
+      - -X github.com/goplus/llgo/xtool/env.buildDate={{.Date}}
+      - -X github.com/goplus/llgo/xtool/env/llvm.ldLLVMConfigBin=/usr/lib/llvm-18/bin/llvm-config
+    env:
+      - CC=x86_64-linux-gnu-gcc
+      - CXX=x86_64-linux-gnu-g++
+      - CGO_CPPFLAGS=--sysroot={{.Env.SYSROOT_LINUX_AMD64}} -I{{.Env.SYSROOT_LINUX_AMD64}}/usr/include/llvm-18 -I{{.Env.SYSROOT_LINUX_AMD64}}/usr/include/llvm-c-18 -D_GNU_SOURCE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS
+      - CGO_LDFLAGS=--sysroot={{.Env.SYSROOT_LINUX_AMD64}} -L{{.Env.SYSROOT_LINUX_AMD64}}/usr/lib/llvm-18/lib -lLLVM-18
+    targets:
+      - linux_amd64
+    mod_timestamp: "{{.CommitTimestamp}}"
+  - id: llgo-linux-arm64
+    main: ./cmd/llgo
+    flags:
+      - -tags=linux,arm64,byollvm
+    ldflags:
+      - -X github.com/goplus/llgo/xtool/env.buildVersion=v{{.Version}}
+      - -X github.com/goplus/llgo/xtool/env.buildDate={{.Date}}
+      - -X github.com/goplus/llgo/xtool/env/llvm.ldLLVMConfigBin=/usr/lib/llvm-18/bin/llvm-config
+    env:
+      - CC=aarch64-linux-gnu-gcc
+      - CXX=aarch64-linux-gnu-g++
+      - CGO_CPPFLAGS=--sysroot={{.Env.SYSROOT_LINUX_ARM64}} -I{{.Env.SYSROOT_LINUX_ARM64}}/usr/include/llvm-18 -I{{.Env.SYSROOT_LINUX_ARM64}}/usr/include/llvm-c-18 -D_GNU_SOURCE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS
+      - CGO_LDFLAGS=--sysroot={{.Env.SYSROOT_LINUX_ARM64}} -L{{.Env.SYSROOT_LINUX_ARM64}}/usr/lib/llvm-18/lib -lLLVM-18
+    targets:
+      - linux_arm64
+    mod_timestamp: "{{.CommitTimestamp}}"
+
+archives:
+  - format: tar.gz
+    name_template: >-
+      {{.ProjectName}}{{.Version}}.{{.Os}}-{{.Arch}}
+      {{- if .Arm}}v{{.Arm}}{{end}}
+    files:
+      - LICENSE
+      - README.md
+
+checksum:
+  name_template: "{{.ProjectName}}{{.Version}}.checksums.txt"
+
+nfpms:
+  - package_name: llgo
+    vendor: goplus
+    homepage: https://github.com/goplus/llgo
+    maintainer: Aofei Sheng <aofei@aofeisheng.com>
+    description: |
+      LLGo is a Go compiler based on LLVM in order to better integrate Go with the C ecosystem including Python. It's a
+      subproject of the Go+ project.
+
+      LLGo aims to expand the boundaries of Go/Go+, providing limitless possibilities such as:
+
+      - Game development
+      - AI and data science
+      - WebAssembly
+      - Embedded development
+      - ...
+    license: Apache-2.0
+    formats:
+      - deb
+      - rpm
+    file_name_template: >-
+      {{.ProjectName}}{{.Version}}.{{.Os}}-{{.Arch}}
+      {{- if .Arm}}v{{.Arm}}{{end}}
+    bindir: /usr/local/bin
+
+snapcrafts:
+  - name: llgo
+    title: A Go compiler based on LLVM
+    summary: A Go compiler based on LLVM
+    description: |
+      LLGo is a Go compiler based on LLVM in order to better integrate Go with the C ecosystem including Python. It's a
+      subproject of the Go+ project.
+
+      LLGo aims to expand the boundaries of Go/Go+, providing limitless possibilities such as:
+
+      - Game development
+      - AI and data science
+      - WebAssembly
+      - Embedded development
+      - ...
+    license: Apache-2.0
+    confinement: classic
+    name_template: >-
+      {{.ProjectName}}{{.Version}}.{{.Os}}-{{.Arch}}
+      {{- if .Arm}}v{{.Arm}}{{end}}
+
+snapshot:
+  name_template: '{{trimprefix .Summary "v"}}'
+
+release:
+  prerelease: auto

README.md

@@ -10,8 +10,42 @@ llgo - A Go compiler based on LLVM
 
 LLGo is a Go compiler based on LLVM in order to better integrate Go with the C ecosystem including Python. It's a subproject of [the Go+ project](https://github.com/goplus/gop).
 
+LLGo aims to expand the boundaries of Go/Go+, providing limitless possibilities such as:
 
-## C standard libary support
+* Game development
+* AI and data science
+* WebAssembly
+* Embedded development
+* ...
+
+How can these be achieved?
+
+```
+LLGo := Go + C + Python
+```
+
+LLGo is compatible with C and Python through the language's **Application Binary Interface (ABI)**, while LLGo is compatible with Go through its **syntax (source code)**.
+
+
+## C/C++ standard libary support
+
+You can import a C/C++ standard library in LLGo!
+
+* [c](https://pkg.go.dev/github.com/goplus/llgo/c)
+* [c/syscall](https://pkg.go.dev/github.com/goplus/llgo/c/syscall)
+* [c/sys](https://pkg.go.dev/github.com/goplus/llgo/c/sys)
+* [c/os](https://pkg.go.dev/github.com/goplus/llgo/c/os)
+* [c/math](https://pkg.go.dev/github.com/goplus/llgo/c/math)
+* [c/math/cmplx](https://pkg.go.dev/github.com/goplus/llgo/c/math/cmplx)
+* [c/math/rand](https://pkg.go.dev/github.com/goplus/llgo/c/math/rand)
+* [c/pthread](https://pkg.go.dev/github.com/goplus/llgo/c/pthread)
+* [c/pthread/sync](https://pkg.go.dev/github.com/goplus/llgo/c/pthread/sync)
+* [c/sync/atomic](https://pkg.go.dev/github.com/goplus/llgo/c/sync/atomic)
+* [c/time](https://pkg.go.dev/github.com/goplus/llgo/c/time)
+* [c/net](https://pkg.go.dev/github.com/goplus/llgo/c/net)
+* [cpp/std](https://pkg.go.dev/github.com/goplus/llgo/cpp/std)
+
+Here is a simple example:
 
 ```go
 package main
@@ -34,46 +68,98 @@ The `_demo` directory contains some C standard libary related demos (it start wi
 To run these demos (If you haven't installed `llgo` yet, please refer to [How to install](#how-to-install)):
 
 ```sh
-export LLGOROOT=`pwd`
 cd <demo-directory>  # eg. cd _demo/hello
 llgo run .
 ```
 
-See [github.com/goplus/llgo/c](https://pkg.go.dev/github.com/goplus/llgo/c) for more detials.
+
+## How support C/C++ and Python
+
+LLGo use `go:linkname` to link an extern symbol througth its ABI:
+
+```go
+import _ "unsafe" // for go:linkname
+
+//go:linkname Sqrt C.sqrt
+func Sqrt(x float64) float64
+```
+
+You can directly integrate it into [your own code](_demo/linkname/linkname.go):
+
+```go
+package main
+
+import _ "unsafe" // for go:linkname
+
+//go:linkname Sqrt C.sqrt
+func Sqrt(x float64) float64
+
+func main() {
+	println("sqrt(2) =", Sqrt(2))
+}
+```
+
+Or put it into a package (see [c/math](c/math/math.go)):
+
+```go
+package main
+
+import "github.com/goplus/llgo/c/math"
+
+func main() {
+	println("sqrt(2) =", math.Sqrt(2))
+}
+```
 
 
 ## Python support
 
 You can import a Python library in LLGo!
 
-And you can import any Python library into `llgo` through a program called `llpyg` (see [Development tools](#development-tools)). The currently imported libraries include:
+And you can import any Python library into `llgo` through a program called `llpyg` (see [Development tools](#development-tools)). The following libraries have been included in `llgo`:
 
-* [sys](https://pkg.go.dev/github.com/goplus/llgo/py/sys)
-* [os](https://pkg.go.dev/github.com/goplus/llgo/py/os)
-* [math](https://pkg.go.dev/github.com/goplus/llgo/py/math)
-* [json](https://pkg.go.dev/github.com/goplus/llgo/py/json)
-* [inspect](https://pkg.go.dev/github.com/goplus/llgo/py/inspect)
-* [statistics](https://pkg.go.dev/github.com/goplus/llgo/py/statistics)
-* [numpy](https://pkg.go.dev/github.com/goplus/llgo/py/numpy)
+* [py](https://pkg.go.dev/github.com/goplus/llgo/py) (abi)
+* [py/std](https://pkg.go.dev/github.com/goplus/llgo/py/std) (builtins)
+* [py/sys](https://pkg.go.dev/github.com/goplus/llgo/py/sys)
+* [py/os](https://pkg.go.dev/github.com/goplus/llgo/py/os)
+* [py/math](https://pkg.go.dev/github.com/goplus/llgo/py/math)
+* [py/json](https://pkg.go.dev/github.com/goplus/llgo/py/json)
+* [py/inspect](https://pkg.go.dev/github.com/goplus/llgo/py/inspect)
+* [py/statistics](https://pkg.go.dev/github.com/goplus/llgo/py/statistics)
+* [py/numpy](https://pkg.go.dev/github.com/goplus/llgo/py/numpy)
+* [py/pandas](https://pkg.go.dev/github.com/goplus/llgo/py/pandas)
+* [py/torch](https://pkg.go.dev/github.com/goplus/llgo/py/torch)
+* [py/matplotlib](https://pkg.go.dev/github.com/goplus/llgo/py/matplotlib)
 
-Here is an example using the Python `math` library:
+Note: For third-party libraries (such as pandas and pytorch), you still need to install the library files.
+
+Here is an example:
 
 ```go
 package main
 
 import (
-	"github.com/goplus/llgo/c"
 	"github.com/goplus/llgo/py"
 	"github.com/goplus/llgo/py/math"
+	"github.com/goplus/llgo/py/std"
 )
 
 func main() {
-	x := math.Sqrt(py.Float(2))
-	c.Printf(c.Str("sqrt(2) = %f\n"), x.Float64())
+	x := math.Sqrt(py.Float(2))       // x = sqrt(2)
+	std.Print(py.Str("sqrt(2) ="), x) // print("sqrt(2) =", x)
 }
 ```
 
-Here, We call `py.Float(2)` to create a Python number 2, and pass it to Python’s `math.sqrt` to get `x`. Then use `x.Float64()` to convert x to Go's `float64` type, and print the value through the C `printf` function.
+It is equivalent to the following Python code:
+
+```py
+import math
+
+x = math.sqrt(2)
+print("sqrt =", x)
+```
+
+Here, We call `py.Float(2)` to create a Python number 2, and pass it to Python’s `math.sqrt` to get `x`. Then we call `std.Print` to print the result.
 
 Let's look at a slightly more complex example. For example, we use `numpy` to calculate:
 
@@ -81,9 +167,9 @@ Let's look at a slightly more complex example. For example, we use `numpy` to ca
 package main
 
 import (
-	"github.com/goplus/llgo/c"
 	"github.com/goplus/llgo/py"
 	"github.com/goplus/llgo/py/numpy"
+	"github.com/goplus/llgo/py/std"
 )
 
 func main() {
@@ -98,7 +184,7 @@ func main() {
 		py.List(3.0, 2.0, 1.0),
 	)
 	x := numpy.Add(a, b)
-	c.Printf(c.Str("a+b = %s\n"), x.Str().CStr())
+	std.Print(py.Str("a+b ="), x)
 }
 ```
 
@@ -111,83 +197,111 @@ The `_pydemo` directory contains some python related demos:
 * [statistics](_pydemo/statistics/statistics.go): define a python list and call `statistics.mean` to get the mean
 * [matrix](_pydemo/matrix/matrix.go): a basic `numpy` demo
 
-To run these demos, you need to set the `LLGO_LIB_PYTHON` environment variable first.
-
-If Python is in the search path for `clang` linking, then `LLGO_LIB_PYTHON` only needs to be set to the name of the Python library. For example:
+To run these demos (If you haven't installed `llgo` yet, please refer to [How to install](#how-to-install)):
 
 ```sh
-export LLGO_LIB_PYTHON=python3.12
-```
-
-You can also specify the path to tell `llgo` where the Python library is located:
-
-```sh
-export LLGO_LIB_PYTHON=/foo/bar/python3.12
-```
-
-For example, `/opt/homebrew/Frameworks/Python.framework/Versions/3.12/libpython3.12.dylib` is a typical python library location under macOS. So we should set it like this:
-
-```sh
-export LLGO_LIB_PYTHON=/opt/homebrew/Frameworks/Python.framework/Versions/3.12/python3.12
-```
-
-Note that the file name must be written in a platform-independent format, using `python3.12` instead of `libpython3.12.dylib`.
-
-Then you can run the demos:
-
-```sh
-export LLGOROOT=`pwd`
 cd <demo-directory>  # eg. cd _pydemo/callpy
 llgo run .
 ```
 
-See [github.com/goplus/llgo/py](https://pkg.go.dev/github.com/goplus/llgo/py) for more detials.
-
 
 ## Other frequently used libraries
 
 LLGo can easily import any libraries from the C ecosystem. Currently, this import process is still manual, but in the future, it will be automated similar to Python library imports.
 
-The currently imported libraries include:
+The currently supported libraries include:
 
-* [llama2.c](https://pkg.go.dev/github.com/goplus/llgo/c/llama2)
-* [cjson](https://pkg.go.dev/github.com/goplus/llgo/c/cjson)
-* [sqlite](https://pkg.go.dev/github.com/goplus/llgo/c/sqlite)
+* [c/bdwgc](https://pkg.go.dev/github.com/goplus/llgo/c/bdwgc)
+* [c/cjson](https://pkg.go.dev/github.com/goplus/llgo/c/cjson)
+* [c/clang](https://pkg.go.dev/github.com/goplus/llgo/c/clang)
+* [c/libuv](https://pkg.go.dev/github.com/goplus/llgo/c/libuv)
+* [c/llama2](https://pkg.go.dev/github.com/goplus/llgo/c/llama2)
+* [c/lua](https://pkg.go.dev/github.com/goplus/llgo/c/lua)
+* [c/neco](https://pkg.go.dev/github.com/goplus/llgo/c/neco)
+* [c/raylib](https://pkg.go.dev/github.com/goplus/llgo/c/raylib)
+* [c/sqlite](https://pkg.go.dev/github.com/goplus/llgo/c/sqlite)
+* [c/zlib](https://pkg.go.dev/github.com/goplus/llgo/c/zlib)
+* [cpp/inih](https://pkg.go.dev/github.com/goplus/llgo/cpp/inih)
+* [cpp/llvm](https://pkg.go.dev/github.com/goplus/llgo/cpp/llvm)
 
 Here are some examples related to them:
 
 * [llama2-c](_demo/llama2-c): inference Llama 2 (It's the first llgo AI example)
 * [mkjson](c/cjson/_demo/mkjson/mkjson.go): create a json object and print it
 * [sqlitedemo](c/sqlite/_demo/sqlitedemo/demo.go): a basic sqlite demo
+* [tetris](c/raylib/_demo/tetris/tetris.go): a tetris game based on raylib
 
 
 ## Go syntax support
 
-The priority of `llgo` feature iteration is:
-
-* Popular C/Python libraries
-* Full Go syntax
-* Go standard libraries
-* Popular Go packages
-
-Common Go syntax is already supported. Except for the following, which needs to be improved:
-
-* interface (Limited support)
-* map (Very limited support)
-* panic (Limited support)
-* recover (Not supported yet)
-* defer (Not supported yet)
-* gc (Not supported yet)
-* chan (Not supported yet)
-* goroutine (Not supported yet)
-* generics (Not supported yet)
-
-Here are some examples related to Go syntax:
+All Go syntax (not including `cgo`) is already supported. Here are some examples:
 
 * [concat](_demo/concat/concat.go): define a variadic function
 * [genints](_demo/genints/genints.go): various forms of closure usage (including C function, recv.method and anonymous function)
+* [errors](_cmptest/errors/errors.go): demo to implement error interface
+* [defer](_cmptest/defer/defer.go): defer demo
+* [goroutine](_demo/goroutine/goroutine.go): goroutine demo
 
 
+## Defer
+
+LLGo `defer` does not support usage in loops. This is not a bug but a feature, because we think that using `defer` in a loop is a very unrecommended practice.
+
+
+### Garbage Collection (GC)
+
+By default, LLGo implements `gc` based on [bdwgc](https://www.hboehm.info/gc/) (also known as [libgc](https://www.hboehm.info/gc/)).
+
+However, you can disable gc by specifying the `nogc` tag. For example:
+
+```sh
+llgo run -tags nogc .
+```
+
+
+## Go packages support
+
+Here are the Go packages that can be imported correctly:
+
+* [unsafe](https://pkg.go.dev/unsafe)
+* [unicode](https://pkg.go.dev/unicode)
+* [unicode/utf8](https://pkg.go.dev/unicode/utf8)
+* [unicode/utf16](https://pkg.go.dev/unicode/utf16)
+* [math](https://pkg.go.dev/math)
+* [math/bits](https://pkg.go.dev/math/bits)
+* [math/cmplx](https://pkg.go.dev/math/cmplx)
+* [errors](https://pkg.go.dev/errors)
+* [context](https://pkg.go.dev/context)
+* [io](https://pkg.go.dev/io)
+* [io/fs](https://pkg.go.dev/io/fs)
+* [log](https://pkg.go.dev/log)
+* [flag](https://pkg.go.dev/flag)
+* [sort](https://pkg.go.dev/sort)
+* [strconv](https://pkg.go.dev/strconv)
+* [strings](https://pkg.go.dev/strings)
+* [sync/atomic](https://pkg.go.dev/sync/atomic)
+* [sync](https://pkg.go.dev/sync) (partially)
+* [syscall](https://pkg.go.dev/syscall) (partially)
+* [runtime](https://pkg.go.dev/runtime) (partially)
+* [os](https://pkg.go.dev/os) (partially)
+* [os/exec](https://pkg.go.dev/os/exec) (partially)
+* [fmt](https://pkg.go.dev/fmt) (partially)
+* [reflect](https://pkg.go.dev/reflect) (partially)
+* [time](https://pkg.go.dev/time) (partially)
+
+
+## Dependencies
+
+- [Go 1.20+](https://go.dev) (build only)
+- [LLVM 18](https://llvm.org)
+- [LLD 18](https://lld.llvm.org)
+- [Clang 18](https://clang.llvm.org)
+- [pkg-config 0.29+](https://www.freedesktop.org/wiki/Software/pkg-config/)
+- [bdwgc/libgc 8.0+](https://www.hboehm.info/gc/)
+- [cJSON 1.7+](https://github.com/DaveGamble/cJSON) (optional, for [github.com/goplus/llgo/c/cjson](https://pkg.go.dev/github.com/goplus/llgo/c/cjson))
+- [SQLite 3](https://www.sqlite.org) (optional, for [github.com/goplus/llgo/c/sqlite](https://pkg.go.dev/github.com/goplus/llgo/c/sqlite))
+- [Python 3.11+](https://www.python.org) (optional, for [github.com/goplus/llgo/py](https://pkg.go.dev/github.com/goplus/llgo/py))
+
 ## How to install
 
 Follow these steps to generate the `llgo` command (its usage is the same as the `go` command):
@@ -195,19 +309,23 @@ Follow these steps to generate the `llgo` command (its usage is the same as the
 ### on macOS
 
 ```sh
-brew update  # execute if needed
-brew install llvm@17
-go install -v ./...
+brew update # execute if needed
+brew install llvm@18 pkg-config libgc
+brew install cjson sqlite python@3.12 # optional
+export PATH=$(brew --prefix llvm@18)/bin:$PATH # you may want to add this to your shell RC file, e.g. ~/.zshrc
+go install -v github.com/goplus/llgo/cmd/llgo@latest
 ```
 
-### on Linux
+### on Linux (Debian/Ubuntu)
 
 ```sh
-echo 'deb http://apt.llvm.org/focal/ llvm-toolchain-focal-17 main' | sudo tee /etc/apt/sources.list.d/llvm.list
+echo "deb http://apt.llvm.org/$(lsb_release -cs)/ llvm-toolchain-$(lsb_release -cs)-18 main" | sudo tee /etc/apt/sources.list.d/llvm.list
 wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key | sudo apt-key add -
-sudo apt-get update  # execute if needed
-sudo apt-get install --no-install-recommends llvm-17-dev
-go install -v ./...
+sudo apt-get update # execute if needed
+sudo apt-get install -y llvm-18-dev clang-18 lld-18 pkg-config libgc-dev
+sudo apt-get install -y libcjson-dev libsqlite3-dev python3.12-dev # optional
+export PATH=/usr/lib/llvm-18/bin:$PATH # you may want to add this to your shell RC file, e.g. ~/.bashrc
+go install -v github.com/goplus/llgo/cmd/llgo@latest
 ```
 
 ### on Windows
@@ -218,16 +336,20 @@ TODO
 ## Development tools
 
 * [pydump](chore/_xtool/pydump): It's the first program compiled by `llgo` (NOT `go`) in a production environment. It outputs symbol information (functions, variables, and constants) from a Python library in JSON format, preparing for the generation of corresponding packages in `llgo`.
-* [llpyg](chore/llpyg): It is used to automatically convert Python libraries into Go packages that `llgo` can import. It depends on `pydump` to accomplish the task.
+* [pysigfetch](https://github.com/goplus/hdq/tree/main/chore/pysigfetch): It generates symbol information by extracting information from Python's documentation site. This tool is not part of the `llgo` project, but we depend on it.
+* [llpyg](chore/llpyg): It is used to automatically convert Python libraries into Go packages that `llgo` can import. It depends on `pydump` and `pysigfetch` to accomplish the task.
 * [llgen](chore/llgen): It is used to compile Go packages into LLVM IR files (*.ll).
 * [ssadump](chore/ssadump): It is a Go SSA builder and interpreter.
 
 How do I generate these tools?
 
 ```sh
-go install -v ./...  # compile all tools except pydump
+git clone https://github.com/goplus/llgo.git
+cd llgo
+go install -v ./chore/...  # compile all tools except pydump
 cd chore/_xtool
 llgo install ./...   # compile pydump
+go install github.com/goplus/hdq/chore/pysigfetch@v0.8.1  # compile pysigfetch
 ```
 
 ## Key modules

_cmptest/_chandemo/chan.go

@@ -0,0 +1,21 @@
+package main
+
+func main() {
+	c1 := make(chan string)
+	c2 := make(chan string, 1)
+	go func() {
+		c1 <- "ch1"
+	}()
+	go func() {
+		c2 <- "ch2"
+	}()
+
+	for i := 0; i < 2; i++ {
+		select {
+		case msg1 := <-c1:
+			println(msg1)
+		case msg2 := <-c2:
+			println(msg2)
+		}
+	}
+}

_cmptest/_osexec/exec.go

@@ -0,0 +1,18 @@
+package main
+
+import (
+	"os"
+	"os/exec"
+	"runtime"
+)
+
+func main() {
+	ls := "ls"
+	if runtime.GOOS == "windows" {
+		ls = "dir"
+	}
+	cmd := exec.Command(ls)
+	cmd.Stdout = os.Stdout
+	cmd.Stderr = os.Stderr
+	cmd.Run()
+}

_cmptest/_pipedemo/pipe.go

@@ -0,0 +1,26 @@
+package main
+
+import (
+	"fmt"
+	"io"
+)
+
+func main() {
+	data := []byte("This is some data that needs to be stored in Body.")
+	pr, pw := io.Pipe()
+	go func() {
+		defer pw.Close()
+		if _, err := pw.Write(data); err != nil {
+			fmt.Println("Error writing to pipe:", err)
+			return
+		}
+	}()
+	defer pr.Close()
+
+	readData, err := io.ReadAll(pr)
+	if err != nil {
+		fmt.Println("Error reading from Body:", err)
+		return
+	}
+	fmt.Println("Body:", string(readData))
+}

_cmptest/_timeout/timer.go

@@ -0,0 +1,19 @@
+package main
+
+import (
+	"fmt"
+	"time"
+)
+
+var c chan int
+
+func handle(int) {}
+
+func main() {
+	select {
+	case m := <-c:
+		handle(m)
+	case <-time.After(10 * time.Second):
+		fmt.Println("timed out")
+	}
+}

_cmptest/chan/chan.go

@@ -0,0 +1,11 @@
+package main
+
+func main() {
+	ch := make(chan int, 10)
+	println(len(ch), cap(ch))
+	go func() {
+		ch <- 100
+	}()
+	n, ok := <-ch
+	println(n, ok)
+}

_cmptest/chansel/chansel.go

@@ -0,0 +1,26 @@
+package main
+
+func fibonacci(c, quit chan int) {
+	x, y := 0, 1
+	for {
+		select {
+		case c <- x:
+			x, y = y, x+y
+		case <-quit:
+			println("quit")
+			return
+		}
+	}
+}
+
+func main() {
+	c := make(chan int)
+	quit := make(chan int)
+	go func() {
+		for i := 0; i < 10; i++ {
+			println(<-c)
+		}
+		close(quit)
+	}()
+	fibonacci(c, quit)
+}

_cmptest/ctxcancel/ctx.go

@@ -0,0 +1,39 @@
+package main
+
+import (
+	"context"
+	"fmt"
+)
+
+func main() {
+	// gen generates integers in a separate goroutine and
+	// sends them to the returned channel.
+	// The callers of gen need to cancel the context once
+	// they are done consuming generated integers not to leak
+	// the internal goroutine started by gen.
+	gen := func(ctx context.Context) <-chan int {
+		dst := make(chan int)
+		n := 1
+		go func() {
+			for {
+				select {
+				case <-ctx.Done():
+					return // returning not to leak the goroutine
+				case dst <- n:
+					n++
+				}
+			}
+		}()
+		return dst
+	}
+
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel() // cancel when we are finished consuming integers
+
+	for n := range gen(ctx) {
+		fmt.Println(n)
+		if n == 5 {
+			break
+		}
+	}
+}

_cmptest/defer/defer.go

@@ -0,0 +1,17 @@
+package main
+
+func f(s string) bool {
+	return len(s) > 2
+}
+
+func main() {
+	defer func() {
+		println("hi")
+	}()
+	if s := "hello"; f(s) {
+		defer println(s)
+	} else {
+		defer println("world")
+	}
+	defer println("bye")
+}

_cmptest/errors/errors.go

@@ -0,0 +1,21 @@
+package main
+
+// New returns an error that formats as the given text.
+// Each call to New returns a distinct error value even if the text is identical.
+func New(text string) error {
+	return &errorString{text}
+}
+
+// errorString is a trivial implementation of error.
+type errorString struct {
+	s string
+}
+
+func (e *errorString) Error() string {
+	return e.s
+}
+
+func main() {
+	err := New("an error")
+	println(err.Error())
+}

_cmptest/flagdemo/flagdemo.go

@@ -0,0 +1,21 @@
+package main
+
+import (
+	"flag"
+	"fmt"
+	"os"
+)
+
+func main() {
+	fmt.Println("args:", os.Args[1:])
+	if len(os.Args) == 1 {
+		os.Args = []string{"flagdemo", "-cpu", "100"}
+	}
+
+	verbose := flag.Bool("v", false, "verbose")
+	cpu := flag.Int("cpu", 1, "cpu number")
+	host := flag.String("host", ":8888", "host")
+	flag.Parse()
+
+	fmt.Println("host:", *host, "cpu:", *cpu, "verbose:", *verbose)
+}

_cmptest/fmtdemo/fmt.go

@@ -0,0 +1,12 @@
+package main
+
+import (
+	"errors"
+	"fmt"
+)
+
+func main() {
+	fmt.Println("Hello, world")
+	fmt.Printf("%f\n", 3.14)
+	fmt.Printf("%v\n", errors.New("error message"))
+}

_cmptest/interf/foo/foo.go

@@ -0,0 +1,9 @@
+package foo
+
+func Bar() any {
+	return struct{ V int }{1}
+}
+
+func F() any {
+	return struct{ v int }{1}
+}

_cmptest/interf/interf.go

@@ -0,0 +1,35 @@
+package main
+
+import (
+	"github.com/goplus/llgo/_cmptest/interf/foo"
+)
+
+func Foo() any {
+	return struct{ v int }{1}
+}
+
+func main() {
+	v := Foo()
+	if x, ok := v.(struct{ v int }); ok {
+		println(x.v)
+	} else {
+		println("Foo: not ok")
+	}
+	bar := foo.Bar()
+	if x, ok := bar.(struct{ V int }); ok {
+		println(x.V)
+	} else {
+		println("Bar: not ok")
+	}
+	if x, ok := foo.F().(struct{ v int }); ok {
+		println(x.v)
+	} else {
+		println("F: not ok")
+	}
+}
+
+/* Expected output:
+1
+1
+F: not ok
+*/

_cmptest/iodemo/io.go

@@ -0,0 +1,14 @@
+package main
+
+import (
+	"io"
+	"os"
+)
+
+func f(w io.Writer) {
+	w.Write([]byte("Hello, world\n"))
+}
+
+func main() {
+	f(os.Stdout)
+}

_cmptest/osdemo/osd.go

@@ -0,0 +1,18 @@
+package main
+
+import (
+	"os"
+)
+
+func main() {
+	home, _ := os.UserHomeDir()
+	println("home:", home)
+
+	cfgdir, _ := os.UserConfigDir()
+	println("cfgdir:", cfgdir)
+
+	cache, _ := os.UserCacheDir()
+	println("cachedir:", cache)
+
+	os.Stdout.Write([]byte("Hello, World\n"))
+}

_cmptest/osproc/exec.go

@@ -0,0 +1,30 @@
+package main
+
+import (
+	"fmt"
+	"os"
+	"os/exec"
+	"runtime"
+)
+
+func main() {
+	ls := "ls"
+	args := []string{ls, "-l"}
+	if runtime.GOOS == "windows" {
+		ls = "dir"
+		args = []string{ls}
+	}
+	lspath, _ := exec.LookPath(ls)
+	if lspath != "" {
+		ls = lspath
+	}
+	proc, err := os.StartProcess(ls, args, &os.ProcAttr{
+		Files: []*os.File{nil, os.Stdout, os.Stderr},
+	})
+	if err != nil {
+		fmt.Println("os.StartProcess error:", err)
+		return
+	}
+	proc.Wait()
+	fmt.Println("proc.Wait done")
+}

_cmptest/reflect/reflect.go

@@ -0,0 +1,13 @@
+package main
+
+import "reflect"
+
+func main() {
+	tyIntSlice := reflect.SliceOf(reflect.TypeOf(0))
+	v := reflect.Zero(tyIntSlice)
+	v = reflect.Append(v, reflect.ValueOf(1), reflect.ValueOf(2), reflect.ValueOf(3))
+	for i, n := 0, v.Len(); i < n; i++ {
+		item := v.Index(i)
+		println(item.Int())
+	}
+}

_cmptest/rtype/rtype.go

@@ -0,0 +1,14 @@
+package main
+
+import "reflect"
+
+func main() {
+	tyIntSlice := reflect.SliceOf(reflect.TypeOf(0))
+	println(tyIntSlice.String())
+
+	v := reflect.Zero(tyIntSlice)
+	println(v.Len())
+
+	v = reflect.ValueOf(100)
+	println(v.Int())
+}

_cmptest/sortdemo/sort.go

@@ -0,0 +1,23 @@
+package main
+
+import "sort"
+
+func main() {
+	vals := []int{32, 58, 25, 92, 45, 78}
+	sort.Ints(vals)
+	for _, v := range vals {
+		println(v)
+	}
+
+	texts := []string{"apple", "banana", "cherry", "date", "elderberry", "fig"}
+	sort.Slice(texts, func(i, j int) bool {
+		leni, lenj := len(texts[i]), len(texts[j])
+		if leni != lenj {
+			return leni < lenj
+		}
+		return texts[i] < texts[j]
+	})
+	for _, v := range texts {
+		println(v)
+	}
+}

_cmptest/strconv/strconv.go

@@ -0,0 +1,12 @@
+package main
+
+import (
+	"fmt"
+	"strconv"
+	"strings"
+)
+
+func main() {
+	fmt.Println(strconv.Itoa(-123))
+	fmt.Println(strings.Split("abc,def,123", ","))
+}

_cmptest/syscall/syscall.go

@@ -0,0 +1,11 @@
+package main
+
+import "syscall"
+
+func main() {
+	wd, err := syscall.Getwd()
+	if err != nil {
+		panic(err)
+	}
+	println("cwd:", wd)
+}

_cmptest/timedemo/time.go

@@ -0,0 +1,8 @@
+package main
+
+import "time"
+
+func main() {
+	t := time.Date(2018, time.January, 1, 2, 3, 4, 5, time.UTC)
+	println(t.String())
+}

_demo/catomic/atomic.go

@@ -0,0 +1,24 @@
+package main
+
+import (
+	"github.com/goplus/llgo/c/sync/atomic"
+)
+
+func main() {
+	var v int64
+
+	atomic.Store(&v, 100)
+	println("store:", atomic.Load(&v))
+
+	ret := atomic.Add(&v, 1)
+	println("ret:", ret, "v:", v)
+
+	ret, _ = atomic.CompareAndExchange(&v, 100, 102)
+	println("ret:", ret, "vs 100, v:", v)
+
+	ret, _ = atomic.CompareAndExchange(&v, 101, 102)
+	println("ret:", ret, "vs 101, v:", v)
+
+	ret = atomic.Sub(&v, 1)
+	println("ret:", ret, "v:", v)
+}

_demo/cchan/cchan.go

@@ -0,0 +1,50 @@
+package main
+
+import (
+	"unsafe"
+
+	"github.com/goplus/llgo/internal/runtime"
+)
+
+const (
+	eltSize = int(unsafe.Sizeof(0))
+)
+
+func doChan(cap int) {
+	c := runtime.NewChan(eltSize, cap)
+	go func() {
+		v := 100
+		runtime.ChanSend(c, unsafe.Pointer(&v), eltSize)
+	}()
+	var ret int
+	runtime.ChanRecv(c, unsafe.Pointer(&ret), eltSize)
+	println(ret)
+}
+
+func main() {
+	doChan(10)
+	doChan(0)
+
+	c := runtime.NewChan(eltSize, 3)
+
+	v := 1
+	runtime.ChanSend(c, unsafe.Pointer(&v), eltSize)
+	v = 2
+	runtime.ChanSend(c, unsafe.Pointer(&v), eltSize)
+	v = 3
+	runtime.ChanSend(c, unsafe.Pointer(&v), eltSize)
+	runtime.ChanClose(c)
+
+	v = 10
+
+	if runtime.ChanTrySend(c, unsafe.Pointer(&v), eltSize) {
+		println("error: chan send to closed chan")
+	}
+
+	for {
+		if ok := runtime.ChanRecv(c, unsafe.Pointer(&v), eltSize); !ok {
+			break
+		}
+		println(v)
+	}
+}

_demo/cchansel/cchansel.go

@@ -0,0 +1,41 @@
+package main
+
+import (
+	"unsafe"
+
+	"github.com/goplus/llgo/internal/runtime"
+)
+
+const (
+	eltSize = int(unsafe.Sizeof(0))
+)
+
+func fibonacci(c, quit *runtime.Chan) {
+	x, y := 0, 1
+	for {
+		isel, _ := runtime.Select(
+			runtime.ChanOp{C: c, Send: true, Val: unsafe.Pointer(&x), Size: int32(eltSize)},
+			runtime.ChanOp{C: quit},
+		)
+		if isel == 0 {
+			x, y = y, x+y
+		} else {
+			println("quit")
+			return
+		}
+	}
+}
+
+func main() {
+	c := runtime.NewChan(eltSize, 0)
+	quit := runtime.NewChan(eltSize, 0)
+	go func() {
+		for i := 0; i < 10; i++ {
+			val := 0
+			runtime.ChanRecv(c, unsafe.Pointer(&val), eltSize)
+			println(val)
+		}
+		runtime.ChanClose(quit)
+	}()
+	fibonacci(c, quit)
+}

_demo/cexec/exec.go

@@ -0,0 +1,11 @@
+package main
+
+import (
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/os"
+)
+
+func main() {
+	ls := c.Str("ls")
+	os.Execlp(ls, ls, c.Str("-l"), nil)
+}

_demo/complex/cmplx.go

@@ -0,0 +1,9 @@
+package main
+
+import (
+	"math/cmplx"
+)
+
+func main() {
+	println("abs(3+4i):", cmplx.Abs(3+4i))
+}

_demo/cppintf/cppintf.go

@@ -0,0 +1,38 @@
+package main
+
+import (
+	"github.com/goplus/llgo/_demo/cppintf/foo"
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/math"
+)
+
+type Bar struct {
+	foo.Callback
+	a c.Int
+}
+
+func NewBar(a c.Int) *Bar {
+	return &Bar{
+		Callback: foo.Callback{
+			Vptr: &foo.CallbackVtbl{
+				Val:  c.Func((*Bar).getA),
+				Calc: c.Func((*Bar).sqrt),
+			},
+		},
+		a: a,
+	}
+}
+
+func (p *Bar) getA() c.Int {
+	return p.a
+}
+
+func (p *Bar) sqrt(v float64) float64 {
+	return math.Sqrt(v)
+}
+
+func main() {
+	bar := NewBar(1)
+	foo.F(&bar.Callback)
+	foo.G(&bar.Callback)
+}

_demo/cppintf/foo/bar/bar.cpp

@@ -0,0 +1,15 @@
+#include <stdio.h>
+#define interface struct
+
+interface ICallback {
+	virtual int val() = 0;
+	virtual double calc(double v) = 0;
+};
+
+extern "C" void f(ICallback* cb) {
+	printf("val: %d\ncalc(2): %lf\n", cb->val(), cb->calc(2));
+}
+
+void g(ICallback* cb) {
+	f(cb);
+}

_demo/cppintf/foo/foo.go

@@ -0,0 +1,25 @@
+package foo
+
+import (
+	"unsafe"
+)
+
+const (
+	LLGoFiles   = "bar/bar.cpp"
+	LLGoPackage = "link"
+)
+
+type Callback struct {
+	Vptr *CallbackVtbl
+}
+
+type CallbackVtbl struct {
+	Val  unsafe.Pointer
+	Calc unsafe.Pointer
+}
+
+//go:linkname F C.f
+func F(cb *Callback)
+
+//go:linkname G C._Z1gP9ICallback
+func G(cb *Callback)

_demo/cppmintf/cpp_multi_intf.go

@@ -0,0 +1,50 @@
+package main
+
+import (
+	"unsafe"
+
+	"github.com/goplus/llgo/_demo/cppmintf/foo"
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/math"
+)
+
+type Bar struct {
+	foo.Callback
+	a c.Int
+}
+
+func NewBar(a c.Int) *Bar {
+	return &Bar{
+		Callback: foo.Callback{
+			ICalc: foo.ICalc{
+				Vptr: &foo.ICalcVtbl{
+					Calc: c.Func((*Bar).sqrt),
+				},
+			},
+			IVal: foo.IVal{
+				Vptr: &foo.IValVtbl{
+					Val: c.Func(bar_IVal_getA),
+				},
+			},
+		},
+		a: a,
+	}
+}
+
+func (p *Bar) getA() c.Int {
+	return p.a
+}
+
+func bar_IVal_getA(this c.Pointer) c.Int {
+	const delta = -int(unsafe.Offsetof(foo.Callback{}.IVal))
+	return (*Bar)(c.Advance(this, delta)).getA()
+}
+
+func (p *Bar) sqrt(v float64) float64 {
+	return math.Sqrt(v)
+}
+
+func main() {
+	bar := NewBar(1)
+	foo.F(&bar.Callback)
+}

_demo/cppmintf/foo/bar/bar.cpp

@@ -0,0 +1,17 @@
+#include <stdio.h>
+#define interface struct
+
+interface ICalc {
+	virtual double calc(double v) = 0;
+};
+
+interface IVal {
+	virtual int val() = 0;
+};
+
+class Callback : public ICalc, public IVal {
+};
+
+extern "C" void f(Callback* cb) {
+	printf("val: %d\ncalc(2): %lf\n", cb->val(), cb->calc(2));
+}

_demo/cppmintf/foo/foo.go

@@ -0,0 +1,42 @@
+package foo
+
+import (
+	"unsafe"
+)
+
+const (
+	LLGoFiles   = "bar/bar.cpp"
+	LLGoPackage = "link"
+)
+
+// -----------------------------------------------------------------------------
+
+type ICalc struct {
+	Vptr *ICalcVtbl
+}
+
+type ICalcVtbl struct {
+	Calc unsafe.Pointer
+}
+
+// -----------------------------------------------------------------------------
+
+type IVal struct {
+	Vptr *IValVtbl
+}
+
+type IValVtbl struct {
+	Val unsafe.Pointer
+}
+
+// -----------------------------------------------------------------------------
+
+type Callback struct {
+	ICalc
+	IVal
+}
+
+//go:linkname F C.f
+func F(cb *Callback)
+
+// -----------------------------------------------------------------------------

_demo/ctime/time.go

@@ -0,0 +1,12 @@
+package main
+
+import "github.com/goplus/llgo/c/time"
+
+func main() {
+	var tv time.Timespec
+	time.ClockGettime(time.CLOCK_REALTIME, &tv)
+	println("REALTIME sec:", tv.Sec, "nsec:", tv.Nsec)
+
+	time.ClockGettime(time.CLOCK_MONOTONIC, &tv)
+	println("MONOTONIC sec:", tv.Sec, "nsec:", tv.Nsec)
+}

_demo/fcntl/fcntl.go

@@ -0,0 +1,77 @@
+package main
+
+import (
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/os"
+	"unsafe"
+)
+
+func main() {
+
+	filename := c.Str("testfile.txt")
+	data := c.Str("Hello, os!")
+	var buffer [20]c.Char
+
+	// Open a file, O_CREAT|O_WRONLY|O_TRUNC means create, write only, or clear the file
+	fd := os.Open(filename, os.O_CREAT|os.O_WRONLY|os.O_TRUNC, 0644)
+	if fd == -1 {
+		c.Printf(c.Str("open error\n"))
+		return
+	}
+
+	// Writing data to a file
+	bytesWritten := os.Write(fd, c.Pointer(data), c.Strlen(data))
+	if bytesWritten == -1 {
+		c.Printf(c.Str("write error\n"))
+		os.Close(fd)
+		return
+	}
+	c.Printf(c.Str("Written %ld bytes to %s\n"), bytesWritten, filename)
+
+	// Get file status flags
+	flags := os.Fcntl(fd, os.F_GETFL)
+	if flags == -1 {
+		c.Printf(c.Str("os error\n"))
+		os.Close(fd)
+		return
+	}
+	c.Printf(c.Str("File flags: %d\n"), flags)
+
+	// Set the file status flag to non-blocking mode
+	if os.Fcntl(fd, os.F_SETFL, flags|os.O_NONBLOCK) == -1 {
+		c.Printf(c.Str("os error\n"))
+		os.Close(fd)
+		return
+	}
+	c.Printf(c.Str("set file status successfully\n"))
+
+
+
+	c.Printf(c.Str("111"))
+	// Close file
+	os.Close(fd)
+
+	// Reopen the file, O_RDONLY means read-only
+	fd = os.Open(filename, os.O_RDONLY)
+	if fd == -1 {
+		c.Printf(c.Str("open error\n"))
+		return
+	}
+
+	// Reading data from a file
+	// &buffer[:][0]
+	// unsafe.SliceData(buffer[:])
+	bytesRead := os.Read(fd, c.Pointer(unsafe.SliceData(buffer[:])), unsafe.Sizeof(buffer)-1)
+	if bytesRead == -1 {
+		c.Printf(c.Str("read error\n"))
+		os.Close(fd)
+		return
+	}
+
+	// Ensure that the buffer is null-terminated
+	buffer[bytesRead] = c.Char(0)
+	c.Printf(c.Str("Read %ld bytes: %s\n"), bytesRead, &buffer[0])
+
+	// Close file
+	os.Close(fd)
+}

_demo/getcwd/getcwd.go

@@ -0,0 +1,11 @@
+package main
+
+import (
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/os"
+)
+
+func main() {
+	wd := os.Getcwd(c.Alloca(os.PATH_MAX), os.PATH_MAX)
+	c.Printf(c.Str("cwd: %s\n"), wd)
+}

_demo/goroutine/goroutine.go

@@ -0,0 +1,12 @@
+package main
+
+func main() {
+	done := false
+	go func() {
+		println("Hello, goroutine")
+		done = true
+	}()
+	for !done {
+		print(".")
+	}
+}

_demo/gotime/time.go

@@ -0,0 +1,7 @@
+package main
+
+import "time"
+
+func main() {
+	println(time.Now().String())
+}

_demo/hello/hello.go

@@ -9,5 +9,5 @@ func main() {
 }
 
 /* Expected output:
-Hello World
+Hello world
 */

_demo/linkname/linkname.go

@@ -0,0 +1,10 @@
+package main
+
+import _ "unsafe" // for go:linkname
+
+//go:linkname Sqrt C.sqrt
+func Sqrt(x float64) float64
+
+func main() {
+	println("sqrt(2) =", Sqrt(2))
+}

_demo/logdemo/log.go

@@ -0,0 +1,9 @@
+package main
+
+import (
+	"log"
+)
+
+func main() {
+	log.Println("Hello")
+}

_demo/math/math.go

@@ -0,0 +1,11 @@
+package main
+
+import (
+	"math"
+)
+
+func main() {
+	println(math.Sqrt(2))
+	println(math.Abs(-1.2))
+	println(math.Ldexp(1.2, 3))
+}

_demo/netdbdemo/netdb.go

@@ -0,0 +1,20 @@
+package main
+
+import (
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/net"
+)
+
+func main() {
+	var hints net.AddrInfo
+	hints.Family = net.AF_UNSPEC
+	hints.SockType = net.SOCK_STREAM
+
+	host := "httpbin.org"
+	port := "80"
+
+	var result *net.AddrInfo
+	c.Printf(c.Str("%d\n"), net.Getaddrinfo(c.Str(host), c.Str(port), &hints, &result))
+
+	c.Printf(c.Str("%d\n"), net.Freeaddrinfo(result))
+}

_demo/oslookpath/lookpath.go

@@ -0,0 +1,19 @@
+package main
+
+import (
+	"fmt"
+	"os/exec"
+	"runtime"
+)
+
+func main() {
+	ls := "ls"
+	if runtime.GOOS == "windows" {
+		ls = "dir"
+	}
+	lspath, _ := exec.LookPath(ls)
+	if lspath != "" {
+		ls = lspath
+	}
+	fmt.Println(ls)
+}

_demo/rand/rand.go

@@ -0,0 +1,15 @@
+package main
+
+import (
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/math/rand"
+)
+
+func main() {
+	var s c.Uint = 6
+	rand.Srand(s)
+	rr := rand.RandR(&s)
+	r := rand.Rand()
+	println("r:", r)
+	println("rr:", rr)
+}

_demo/setjmp/setjmp.go

@@ -0,0 +1,16 @@
+package main
+
+import (
+	"github.com/goplus/llgo/c/setjmp"
+)
+
+func main() {
+	var jb setjmp.SigjmpBuf
+	switch ret := setjmp.Sigsetjmp(&jb, 0); ret {
+	case 0:
+		println("Hello, setjmp!")
+		setjmp.Siglongjmp(&jb, 1)
+	default:
+		println("exception:", ret)
+	}
+}

_demo/socket/client/client.go

@@ -0,0 +1,32 @@
+package main
+
+import (
+	"unsafe"
+
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/net"
+	"github.com/goplus/llgo/c/os"
+)
+
+func main() {
+	sockfd := net.Socket(net.AF_INET, net.SOCK_STREAM, 0)
+	msg := c.Str("Hello, World!")
+	defer os.Close(sockfd)
+
+	server := net.GetHostByName(c.Str("localhost"))
+	if server == nil {
+		c.Perror(c.Str("hostname get error"))
+		return
+	}
+
+	servAddr := &net.SockaddrIn{}
+	servAddr.Family = net.AF_INET
+	servAddr.Port = net.Htons(uint16(1234))
+	c.Memcpy(unsafe.Pointer(&servAddr.Addr.Addr), unsafe.Pointer(*server.AddrList), uintptr(server.Length))
+
+	if res := net.Connect(sockfd, (*net.SockAddr)(unsafe.Pointer(servAddr)), c.Uint(16)); res < 0 {
+		c.Perror(c.Str("connect error"))
+		return
+	}
+	os.Write(sockfd, unsafe.Pointer(msg), c.Strlen(msg))
+}

_demo/socket/server/server.go

@@ -0,0 +1,43 @@
+package main
+
+import (
+	"unsafe"
+
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/net"
+	"github.com/goplus/llgo/c/os"
+)
+
+func main() {
+	var buffer [256]c.Char
+
+	sockfd := net.Socket(net.AF_INET, net.SOCK_STREAM, 0)
+	defer os.Close(sockfd)
+
+	servAddr := &net.SockaddrIn{
+		Family: net.AF_INET,
+		Port:   net.Htons(uint16(1234)),
+		Addr:   net.InAddr{Addr: 0x00000000},
+		Zero:   [8]c.Char{0, 0, 0, 0, 0, 0, 0, 0},
+	}
+	if res := net.Bind(sockfd, servAddr, c.Uint(unsafe.Sizeof(*servAddr))); res < 0 {
+		c.Perror(c.Str("bind error"))
+		return
+	}
+
+	if net.Listen(sockfd, 5) < 0 {
+		c.Printf(c.Str("listen error"))
+		return
+	}
+	c.Printf(c.Str("Listening on port 1234...\n"))
+
+	cliAddr, clilen := &net.SockaddrIn{}, c.Uint(unsafe.Sizeof(servAddr))
+
+	newsockfd := net.Accept(sockfd, cliAddr, &clilen)
+	defer os.Close(newsockfd)
+	c.Printf(c.Str("Connection accepted."))
+
+	os.Read(newsockfd, unsafe.Pointer(unsafe.SliceData(buffer[:])), 256)
+	c.Printf(c.Str("Received: %s"), &buffer[0])
+
+}

_demo/sysexec/exec.go

@@ -0,0 +1,26 @@
+package main
+
+import (
+	"fmt"
+	"os/exec"
+	"runtime"
+	"syscall"
+)
+
+func main() {
+	ls := "ls"
+	args := []string{ls, "-l"}
+	if runtime.GOOS == "windows" {
+		ls = "dir"
+		args = []string{ls}
+	}
+	lspath, _ := exec.LookPath(ls)
+	if lspath != "" {
+		ls = lspath
+	}
+	err := syscall.Exec(ls, args, nil)
+	if err != nil {
+		fmt.Println("syscall.Exec error:", err)
+		return
+	}
+}

_demo/thread/thd.go

@@ -0,0 +1,25 @@
+package main
+
+import (
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/pthread"
+)
+
+var key pthread.Key
+
+func main() {
+	key.Create(nil)
+	key.Set(c.Pointer(c.Str("main value\n")))
+
+	var thd pthread.Thread
+	pthread.Create(&thd, nil, func(arg c.Pointer) c.Pointer {
+		key.Set(c.Pointer(c.Str("thread value\n")))
+		c.Printf(c.Str("Hello, thread\nTLS: %s"), key.Get())
+		return c.Pointer(c.Str("Back to main\n"))
+	}, nil)
+
+	var retval c.Pointer
+	pthread.Join(thd, &retval)
+
+	c.Printf(c.Str("%sTLS: %s"), retval, key.Get())
+}

_demo/timedur/timedur.go

@@ -0,0 +1,11 @@
+package main
+
+import (
+	"fmt"
+	"time"
+)
+
+func main() {
+	t := time.Now().Add(time.Second * 5)
+	fmt.Println(time.Until(t))
+}

_pydemo/callpy/callpy.go

@@ -1,12 +1,12 @@
 package main
 
 import (
-	"github.com/goplus/llgo/c"
 	"github.com/goplus/llgo/py"
 	"github.com/goplus/llgo/py/math"
+	"github.com/goplus/llgo/py/std"
 )
 
 func main() {
-	x := math.Sqrt(py.Float(2))
-	c.Printf(c.Str("sqrt(2) = %f\n"), x.Float64())
+	x := math.Sqrt(py.Float(2))       // x = sqrt(2)
+	std.Print(py.Str("sqrt(2) ="), x) // print("sqrt(2) =", x)
 }

_pydemo/matrix/matrix.go

@@ -1,9 +1,9 @@
 package main
 
 import (
-	"github.com/goplus/llgo/c"
 	"github.com/goplus/llgo/py"
 	"github.com/goplus/llgo/py/numpy"
+	"github.com/goplus/llgo/py/std"
 )
 
 func main() {
@@ -18,5 +18,5 @@ func main() {
 		py.List(3.0, 2.0, 1.0),
 	)
 	x := numpy.Add(a, b)
-	c.Printf(c.Str("a+b = %s\n"), x.Str().CStr())
+	std.Print(py.Str("a+b ="), x)
 }

_pydemo/max/max.go

@@ -0,0 +1,15 @@
+package main
+
+import (
+	"github.com/goplus/llgo/py"
+	"github.com/goplus/llgo/py/std"
+)
+
+func main() {
+	x := std.Max(py.Float(3.0), py.Float(9.0), py.Float(23.0), py.Float(100.0))
+	std.Print(x)
+
+	list := py.List(3.0, 9.0, 23.0, 100.0)
+	y := std.Max(std.Iter(list))
+	std.Print(y)
+}

_pydemo/print/print.go

@@ -0,0 +1,11 @@
+package main
+
+import (
+	"github.com/goplus/llgo/py"
+	"github.com/goplus/llgo/py/std"
+)
+
+func main() {
+	x := py.Float(3.14)
+	std.Print(x)
+}

_pydemo/tensor/tensor.go

@@ -0,0 +1,16 @@
+package main
+
+import (
+	"github.com/goplus/llgo/py"
+	"github.com/goplus/llgo/py/std"
+	"github.com/goplus/llgo/py/torch"
+)
+
+func main() {
+	data := py.List(
+		py.List(1.0, 2.0),
+		py.List(3.0, 4.0),
+	)
+	x := torch.Tensor(data)
+	std.Print(x)
+}

c/bdwgc/_test/bdwgc.go

@@ -0,0 +1,101 @@
+package main
+
+import (
+	"unsafe"
+
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/bdwgc"
+	"github.com/goplus/llgo/c/bdwgc/_test/testing"
+)
+
+// ------ Test malloc ------
+
+func TestMalloc(t *testing.T) {
+	pn := (*int)(bdwgc.Malloc(unsafe.Sizeof(int(0))))
+	*pn = 1 << 30
+	c.Printf(c.Str("value: %d, %x, %p, %p\n"), *pn, *pn, pn, &pn)
+
+	pl := (*int64)(bdwgc.Realloc(c.Pointer(pn), unsafe.Sizeof(int64(0))))
+	*pl = 1 << 60
+	c.Printf(c.Str("value: %lld, %llx, %p, %p\n"), *pl, *pl, pl, &pl)
+
+	bdwgc.Free(c.Pointer(pl))
+}
+
+// ------ Test finalizer ------
+
+const (
+	RETURN_VALUE_FREED = 1 << 31
+)
+
+var called uint = 0
+
+func setReturnValueFreed(pobj c.Pointer, clientData c.Pointer) {
+	called |= RETURN_VALUE_FREED
+	c.Printf(c.Str("called: %x\n"), called)
+}
+
+func setLoopValueFreed(pobj c.Pointer, clientData c.Pointer) {
+	pmask := (*uint)(clientData)
+	called |= *pmask
+	c.Printf(c.Str("called: %x\n"), called)
+}
+
+func isCalled(mask uint) bool {
+	return called&mask != 0
+}
+
+func returnValue() *int {
+	pn := bdwgc.Malloc(unsafe.Sizeof(int(0)))
+	bdwgc.RegisterFinalizer(pn, setReturnValueFreed, nil, nil, nil)
+	return (*int)(pn)
+}
+
+func callFunc() {
+	pn := returnValue()
+	*pn = 1 << 30
+	c.Printf(c.Str("value: %d, %x, %p, %p\n"), *pn, *pn, pn, &pn)
+	bdwgc.Gcollect()
+	check(!isCalled(RETURN_VALUE_FREED), c.Str("finalizer should not be called"))
+}
+
+func loop() {
+	for i := 0; i < 5; i++ {
+		p := bdwgc.Malloc(unsafe.Sizeof(int(0)))
+		pn := (*int)(p)
+		*pn = i
+		c.Printf(c.Str("value: %d, %x, %p, %p\n"), *pn, *pn, pn, &pn)
+		pflag := (*uint)(c.Malloc(unsafe.Sizeof(uint(0))))
+		*pflag = 1 << i
+		bdwgc.RegisterFinalizer(p, setLoopValueFreed, c.Pointer(pflag), nil, nil)
+		bdwgc.Gcollect()
+		check(!isCalled(1<<i), c.Str("finalizer should not be called"))
+		for j := 0; j < i; j++ {
+			check(isCalled(1<<j), c.Str("finalizers of previous objects should be called"))
+		}
+	}
+}
+
+// Clear stack to avoid reference
+func clearStack() {
+	p := c.Alloca(128)
+	c.Memset(p, 0, 128)
+}
+
+func check(b bool, msg *c.Char) {
+	if !b {
+		c.Printf(c.Str("check failed: %s\n"), msg)
+		panic("check failed")
+	}
+}
+
+func TestFinalizer(t *testing.T) {
+	bdwgc.Init()
+
+	callFunc()
+	clearStack()
+	bdwgc.Gcollect()
+	check(isCalled(RETURN_VALUE_FREED), c.Str("finalizer should be called"))
+
+	loop()
+}

c/bdwgc/_test/main.go

@@ -0,0 +1,31 @@
+package main
+
+import (
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/bdwgc/_test/testing"
+)
+
+type TestCase struct {
+	Name string
+	F    func(*testing.T)
+}
+
+func main() {
+	tests := []TestCase{
+		{"TestMalloc", TestMalloc},
+		{"TestFinalizer", TestFinalizer},
+	}
+	if c.Argc == 1 {
+		for _, test := range tests {
+			c.Printf(c.Str("%s\n"), c.AllocaCStr(test.Name))
+		}
+		return
+	}
+	c.Fprintf(c.Stderr, c.Str("arg: %s\n"), c.Index(c.Argv, 1))
+	idx := int(c.Atoi(c.Index(c.Argv, 1)))
+	if idx < 0 || idx >= len(tests) {
+		c.Printf(c.Str("invalid test index %d"), idx)
+		panic("invalid test index")
+	}
+	tests[idx].F(nil)
+}

c/bdwgc/_test/testing/testing.go

@@ -0,0 +1,4 @@
+package testing
+
+type T struct {
+}

c/bdwgc/bdwgc.go

@@ -0,0 +1,103 @@
+/*
+ * Copyright (c) 2024 The GoPlus Authors (goplus.org). All rights reserved.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package bdwgc
+
+import (
+	_ "unsafe"
+
+	"github.com/goplus/llgo/c"
+)
+
+const (
+	LLGoPackage = "link: $(pkg-config --libs bdw-gc); -lgc"
+)
+
+// -----------------------------------------------------------------------------
+
+//go:linkname Init C.GC_init
+func Init()
+
+//go:linkname Malloc C.GC_malloc
+func Malloc(size uintptr) c.Pointer
+
+//go:linkname Realloc C.GC_realloc
+func Realloc(ptr c.Pointer, size uintptr) c.Pointer
+
+//go:linkname Free C.GC_free
+func Free(ptr c.Pointer)
+
+// -----------------------------------------------------------------------------
+
+//go:linkname RegisterFinalizer C.GC_register_finalizer
+func RegisterFinalizer(
+	obj c.Pointer,
+	fn func(c.Pointer, c.Pointer), cd c.Pointer,
+	oldFn *func(c.Pointer, c.Pointer), oldCd *c.Pointer)
+
+//go:linkname RegisterFinalizerNoOrder C.GC_register_finalizer_no_order
+func RegisterFinalizerNoOrder(
+	obj c.Pointer,
+	fn func(c.Pointer, c.Pointer), cd c.Pointer,
+	oldFn *func(c.Pointer, c.Pointer), oldCd *c.Pointer)
+
+//go:linkname RegisterFinalizerIgnoreSelf C.GC_register_finalizer_ignore_self
+func RegisterFinalizerIgnoreSelf(
+	obj c.Pointer,
+	fn func(c.Pointer, c.Pointer), cd c.Pointer,
+	oldFn *func(c.Pointer, c.Pointer), oldCd *c.Pointer)
+
+//go:linkname RegisterFinalizerUnreachable C.GC_register_finalizer_unreachable
+func RegisterFinalizerUnreachable(
+	obj c.Pointer,
+	fn func(c.Pointer, c.Pointer), cd c.Pointer,
+	oldFn *func(c.Pointer, c.Pointer), oldCd *c.Pointer)
+
+// -----------------------------------------------------------------------------
+
+//go:linkname Enable C.GC_enable
+func Enable()
+
+//go:linkname Disable C.GC_disable
+func Disable()
+
+//go:linkname IsDisabled C.GC_is_disabled
+func IsDisabled() c.Int
+
+//go:linkname Gcollect C.GC_gcollect
+func Gcollect()
+
+//go:linkname GetMemoryUse C.GC_get_memory_use
+func GetMemoryUse() uintptr
+
+// -----------------------------------------------------------------------------
+
+//go:linkname EnableIncremental C.GC_enable_incremental
+func EnableIncremental()
+
+//go:linkname IsIncrementalMode C.GC_is_incremental_mode
+func IsIncrementalMode() c.Int
+
+//go:linkname IncrementalProtectionNeeds C.GC_incremental_protection_needs
+func IncrementalProtectionNeeds() c.Int
+
+//go:linkname StartIncrementalCollection C.GC_start_incremental_collection
+func StartIncrementalCollection()
+
+//go:linkname CollectALittle C.GC_collect_a_little
+func CollectALittle()
+
+// -----------------------------------------------------------------------------

c/bitcast/_cast/cast.c

@@ -0,0 +1,17 @@
+typedef union {
+    double d;
+    float  f;
+    long   v;
+} castUnion;
+
+double llgoToFloat64(long v) {
+    castUnion k;
+    k.v = v;
+    return k.d;
+}
+
+float llgoToFloat32(long v) {
+    castUnion k;
+    k.v = v;
+    return k.f;
+}

c/bitcast/bitcast.go

@@ -0,0 +1,30 @@
+/*
+ * Copyright (c) 2024 The GoPlus Authors (goplus.org). All rights reserved.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package bitcast
+
+import _ "unsafe"
+
+const (
+	LLGoFiles   = "_cast/cast.c"
+	LLGoPackage = "link"
+)
+
+//go:linkname ToFloat64 C.llgoToFloat64
+func ToFloat64(v uintptr) float64
+
+//go:linkname ToFloat32 C.llgoToFloat32
+func ToFloat32(v uintptr) float32

c/c.go

@@ -17,6 +17,9 @@
 package c
 
 // typedef unsigned int uint;
+// typedef unsigned long ulong;
+// typedef unsigned long long ulonglong;
+// typedef long long longlong;
 import "C"
 import "unsafe"
 
@@ -25,17 +28,23 @@ const (
 )
 
 type (
-	Char      = int8
-	Int       = C.int
-	Uint      = C.uint
-	Long      = int32
-	Ulong     = uint32
-	LongLong  = int64
-	UlongLong = uint64
-	Float     = float32
-	Double    = float64
-	Pointer   = unsafe.Pointer
-	FilePtr   = unsafe.Pointer
+	Void    = [0]byte
+	Char    = int8
+	Float   = float32
+	Double  = float64
+	Pointer = unsafe.Pointer
+	FilePtr = unsafe.Pointer
+)
+
+type (
+	Int  C.int
+	Uint C.uint
+
+	Long  C.long
+	Ulong C.ulong
+
+	LongLong  C.longlong
+	UlongLong C.ulonglong
 )
 
 type integer interface {
@@ -45,8 +54,11 @@ type integer interface {
 //go:linkname Str llgo.cstr
 func Str(string) *Char
 
+//go:linkname Func llgo.funcAddr
+func Func(any) Pointer
+
 // llgo:link Advance llgo.advance
-func Advance[PtrT any](ptr PtrT, offset int) PtrT { return ptr }
+func Advance[PtrT any, I integer](ptr PtrT, offset I) PtrT { return ptr }
 
 // llgo:link Index llgo.index
 func Index[T any, I integer](ptr *T, offset I) T { return *ptr }
@@ -57,27 +69,114 @@ func Alloca(size uintptr) Pointer
 //go:linkname AllocaCStr llgo.allocaCStr
 func AllocaCStr(s string) *Char
 
-//go:linkname Unreachable llgo.unreachable
-func Unreachable()
+//go:linkname AllocaCStrs llgo.allocaCStrs
+func AllocaCStrs(strs []string, endWithNil bool) **Char
+
+// TODO(xsw):
+// llgo:link AllocaNew llgo.allocaNew
+func AllocaNew[T any]() *T { return nil }
 
 //go:linkname Malloc C.malloc
 func Malloc(size uintptr) Pointer
 
+//go:linkname Free C.free
+func Free(ptr Pointer)
+
 //go:linkname Memcpy C.memcpy
 func Memcpy(dst, src Pointer, n uintptr) Pointer
 
+//go:linkname Memmove C.memmove
+func Memmove(dst, src Pointer, n uintptr) Pointer
+
 //go:linkname Memset C.memset
 func Memset(s Pointer, c Int, n uintptr) Pointer
 
+//go:linkname Memchr C.memchr
+func Memchr(s Pointer, c Int, n uintptr) Pointer
+
+//go:linkname Memcmp C.memcmp
+func Memcmp(s1, s2 Pointer, n uintptr) Int
+
 // -----------------------------------------------------------------------------
 
+//go:linkname Strlen C.strlen
+func Strlen(s *Char) uintptr
+
+//go:linkname Strcpy C.strcpy
+func Strcpy(dst, src *Char) *Char
+
+//go:linkname Strncpy C.strncpy
+func Strncpy(dst, src *Char, n uintptr) *Char
+
+//go:linkname Strcat C.strcat
+func Strcat(dst, src *Char) *Char
+
+//go:linkname Strncat C.strncat
+func Strncat(dst, src *Char, n uintptr) *Char
+
+//go:linkname Strcmp C.strcmp
+func Strcmp(s1, s2 *Char) Int
+
+//go:linkname Strncmp C.strncmp
+func Strncmp(s1, s2 *Char, n uintptr) Int
+
+//go:linkname Strchr C.strchr
+func Strchr(s *Char, c Int) *Char
+
+//go:linkname Strrchr C.strrchr
+func Strrchr(s *Char, c Int) *Char
+
+//go:linkname Strstr C.strstr
+func Strstr(s1, s2 *Char) *Char
+
+//go:linkname Strdup C.strdup
+func Strdup(s *Char) *Char
+
+//go:linkname Strndup C.strndup
+func Strndup(s *Char, n uintptr) *Char
+
+//go:linkname Strtok C.strtok
+func Strtok(s, delim *Char) *Char
+
+//go:linkname Strerror C.strerror
+func Strerror(errnum Int) *Char
+
+//go:linkname Sprintf C.sprintf
+func Sprintf(s *Char, format *Char, __llgo_va_list ...any) Int
+
+//go:linkname Snprintf C.snprintf
+func Snprintf(s *Char, n uintptr, format *Char, __llgo_va_list ...any) Int
+
+//go:linkname Vsnprintf C.vsnprintf
+func Vsnprintf(s *Char, n uintptr, format *Char, ap Pointer) Int
+
+// -----------------------------------------------------------------------------
+
+// GoString converts a C string to a Go string.
+// TODO(xsw): any => int
+//
+//go:linkname GoString llgo.string
+func GoString(cstr *Char, __llgo_va_list /* n */ ...any) string
+
 //go:linkname GoStringData llgo.stringData
 func GoStringData(string) *Char
 
+//go:linkname GoDeferData llgo.deferData
+func GoDeferData() Pointer
+
 // -----------------------------------------------------------------------------
 
-//go:linkname Remove C.remove
-func Remove(path *Char) Int
+//go:linkname AllocaSigjmpBuf llgo.sigjmpbuf
+func AllocaSigjmpBuf() Pointer
+
+//go:linkname Sigsetjmp llgo.sigsetjmp
+func Sigsetjmp(jb Pointer, savemask Int) Int
+
+//go:linkname Siglongjmp llgo.siglongjmp
+func Siglongjmp(jb Pointer, retval Int)
+
+//go:linkname Unreachable llgo.unreachable
+func Unreachable()
 
 // -----------------------------------------------------------------------------
 
@@ -94,14 +193,10 @@ func Qsort(base Pointer, count, elem uintptr, compar func(a, b Pointer) Int)
 
 // -----------------------------------------------------------------------------
 
-//go:linkname Stdin __stdinp
-var Stdin FilePtr
+//go:linkname Atoi C.atoi
+func Atoi(s *Char) Int
 
-//go:linkname Stdout __stdoutp
-var Stdout FilePtr
-
-//go:linkname Stderr __stderrp
-var Stderr FilePtr
+// -----------------------------------------------------------------------------
 
 //go:linkname Printf C.printf
 func Printf(format *Char, __llgo_va_list ...any) Int
@@ -109,11 +204,35 @@ func Printf(format *Char, __llgo_va_list ...any) Int
 //go:linkname Fprintf C.fprintf
 func Fprintf(fp FilePtr, format *Char, __llgo_va_list ...any) Int
 
+//go:linkname Fwrite C.fwrite
+func Fwrite(data Pointer, size, count uintptr, fp FilePtr) uintptr
+
+//go:linkname Fputc C.fputc
+func Fputc(c Int, fp FilePtr) Int
+
+//go:linkname Fputs C.fputs
+func Fputs(s *Char, fp FilePtr) Int
+
+//go:linkname Fflush C.fflush
+func Fflush(fp FilePtr) Int
+
+//go:linkname Fopen C.fopen
+func Fopen(c *Char, mod *Char) FilePtr
+
+//go:linkname Fclose C.fclose
+func Fclose(fp FilePtr) Int
+
+//go:linkname Perror C.perror
+func Perror(s *Char)
+
 // -----------------------------------------------------------------------------
 
 //go:linkname Time C.time
 func Time(*int32) int32
 
+//go:linkname Usleep C.usleep
+func Usleep(useconds Uint) Int
+
 // -----------------------------------------------------------------------------
 
 type Option struct {

c/c_default.go

@@ -0,0 +1,31 @@
+//go:build !linux
+// +build !linux
+
+/*
+ * Copyright (c) 2024 The GoPlus Authors (goplus.org). All rights reserved.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package c
+
+import _ "unsafe"
+
+//go:linkname Stdin __stdinp
+var Stdin FilePtr
+
+//go:linkname Stdout __stdoutp
+var Stdout FilePtr
+
+//go:linkname Stderr __stderrp
+var Stderr FilePtr

c/c_linux.go

@@ -0,0 +1,31 @@
+//go:build linux
+// +build linux
+
+/*
+ * Copyright (c) 2024 The GoPlus Authors (goplus.org). All rights reserved.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package c
+
+import _ "unsafe"
+
+//go:linkname Stdin stdin
+var Stdin FilePtr
+
+//go:linkname Stdout stdout
+var Stdout FilePtr
+
+//go:linkname Stderr stderr
+var Stderr FilePtr

c/cjson/README.md

@@ -1,22 +1,18 @@
 LLGo wrapper of DaveGamble/cJSON
 =====
-[![Build Status](https://github.com/goplus/cjson/actions/workflows/go.yml/badge.svg)](https://github.com/goplus/cjson/actions/workflows/go.yml)
-[![GitHub release](https://img.shields.io/github/v/tag/goplus/cjson.svg?label=release)](https://github.com/goplus/cjson/releases)
-[![GoDoc](https://pkg.go.dev/badge/github.com/goplus/cjson.svg)](https://pkg.go.dev/github.com/goplus/cjson)
-[![Compiler](https://img.shields.io/badge/compiler-llgo-darkgreen.svg)](https://github.com/goplus/llgo)
-[![Language](https://img.shields.io/badge/language-Go+-blue.svg)](https://github.com/goplus/gop)
 
 ## How to install
 
+### on macOS (Homebrew)
+
 ```sh
-git clone https://github.com/goplus/cjson.git
-cd cjson
-git submodule init
-git submodule update
-mkdir build.dir
-cd build.dir
-cmake ../cJSON
-sudo make install
+brew install cjson 
+```
+
+### on Linux (Debian/Ubuntu)
+
+```sh
+apt-get install -y libcjson-dev
 ```
 
 ## Demos

c/cjson/_demo/mkjson/mkjson.go

@@ -22,6 +22,22 @@ func main() {
 
 	mod.SetItem(c.Str("items"), syms)
 
-	c.Printf(c.Str("%s\n"), mod.CStr())
+	cstr := mod.CStr()
+	str := c.GoString(cstr)
+	c.Printf(c.Str("%s\n"), cstr)
+	cjson.FreeCStr(cstr)
+
+	mod.Delete()
+
+	cjsonLoad(str)
+}
+
+func cjsonLoad(str string) {
+	mod := cjson.ParseString(str)
+
+	cstr := mod.Print()
+	c.Printf(c.Str("%s\n"), cstr)
+	cjson.FreeCStr(cstr)
+
 	mod.Delete()
 }

c/cjson/cjson.go

@@ -17,13 +17,13 @@
 package cjson
 
 import (
-	_ "unsafe"
+	"unsafe"
 
 	"github.com/goplus/llgo/c"
 )
 
 const (
-	LLGoPackage = "link: cjson"
+	LLGoPackage = "link: $(pkg-config --libs libcjson); -lcjson"
 )
 
 // llgo:type C
@@ -31,6 +31,20 @@ type JSON struct {
 	Unused [0]byte
 }
 
+//go:linkname Parse C.cJSON_Parse
+func Parse(value *c.Char) *JSON
+
+//go:linkname ParseWithLength C.cJSON_ParseWithLength
+func ParseWithLength(value *byte, valueLength uintptr) *JSON
+
+func ParseBytes(value []byte) *JSON {
+	return ParseWithLength(unsafe.SliceData(value), uintptr(len(value)))
+}
+
+func ParseString(value string) *JSON {
+	return ParseWithLength(unsafe.StringData(value), uintptr(len(value)))
+}
+
 //go:linkname Null C.cJSON_CreateNull
 func Null() *JSON
 
@@ -119,3 +133,9 @@ func (o *JSON) PrintUnformatted() *c.Char { return nil }
 //
 // llgo:link (*JSON).PrintBuffered C.cJSON_PrintBuffered
 func (o *JSON) PrintBuffered(prebuffer c.Int, fmt c.Int) *c.Char { return nil }
+
+//go:linkname Free C.cJSON_free
+func Free(ptr unsafe.Pointer)
+
+//go:linkname FreeCStr C.cJSON_free
+func FreeCStr(*c.Char)

c/clang/_demo/castdump/astdump.go

@@ -0,0 +1,61 @@
+package main
+
+import (
+	"fmt"
+	"os"
+
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/clang"
+)
+
+func visit(cursor, parent clang.Cursor, clientData c.Pointer) clang.ChildVisitResult {
+	depth := *(*c.Uint)(clientData)
+	printAST(cursor, depth+1)
+	return clang.ChildVisit_Continue
+}
+
+func printAST(cursor clang.Cursor, depth c.Uint) {
+	cursorKind := cursor.Kind.String()
+
+	cursorSpelling := cursor.String()
+
+	for i := c.Uint(0); i < depth; i++ {
+		c.Fputs(c.Str("  "), c.Stdout)
+	}
+
+	c.Printf(c.Str("%s: %s\n"), cursorKind.CStr(), cursorSpelling.CStr())
+
+	cursorKind.Dispose()
+	cursorSpelling.Dispose()
+
+	clang.VisitChildren(cursor, visit, c.Pointer(&depth))
+}
+
+func main() {
+	if c.Argc != 2 {
+		fmt.Fprintln(os.Stderr, "Usage: castdump <headerFile>")
+		return
+	}
+	sourceFile := *c.Advance(c.Argv, 1)
+
+	index := clang.CreateIndex(0, 0)
+
+	unit := index.ParseTranslationUnit(
+		sourceFile,
+		nil, 0,
+		nil, 0,
+		clang.TranslationUnit_None,
+	)
+
+	if unit == nil {
+		println("Unable to parse translation unit. Quitting.")
+		c.Exit(1)
+	}
+
+	cursor := unit.Cursor()
+
+	printAST(cursor, 0)
+
+	unit.Dispose()
+	index.Dispose()
+}

c/clang/_demo/symboldump/symboldump.go

@@ -0,0 +1,131 @@
+package main
+
+import (
+	"fmt"
+	"os"
+	"unsafe"
+
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/clang"
+)
+
+type Context struct {
+	namespaceName string
+	className     string
+}
+
+func newContext() *Context {
+	return &Context{}
+}
+
+func (c *Context) setNamespaceName(name string) {
+	c.namespaceName = name
+}
+
+func (c *Context) setClassName(name string) {
+	c.className = name
+}
+
+var context = newContext()
+
+func print_cursor_info(cursor clang.Cursor) {
+	loc := cursor.Location()
+	var file clang.File
+	var line, column c.Uint
+
+	loc.SpellingLocation(&file, &line, &column, nil)
+	filename := file.FileName()
+
+	c.Printf(c.Str("%s:%d:%d\n"), filename.CStr(), line, column)
+
+	cursorStr := cursor.String()
+	symbol := cursor.Mangling()
+	defer symbol.Dispose()
+	defer cursorStr.Dispose()
+	defer filename.Dispose()
+
+	if context.namespaceName != "" && context.className != "" {
+		fmt.Printf("%s:%s:", context.namespaceName, context.className)
+	} else if context.namespaceName != "" {
+		fmt.Printf("%s:", context.namespaceName)
+	}
+	c.Printf(c.Str("%s\n"), cursorStr.CStr())
+
+	if cursor.Kind == clang.CXXMethod || cursor.Kind == clang.FunctionDecl {
+		c.Printf(c.Str("symbol:%s\n"), symbol.CStr())
+
+		typeStr := cursor.ResultType().String()
+		defer typeStr.Dispose()
+		c.Printf(c.Str("Return Type: %s\n"), typeStr.CStr())
+		c.Printf(c.Str("Parameters(%d): ( "), cursor.NumArguments())
+
+		for i := 0; i < int(cursor.NumArguments()); i++ {
+			argCurSor := cursor.Argument(c.Uint(i))
+			argType := argCurSor.Type().String()
+			argName := argCurSor.String()
+			c.Printf(c.Str("%s %s"), argType.CStr(), argName.CStr())
+			if i < int(cursor.NumArguments())-1 {
+				c.Printf(c.Str(", "))
+			}
+
+			argType.Dispose()
+			argName.Dispose()
+		}
+
+		c.Printf(c.Str(" )\n"))
+		println("--------------------------------")
+	}
+}
+
+func visit(cursor, parent clang.Cursor, clientData c.Pointer) clang.ChildVisitResult {
+	if cursor.Kind == clang.Namespace {
+		nameStr := cursor.String()
+		context.setNamespaceName(c.GoString(nameStr.CStr()))
+		clang.VisitChildren(cursor, visit, nil)
+		context.setNamespaceName("")
+	} else if cursor.Kind == clang.ClassDecl {
+		nameStr := cursor.String()
+		context.setClassName(c.GoString(nameStr.CStr()))
+		clang.VisitChildren(cursor, visit, nil)
+		context.setClassName("")
+	} else if cursor.Kind == clang.CXXMethod || cursor.Kind == clang.FunctionDecl {
+		print_cursor_info(cursor)
+	}
+
+	return clang.ChildVisit_Continue
+}
+
+func parse(filename *c.Char) {
+	index := clang.CreateIndex(0, 0)
+	args := make([]*c.Char, 3)
+	args[0] = c.Str("-x")
+	args[1] = c.Str("c++")
+	args[2] = c.Str("-std=c++11")
+	unit := index.ParseTranslationUnit(
+		filename,
+		unsafe.SliceData(args), 3,
+		nil, 0,
+		clang.TranslationUnit_None,
+	)
+
+	if unit == nil {
+		println("Unable to parse translation unit. Quitting.")
+		c.Exit(1)
+	}
+
+	cursor := unit.Cursor()
+
+	clang.VisitChildren(cursor, visit, nil)
+	unit.Dispose()
+	index.Dispose()
+}
+
+func main() {
+	if c.Argc != 2 {
+		fmt.Fprintln(os.Stderr, "Usage: <C++ header file>\n")
+		return
+	} else {
+		sourceFile := *c.Advance(c.Argv, 1)
+		parse(sourceFile)
+	}
+}

c/clang/_wrap/cursor.cpp

@@ -0,0 +1,50 @@
+#include <clang-c/Index.h>
+#include <stdio.h>
+
+typedef enum CXChildVisitResult (*wrap_CXCursorVisitor)(CXCursor *cursor, CXCursor *parent, CXClientData client_data);
+
+typedef struct {
+    CXClientData data;
+    wrap_CXCursorVisitor visitor;
+} wrap_data;
+
+CXChildVisitResult wrap_visitor(CXCursor cursor, CXCursor parent, CXClientData data) {
+    wrap_data *d = (wrap_data *)(data);
+    return d->visitor(&cursor, &parent, d->data);
+}
+
+extern "C" {
+
+CXString wrap_clang_getCursorSpelling(CXCursor *cur) { return clang_getCursorSpelling(*cur); }
+
+CXString wrap_clang_Cursor_getMangling(CXCursor *cur) { return clang_Cursor_getMangling(*cur); }
+
+int wrap_clang_Cursor_getNumArguments(CXCursor *cur) { return clang_Cursor_getNumArguments(*cur); }
+
+void wrap_clang_Cursor_getArgument(CXCursor *C, unsigned i, CXCursor *argCur) {
+    *argCur = clang_Cursor_getArgument(*C, i);
+}
+
+void wrap_clang_getTranslationUnitCursor(CXTranslationUnit uint, CXCursor *cur) {
+    *cur = clang_getTranslationUnitCursor(uint);
+}
+
+void wrap_clang_getCursorType(CXCursor *cur, CXType *typ) { *typ = clang_getCursorType(*cur); }
+
+void wrap_clang_getCursorResultType(CXCursor *cur, CXType *typ) { *typ = clang_getCursorResultType(*cur); }
+
+CXString wrap_clang_getTypeSpelling(CXType *typ) { return clang_getTypeSpelling(*typ); }
+
+void wrap_clang_getCursorLocation(CXCursor *cur, CXSourceLocation *loc) { *loc = clang_getCursorLocation(*cur); }
+
+void wrap_clang_getSpellingLocation(CXSourceLocation *loc, CXFile *file, unsigned *line, unsigned *column,
+                                    unsigned *offset) {
+    clang_getSpellingLocation(*loc, file, line, column, offset);
+}
+
+unsigned wrap_clang_visitChildren(CXCursor *parent, wrap_CXCursorVisitor visitor, CXClientData client_data) {
+    wrap_data data = {client_data, visitor};
+    return clang_visitChildren(*parent, wrap_visitor, CXClientData(&data));
+}
+
+} // extern "C"

c/clang/_wrap/llgo_check.cpp

@@ -0,0 +1,7 @@
+#include <stdio.h>
+#include <clang-c/Index.h>
+
+int main() {
+    printf("sizeof(clang.Cursor) = %lu\n", sizeof(CXCursor));
+    return 0;
+}

c/clang/_wrap/llgo_check.go

@@ -0,0 +1,16 @@
+package main
+
+import (
+	"unsafe"
+
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/clang"
+)
+
+const (
+	LLGoCFlags = "-I$(llvm-config --includedir)"
+)
+
+func main() {
+	c.Printf(c.Str("sizeof(clang.Cursor) = %lu\n"), unsafe.Sizeof(clang.Cursor{}))
+}

c/clang/basic.go

@@ -0,0 +1,59 @@
+/*
+ * Copyright (c) 2024 The GoPlus Authors (goplus.org). All rights reserved.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package clang
+
+import (
+	_ "unsafe"
+
+	"github.com/goplus/llgo/c"
+)
+
+/**
+ * A character string.
+ *
+ * The \c CXString type is used to return strings from the interface when
+ * the ownership of that string might differ from one call to the next.
+ * Use \c clang_getCString() to retrieve the string data and, once finished
+ * with the string data, call \c clang_disposeString() to free the string.
+ */
+type String struct {
+	Data         c.Pointer
+	PrivateFlags c.Uint
+}
+
+/**
+ * Retrieve the character data associated with the given string.
+ */
+// llgo:link String.CStr C.clang_getCString
+func (String) CStr() *c.Char { return nil }
+
+/**
+ * Free the given string.
+ */
+// llgo:link String.Dispose C.clang_disposeString
+func (String) Dispose() {}
+
+type StringSet struct {
+	Strings *String
+	Count   c.Uint
+}
+
+/**
+ * Free the given string set.
+ */
+// llgo:link (*StringSet).Dispose C.clang_disposeStringSet
+func (*StringSet) Dispose() {}

c/libuv/README.md

@@ -0,0 +1,44 @@
+LLGo wrapper of libuv
+=====
+
+## How to install
+
+### on macOS (Homebrew)
+
+```sh
+brew install libuv
+```
+
+### on Linux (Debian/Ubuntu)
+
+```sh
+apt-get install -y libuv1-dev
+```
+
+### on Linux (CentOS/RHEL)
+
+```sh
+yum install -y libuv-devel
+```
+
+### on Linux (Arch Linux)
+
+```sh
+pacman -S libuv
+```
+
+## Demos
+
+The `_demo` directory contains our demos (it start with `_` to prevent the `go` command from compiling it):
+
+* [async_fs](_demo/async_fs/async_fs.go): a simple async file read demo
+* [echo_server](_demo/echo_server/echo_server.go): a basic async tcp echo server
+
+### How to run demos
+
+To run the demos in directory `_demo`:
+
+```sh
+cd <demo-directory>  # eg. cd _demo/sqlitedemo
+llgo run .
+```

c/libuv/_demo/async_fs/async_fs.go

@@ -0,0 +1,97 @@
+package main
+
+import (
+	"unsafe"
+
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/libuv"
+	"github.com/goplus/llgo/c/os"
+)
+
+const BUFFER_SIZE = 1024
+
+var (
+	loop     *libuv.Loop
+	openReq  libuv.Fs
+	readReq  libuv.Fs
+	closeReq libuv.Fs
+
+	buffer [BUFFER_SIZE]c.Char
+	iov    libuv.Buf
+)
+
+func main() {
+	// Print the libuv version
+	c.Printf(c.Str("libuv version: %d\n"), libuv.Version())
+
+	// Initialize the loop
+	loop = libuv.DefaultLoop()
+
+	// Open the file
+	libuv.FsOpen(loop, &openReq, c.Str("example.txt"), os.O_RDONLY, 0, onOpen)
+
+	// Run the loop
+	libuv.Run(loop, libuv.RUN_DEFAULT)
+
+	// Cleanup
+	defer cleanup()
+}
+
+func onOpen(req *libuv.Fs) {
+	// Check for errors
+	if libuv.FsGetResult(req) < 0 {
+		c.Fprintf(c.Stderr, c.Str("Error opening file: %s\n"), libuv.Strerror(libuv.Errno(libuv.LoopClose(loop))))
+		libuv.LoopClose(loop)
+		return
+	}
+	// Init buffer
+	iov = libuv.InitBuf((*c.Char)(unsafe.Pointer(&buffer[0])), c.Uint(unsafe.Sizeof(buffer)))
+	// Read the file
+	readRes := libuv.FsRead(loop, &readReq, libuv.UvFile(libuv.FsGetResult(req)), &iov, 1, -1, onRead)
+	if readRes != 0 {
+		c.Printf(c.Str("Error in FsRead: %s (code: %d)\n"), libuv.Strerror(libuv.Errno(readRes)), readRes)
+		libuv.LoopClose(loop)
+		return
+	}
+}
+
+func onRead(req *libuv.Fs) {
+	// Check for errors
+	if libuv.FsGetResult(req) < 0 {
+		c.Fprintf(c.Stderr, c.Str("Read error: %s\n"), libuv.Strerror(libuv.Errno(libuv.FsGetResult(req))))
+		libuv.LoopClose(loop)
+	} else if libuv.FsGetResult(req) == 0 {
+		c.Printf(c.Str("EOF\n"))
+		// Close the file
+		closeRes := libuv.FsClose(loop, &closeReq, libuv.UvFile(libuv.FsGetResult(&openReq)), onClose)
+		if closeRes != 0 {
+			// Print the content
+			c.Printf(c.Str("Error in FsClose: %s (code: %d)\n"), libuv.Strerror(libuv.Errno(closeRes)), closeRes)
+			libuv.LoopClose(loop)
+			return
+		}
+	} else {
+		c.Printf(c.Str("Read %d bytes\n"), libuv.FsGetResult(req))
+		c.Printf(c.Str("Read content: %.*s\n"), libuv.FsGetResult(req), (*c.Char)(unsafe.Pointer(&buffer[0])))
+		libuv.LoopClose(loop)
+	}
+}
+
+func onClose(req *libuv.Fs) {
+	// Check for errors
+	if libuv.FsGetResult(req) < 0 {
+		c.Fprintf(c.Stderr, c.Str("Error closing file: %s\n"), libuv.Strerror(libuv.Errno(libuv.FsGetResult(req))))
+	} else {
+		c.Printf(c.Str("\nFile closed successfully.\n"))
+	}
+	libuv.LoopClose(loop)
+}
+
+func cleanup() {
+	// Cleanup the requests
+	libuv.FsReqCleanup(&openReq)
+	libuv.FsReqCleanup(&readReq)
+	libuv.FsReqCleanup(&closeReq)
+	// Close the loop
+	libuv.LoopClose(loop)
+}

c/libuv/_demo/async_fs/example.txt

@@ -0,0 +1 @@
+123

c/libuv/_demo/echo_server/echo_server.go

@@ -0,0 +1,120 @@
+package main
+
+import (
+	"unsafe"
+
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/libuv"
+	"github.com/goplus/llgo/c/net"
+)
+
+var DEFAULT_PORT c.Int = 8080
+var DEFAULT_BACKLOG c.Int = 128
+
+var loop *libuv.Loop
+
+type WriteReq struct {
+	Req libuv.Write
+	Buf libuv.Buf
+}
+
+func main() {
+	// Initialize the default event loop
+	loop = libuv.DefaultLoop()
+
+	// Initialize a TCP server
+	var server libuv.Tcp
+	libuv.InitTcp(loop, &server)
+
+	// Set up the address to bind the server to
+	var addr net.SockaddrIn
+	libuv.Ip4Addr(c.Str("0.0.0.0"), DEFAULT_PORT, &addr)
+	c.Printf(c.Str("Listening on %s:%d\n"), c.Str("0.0.0.0"), DEFAULT_PORT)
+
+	// Bind the server to the specified address and port
+	(&server).Bind((*net.SockAddr)(c.Pointer(&addr)), 0)
+	res := (*libuv.Stream)(&server).Listen(DEFAULT_BACKLOG, OnNewConnection)
+	if res != 0 {
+		c.Fprintf(c.Stderr, c.Str("Listen error: %s\n"), libuv.Strerror((libuv.Errno(res))))
+		return
+	}
+
+	// Start listening for incoming connections
+	libuv.Run(loop, libuv.RUN_DEFAULT)
+}
+
+func FreeWriteReq(req *libuv.Write) {
+	wr := (*WriteReq)(c.Pointer(req))
+	// Free the buffer base and the WriteReq itself.
+	c.Free(c.Pointer(wr.Buf.Base))
+	c.Free(c.Pointer(wr))
+}
+
+func AllocBuffer(handle *libuv.Handle, suggestedSize uintptr, buf *libuv.Buf) {
+	// Allocate memory for the buffer based on the suggested size.
+	buf.Base = (*c.Char)(c.Malloc(suggestedSize))
+	buf.Len = suggestedSize
+}
+
+func EchoWrite(req *libuv.Write, status c.Int) {
+	if status != 0 {
+		c.Fprintf(c.Stderr, c.Str("Write error: %s\n"), libuv.Strerror((libuv.Errno(status))))
+	}
+	FreeWriteReq(req)
+}
+
+func EchoRead(client *libuv.Stream, nread c.Long, buf *libuv.Buf) {
+	if nread > 0 {
+		req := (*WriteReq)(c.Malloc(unsafe.Sizeof(WriteReq{})))
+		if req == nil {
+			c.Fprintf(c.Stderr, c.Str("Failed to allocate memory for write request\n"))
+			c.Free(c.Pointer(buf.Base))
+			return
+		}
+		// Initialize the buffer with the data read.
+		req.Buf = libuv.InitBuf(buf.Base, c.Uint(nread))
+		// Write the data back to the client.
+		(&req.Req).Write(client, &req.Buf, 1, EchoWrite)
+		return
+	}
+	if nread < 0 {
+		// Handle read errors and EOF.
+		if (libuv.Errno)(nread) != libuv.EOF {
+			c.Fprintf(c.Stderr, c.Str("Read error: %s\n"), libuv.Strerror((libuv.Errno)(nread)))
+		}
+		(*libuv.Handle)(c.Pointer(client)).Close(nil)
+	}
+	// Free the buffer if it's no longer needed.
+	if buf.Base != nil {
+		c.Free(c.Pointer(buf.Base))
+	}
+}
+
+func OnNewConnection(server *libuv.Stream, status c.Int) {
+	if status < 0 {
+		c.Fprintf(c.Stderr, c.Str("New connection error: %s\n"), libuv.Strerror(libuv.Errno(status)))
+		return
+	}
+
+	// Allocate memory for a new client.
+	client := (*libuv.Tcp)(c.Malloc(unsafe.Sizeof(libuv.Tcp{})))
+
+	if client == nil {
+		c.Fprintf(c.Stderr, c.Str("Failed to allocate memory for client\n"))
+		return
+	}
+
+	// Initialize the client TCP handle.
+	if libuv.InitTcp(loop, client) < 0 {
+		c.Fprintf(c.Stderr, c.Str("Failed to initialize client\n"))
+		c.Free(c.Pointer(client))
+		return
+	}
+
+	// Accept the new connection and start reading data.
+	if server.Accept((*libuv.Stream)(client)) == 0 {
+		(*libuv.Stream)(client).StartRead(AllocBuffer, EchoRead)
+	} else {
+		(*libuv.Handle)(c.Pointer(client)).Close(nil)
+	}
+}

c/libuv/error.go

@@ -0,0 +1,118 @@
+package libuv
+
+import (
+	_ "unsafe"
+
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/net"
+	"github.com/goplus/llgo/c/syscall"
+)
+
+const (
+	E2BIG           = Errno(syscall.E2BIG)
+	EACCES          = Errno(syscall.EACCES)
+	EADDRINUSE      = Errno(syscall.EADDRINUSE)
+	EADDRNOTAVAIL   = Errno(syscall.EADDRNOTAVAIL)
+	EAFNOSUPPORT    = Errno(syscall.EAFNOSUPPORT)
+	EAGAIN          = Errno(syscall.EAGAIN)
+	EALREADY        = Errno(syscall.EALREADY)
+	EBADF           = Errno(syscall.EBADF)
+	EBUSY           = Errno(syscall.EBUSY)
+	ECANCELED       = Errno(syscall.ECANCELED)
+	ECONNABORTED    = Errno(syscall.ECONNABORTED)
+	ECONNREFUSED    = Errno(syscall.ECONNREFUSED)
+	ECONNRESET      = Errno(syscall.ECONNRESET)
+	EDESTADDRREQ    = Errno(syscall.EDESTADDRREQ)
+	EEXIST          = Errno(syscall.EEXIST)
+	EFAULT          = Errno(syscall.EFAULT)
+	EFBIG           = Errno(syscall.EFBIG)
+	EHOSTUNREACH    = Errno(syscall.EHOSTUNREACH)
+	EINTR           = Errno(syscall.EINTR)
+	EINVAL          = Errno(syscall.EINVAL)
+	EIO             = Errno(syscall.EIO)
+	EISCONN         = Errno(syscall.EISCONN)
+	EISDIR          = Errno(syscall.EISDIR)
+	ELOOP           = Errno(syscall.ELOOP)
+	EMFILE          = Errno(syscall.EMFILE)
+	EMSGSIZE        = Errno(syscall.EMSGSIZE)
+	ENAMETOOLONG    = Errno(syscall.ENAMETOOLONG)
+	ENETDOWN        = Errno(syscall.ENETDOWN)
+	ENETUNREACH     = Errno(syscall.ENETUNREACH)
+	ENFILE          = Errno(syscall.ENFILE)
+	ENOBUFS         = Errno(syscall.ENOBUFS)
+	ENODEV          = Errno(syscall.ENODEV)
+	ENOENT          = Errno(syscall.ENOENT)
+	ENOMEM          = Errno(syscall.ENOMEM)
+	ENOPROTOOPT     = Errno(syscall.ENOPROTOOPT)
+	ENOSPC          = Errno(syscall.ENOSPC)
+	ENOSYS          = Errno(syscall.ENOSYS)
+	ENOTCONN        = Errno(syscall.ENOTCONN)
+	ENOTDIR         = Errno(syscall.ENOTDIR)
+	ENOTEMPTY       = Errno(syscall.ENOTEMPTY)
+	ENOTSOCK        = Errno(syscall.ENOTSOCK)
+	ENOTSUP         = Errno(syscall.ENOTSUP)
+	EOVERFLOW       = Errno(syscall.EOVERFLOW)
+	EPERM           = Errno(syscall.EPERM)
+	EPIPE           = Errno(syscall.EPIPE)
+	EPROTO          = Errno(syscall.EPROTO)
+	EPROTONOSUPPORT = Errno(syscall.EPROTONOSUPPORT)
+	EPROTOTYPE      = Errno(syscall.EPROTOTYPE)
+	ERANGE          = Errno(syscall.ERANGE)
+	EROFS           = Errno(syscall.EROFS)
+	ESHUTDOWN       = Errno(syscall.ESHUTDOWN)
+	ESPIPE          = Errno(syscall.ESPIPE)
+	ESRCH           = Errno(syscall.ESRCH)
+	ETIMEDOUT       = Errno(syscall.ETIMEDOUT)
+	ETXTBSY         = Errno(syscall.ETXTBSY)
+	EXDEV           = Errno(syscall.EXDEV)
+	ENXIO           = Errno(syscall.ENXIO)
+	EMLINK          = Errno(syscall.EMLINK)
+	EHOSTDOWN       = Errno(syscall.EHOSTDOWN)
+	ENOTTY          = Errno(syscall.ENOTTY)
+	//EFTYPE        = Errno(syscall.EFTYPE)
+	EILSEQ          = Errno(syscall.EILSEQ)
+	ESOCKTNOSUPPORT = Errno(syscall.ESOCKTNOSUPPORT)
+)
+
+const (
+	EAI_ADDRFAMILY = Errno(net.EAI_ADDRFAMILY)
+	EAI_AGAIN      = Errno(net.EAI_AGAIN)
+	EAI_BADFLAGS   = Errno(net.EAI_BADFLAGS)
+	EAI_BADHINTS   = Errno(net.EAI_BADHINTS)
+	EAI_FAIL       = Errno(net.EAI_FAIL)
+	EAI_FAMILY     = Errno(net.EAI_FAMILY)
+	EAI_MEMORY     = Errno(net.EAI_MEMORY)
+	EAI_NODATA     = Errno(net.EAI_NODATA)
+	EAI_NONAME     = Errno(net.EAI_NONAME)
+	EAI_OVERFLOW   = Errno(net.EAI_OVERFLOW)
+	EAI_PROTOCOL   = Errno(net.EAI_PROTOCOL)
+	EAI_SERVICE    = Errno(net.EAI_SERVICE)
+	EAI_SOCKTYPE   = Errno(net.EAI_SOCKTYPE)
+)
+
+const (
+	EAI_CANCELED Errno = -3003
+	ECHARSET     Errno = -4080
+	ENONET       Errno = -4056
+	UNKNOWN      Errno = -4094
+	EOF          Errno = -1
+	EREMOTEIO    Errno = -4030
+	ERRNO_MAX    Errno = EOF - 1
+)
+
+type Errno c.Int
+
+//go:linkname TranslateSysError C.uv_translate_sys_error
+func TranslateSysError(sysErrno c.Int) Errno
+
+//go:linkname Strerror C.uv_strerror
+func Strerror(err Errno) *c.Char
+
+//go:linkname StrerrorR C.uv_strerror_r
+func StrerrorR(err Errno, buf *c.Char, bufLen uintptr) *c.Char
+
+//go:linkname ErrName C.uv_err_name
+func ErrName(err Errno) *c.Char
+
+//go:linkname ErrNameR C.uv_err_name_r
+func ErrNameR(err Errno, buf *c.Char, bufLen uintptr) *c.Char

c/libuv/fs.go

@@ -0,0 +1,277 @@
+package libuv
+
+import (
+	_ "unsafe"
+
+	"github.com/goplus/llgo/c"
+)
+
+const (
+	FS_UNKNOWN   FsType = -1
+	FS_CUSTOM    FsType = 0
+	FS_OPEN      FsType = 1
+	FS_CLOSE     FsType = 2
+	FS_READ      FsType = 3
+	FS_WRITE     FsType = 4
+	FS_SENDFILE  FsType = 5
+	FS_STAT      FsType = 6
+	FS_LSTAT     FsType = 7
+	FS_FSTAT     FsType = 8
+	FS_FTRUNCATE FsType = 9
+	FS_UTIME     FsType = 10
+	FS_FUTIME    FsType = 11
+	FS_ACCESS    FsType = 12
+	FS_CHMOD     FsType = 13
+	FS_FCHMOD    FsType = 14
+	FS_FSYNC     FsType = 15
+	FS_FDATASYNC FsType = 16
+	FS_UNLINK    FsType = 17
+	FS_RMDIR     FsType = 18
+	FS_MKDIR     FsType = 19
+	FS_MKDTEMP   FsType = 20
+	FS_RENAME    FsType = 21
+	FS_SCANDIR   FsType = 22
+	FS_LINK      FsType = 23
+	FS_SYMLINK   FsType = 24
+	FS_READLINK  FsType = 25
+	FS_CHOWN     FsType = 26
+	FS_FCHOWN    FsType = 27
+	FS_REALPATH  FsType = 28
+	FS_COPYFILE  FsType = 29
+	FS_LCHOWN    FsType = 30
+	FS_OPENDIR   FsType = 31
+	FS_READDIR   FsType = 32
+	FS_CLOSEDIR  FsType = 33
+	FS_STATFS    FsType = 34
+	FS_MKSTEMP   FsType = 35
+	FS_LUTIME    FsType = 36
+)
+
+const (
+	DirentUnknown DirentType = iota
+	DirentFile
+	DirentDir
+	DirentLink
+	DirentFifo
+	DirentSocket
+	DirentChar
+	DirentBlock
+)
+
+type FsType c.Int
+
+type DirentType c.Int
+
+type UvFile c.Int
+
+// ----------------------------------------------
+
+/* Handle types. */
+
+type Fs struct {
+	Unused [0]byte
+}
+
+type FsEvent struct {
+	Unused [0]byte
+}
+
+type FsPoll struct {
+	Unused [0]byte
+}
+
+type Dirent struct {
+	Name *c.Char
+	Type DirentType
+}
+
+type File struct {
+	Loop *Loop
+	Req  *Fs
+}
+
+type Stat struct {
+	Unused [0]byte
+}
+
+// ----------------------------------------------
+
+/* Function type */
+
+// llgo:type C
+type FsCb func(req *Fs)
+
+// llgo:type C
+type FsEventCb func(handle *FsEvent, filename *c.Char, events c.Int, status c.Int)
+
+// llgo:type C
+type FsPollCb func(handle *FsPoll, status c.Int, events c.Int)
+
+// ----------------------------------------------
+
+/* Fs related function and method */
+
+//go:linkname FsGetType C.uv_fs_get_type
+func FsGetType(req *Fs) *FsType
+
+//go:linkname FsGetPath C.uv_fs_get_path
+func FsGetPath(req *Fs) *c.Char
+
+//go:linkname FsGetResult C.uv_fs_get_result
+func FsGetResult(req *Fs) c.Int
+
+//go:linkname FsGetPtr C.uv_fs_get_ptr
+func FsGetPtr(req *Fs) c.Pointer
+
+//go:linkname FsGetSystemError C.uv_fs_get_system_error
+func FsGetSystemError(req *Fs) c.Int
+
+//go:linkname FsGetStatBuf C.uv_fs_get_statbuf
+func FsGetStatBuf(req *Fs) *Stat
+
+//go:linkname FsReqCleanup C.uv_fs_req_cleanup
+func FsReqCleanup(req *Fs)
+
+//go:linkname DefaultLoop C.uv_default_loop
+func DefaultLoop() *Loop
+
+//go:linkname FsOpen C.uv_fs_open
+func FsOpen(loop *Loop, req *Fs, path *c.Char, flags c.Int, mode c.Int, cb FsCb) c.Int
+
+//go:linkname FsClose C.uv_fs_close
+func FsClose(loop *Loop, req *Fs, file UvFile, cb FsCb) c.Int
+
+//go:linkname FsRead C.uv_fs_read
+func FsRead(loop *Loop, req *Fs, file UvFile, bufs *Buf, nbufs c.Uint, offset c.LongLong, cb FsCb) c.Int
+
+//go:linkname FsWrite C.uv_fs_write
+func FsWrite(loop *Loop, req *Fs, file UvFile, bufs *Buf, nbufs c.Uint, offset c.LongLong, cb FsCb) c.Int
+
+//go:linkname FsUnlink C.uv_fs_unlink
+func FsUnlink(loop *Loop, req *Fs, path *c.Char, cb FsCb) c.Int
+
+//go:linkname FsMkdir C.uv_fs_mkdir
+func FsMkdir(loop *Loop, req *Fs, path *c.Char, mode c.Int, cb FsCb) c.Int
+
+//go:linkname FsMkdtemp C.uv_fs_mkdtemp
+func FsMkdtemp(loop *Loop, req *Fs, tpl *c.Char, cb FsCb) c.Int
+
+//go:linkname FsMkStemp C.uv_fs_mkstemp
+func FsMkStemp(loop *Loop, req *Fs, tpl *c.Char, cb FsCb) c.Int
+
+//go:linkname FsRmdir C.uv_fs_rmdir
+func FsRmdir(loop *Loop, req *Fs, path *c.Char, cb FsCb) c.Int
+
+//go:linkname FsStat C.uv_fs_stat
+func FsStat(loop *Loop, req *Fs, path *c.Char, cb FsCb) c.Int
+
+//go:linkname FsFstat C.uv_fs_fstat
+func FsFstat(loop *Loop, req *Fs, file UvFile, cb FsCb) c.Int
+
+//go:linkname FsRename C.uv_fs_rename
+func FsRename(loop *Loop, req *Fs, path *c.Char, newPath *c.Char, cb FsCb) c.Int
+
+//go:linkname FsFsync C.uv_fs_fsync
+func FsFsync(loop *Loop, req *Fs, file UvFile, cb FsCb) c.Int
+
+//go:linkname FsFdatasync C.uv_fs_fdatasync
+func FsFdatasync(loop *Loop, req *Fs, file UvFile, cb FsCb) c.Int
+
+//go:linkname FsFtruncate C.uv_fs_ftruncate
+func FsFtruncate(loop *Loop, req *Fs, file UvFile, offset c.LongLong, cb FsCb) c.Int
+
+//go:linkname FsSendfile C.uv_fs_sendfile
+func FsSendfile(loop *Loop, req *Fs, outFd c.Int, inFd c.Int, inOffset c.LongLong, length c.Int, cb FsCb) c.Int
+
+//go:linkname FsAccess C.uv_fs_access
+func FsAccess(loop *Loop, req *Fs, path *c.Char, flags c.Int, cb FsCb) c.Int
+
+//go:linkname FsChmod C.uv_fs_chmod
+func FsChmod(loop *Loop, req *Fs, path *c.Char, mode c.Int, cb FsCb) c.Int
+
+//go:linkname FsFchmod C.uv_fs_fchmod
+func FsFchmod(loop *Loop, req *Fs, file UvFile, mode c.Int, cb FsCb) c.Int
+
+//go:linkname FsUtime C.uv_fs_utime
+func FsUtime(loop *Loop, req *Fs, path *c.Char, atime c.Int, mtime c.Int, cb FsCb) c.Int
+
+//go:linkname FsFutime C.uv_fs_futime
+func FsFutime(loop *Loop, req *Fs, file UvFile, atime c.Int, mtime c.Int, cb FsCb) c.Int
+
+//go:linkname FsLutime C.uv_fs_lutime
+func FsLutime(loop *Loop, req *Fs, path *c.Char, atime c.Int, mtime c.Int, cb FsCb) c.Int
+
+//go:linkname FsLink C.uv_fs_link
+func FsLink(loop *Loop, req *Fs, path *c.Char, newPath *c.Char, cb FsCb) c.Int
+
+//go:linkname FsSymlink C.uv_fs_symlink
+func FsSymlink(loop *Loop, req *Fs, path *c.Char, newPath *c.Char, flags c.Int, cb FsCb) c.Int
+
+//go:linkname FsReadlink C.uv_fs_read
+func FsReadlink(loop *Loop, req *Fs, path *c.Char, cb FsCb) c.Int
+
+//go:linkname FsRealpath C.uv_fs_realpath
+func FsRealpath(loop *Loop, req *Fs, path *c.Char, cb FsCb) c.Int
+
+//go:linkname FsCopyfile C.uv_fs_copyfile
+func FsCopyfile(loop *Loop, req *Fs, path *c.Char, newPath *c.Char, flags c.Int, cb FsCb) c.Int
+
+//go:linkname FsScandir C.uv_fs_scandir
+func FsScandir(loop *Loop, req *Fs, path *c.Char, flags c.Int, cb FsCb) c.Int
+
+//go:linkname FsScandirNext C.uv_fs_scandir_next
+func FsScandirNext(req *Fs, ent *Dirent) c.Int
+
+//go:linkname FsOpenDir C.uv_fs_opendir
+func FsOpenDir(loop *Loop, req *Fs, path *c.Char, cb FsCb) c.Int
+
+//go:linkname FsReaddir C.uv_fs_readdir
+func FsReaddir(loop *Loop, req *Fs, dir c.Int, cb FsCb) c.Int
+
+//go:linkname FsCloseDir C.uv_fs_closedir
+func FsCloseDir(loop *Loop, req *Fs) c.Int
+
+//go:linkname FsStatfs C.uv_fs_statfs
+func FsStatfs(loop *Loop, req *Fs, path *c.Char, cb FsCb) c.Int
+
+//go:linkname FsChown C.uv_fs_chown
+func FsChown(loop *Loop, req *Fs, path *c.Char, uid c.Int, gid c.Int, cb FsCb) c.Int
+
+//go:linkname FsFchown C.uv_fs_fchown
+func FsFchown(loop *Loop, req *Fs, file UvFile, uid c.Int, gid c.Int, cb FsCb) c.Int
+
+//go:linkname FsLchown C.uv_fs_lchown
+func FsLchown(loop *Loop, req *Fs, path *c.Char, uid c.Int, gid c.Int, cb FsCb) c.Int
+
+//go:linkname FsLstat C.uv_fs_lstat
+func FsLstat(loop *Loop, req *Fs, path *c.Char, cb FsCb) c.Int
+
+//go:linkname FsEventInit C.uv_fs_event_init
+func FsEventInit(loop *Loop, handle *FsEvent) c.Int
+
+//go:linkname FsEventStart C.uv_fs_event_start
+func FsEventStart(handle *FsEvent, cb FsEventCb, path *c.Char, flags c.Int) c.Int
+
+//go:linkname FsEventStop C.uv_fs_event_stop
+func FsEventStop(handle *FsEvent) c.Int
+
+//go:linkname FsEventClose C.uv_fs_event_close
+func FsEventClose(handle *FsEvent) c.Int
+
+//go:linkname FsEventGetpath C.uv_fs_event_getpath
+func FsEventGetpath(handle *FsEvent) *c.Char
+
+//go:linkname FsPollInit C.uv_fs_poll_init
+func FsPollInit(loop *Loop, handle *FsPoll) c.Int
+
+//go:linkname FsPollStart C.uv_fs_poll_start
+func FsPollStart(handle *FsPoll, cb FsPollCb, path *c.Char, interval uint) c.Int
+
+//go:linkname FsPollStop C.uv_fs_poll_stop
+func FsPollStop(handle *FsPoll) c.Int
+
+//go:linkname FsPollClose C.uv_fs_poll_close
+func FsPollClose(handle *FsPoll) c.Int
+
+//go:linkname FsPollGetPath C.uv_fs_poll_getpath
+func FsPollGetPath(handle *FsPoll) *c.Char

c/libuv/libuv.go

@@ -0,0 +1,488 @@
+package libuv
+
+import (
+	_ "unsafe"
+
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/net"
+)
+
+const (
+	LLGoPackage = "link: $(pkg-config --libs libuv); -luv"
+)
+
+// ----------------------------------------------
+const (
+	RUN_DEFAULT RunMode = iota
+	RUN_ONCE
+	RUN_NOWAIT
+)
+
+const (
+	LOOP_BLOCK_SIGNAL LoopOption = iota
+	METRICS_IDLE_TIME
+)
+
+const (
+	UV_LEAVE_GROUP Membership = iota
+	UV_JOIN_GROUP
+)
+
+const (
+	UNKNOWN_HANDLE HandleType = iota
+	ASYNC
+	CHECK
+	FS_EVENT
+	FS_POLL
+	HANDLE
+	IDLE
+	NAMED_PIPE
+	POLL
+	PREPARE
+	PROCESS
+	STREAM
+	TCP
+	TIMER
+	TTY
+	UDP
+	SIGNAL
+	FILE
+	HANDLE_TYPE_MAX
+)
+
+const (
+	UNKNOWN_REQ ReqType = iota
+	REQ
+	CONNECT
+	WRITE
+	SHUTDOWN
+	UDP_SEND
+	FS
+	WORK
+	GETADDRINFO
+	GETNAMEINFO
+	RANDOM
+	REQ_TYPE_PRIVATE
+	REQ_TYPE_MAX
+)
+
+const (
+	READABLE PollEvent = 1 << iota
+	WRITABLE
+	DISCONNECT
+	PRIPRIORITIZED
+)
+
+type RunMode c.Int
+
+type LoopOption c.Int
+
+type Membership c.Int
+
+type HandleType c.Int
+
+type ReqType c.Int
+
+type OsSock c.Int
+
+type OsFd c.Int
+
+type PollEvent c.Int
+
+// ----------------------------------------------
+
+/* Handle types. */
+
+type Loop struct {
+	Unused [0]byte
+}
+
+type Handle struct {
+	Unused [96]byte
+}
+
+type Dir struct {
+	Unused [0]byte
+}
+
+type Stream struct {
+	Unused [264]byte
+}
+
+type Pipe struct {
+	Unused [0]byte
+}
+
+type Tty struct {
+	Unused [0]byte
+}
+
+type Poll struct {
+	Unused [0]byte
+}
+
+type Prepare struct {
+	Unused [0]byte
+}
+
+type Check struct {
+	Unused [0]byte
+}
+
+type Idle struct {
+	Unused [0]byte
+}
+
+type Async struct {
+	Unused [0]byte
+}
+
+type Process struct {
+	Unused [0]byte
+}
+
+/* Request types. */
+
+type Req struct {
+	Unused [0]byte
+}
+
+type GetAddrInfo struct {
+	Unused [0]byte
+}
+
+type GetNameInfo struct {
+	Unused [0]byte
+}
+
+type Shutdown struct {
+	Unused [0]byte
+}
+
+type Write struct {
+	Unused [192]byte
+}
+
+type Connect struct {
+	Unused [0]byte
+}
+
+type Buf struct {
+	Base *c.Char
+	Len  uintptr
+} // ----------------------------------------------
+
+/* Function type */
+
+// llgo:type C
+type MallocFunc func(size uintptr) c.Pointer
+
+// llgo:type C
+type ReallocFunc func(ptr c.Pointer, size uintptr) c.Pointer
+
+// llgo:type C
+type CallocFunc func(count uintptr, size uintptr) c.Pointer
+
+// llgo:type C
+type FreeFunc func(ptr c.Pointer)
+
+// llgo:type C
+type AllocCb func(handle *Handle, suggestedSize uintptr, buf *Buf)
+
+// llgo:type C
+type ReadCb func(stream *Stream, nread c.Long, buf *Buf)
+
+// llgo:type C
+type WriteCb func(req *Write, status c.Int)
+
+// llgo:type C
+type GetaddrinfoCb func(req *GetAddrInfo, status c.Int, res *net.AddrInfo)
+
+// llgo:type C
+type GetnameinfoCb func(req *GetNameInfo, status c.Int, hostname *c.Char, service *c.Char)
+
+// llgo:type C
+type ConnectionCb func(server *Stream, status c.Int)
+
+// llgo:type C
+type ShutdownCb func(req *Shutdown, status c.Int)
+
+// llgo:type C
+type WalkCb func(handle *Handle, arg c.Pointer)
+
+// llgo:type C
+type PollCb func(handle *Poll, status c.Int, events c.Int)
+
+// ----------------------------------------------
+
+//go:linkname Version C.uv_version
+func Version() c.Uint
+
+//go:linkname VersionString C.uv_version_string
+func VersionString() *c.Char
+
+//go:linkname LibraryShutdown C.uv_library_shutdown
+func LibraryShutdown()
+
+//go:linkname ReplaceAllocator C.uv_replace_allocator
+func ReplaceAllocator(mallocFunc MallocFunc, reallocFunc ReallocFunc, callocFunc CallocFunc, freeFunc FreeFunc) c.Int
+
+// ----------------------------------------------
+
+// llgo:link (*Shutdown).Shutdown C.uv_shutdown
+func (shutdown *Shutdown) Shutdown(stream *Stream, shutdownCb ShutdownCb) c.Int {
+	return 0
+}
+
+// ----------------------------------------------
+
+/* Handle related function and method */
+
+// llgo:link (*Handle).Ref C.uv_ref
+func (handle *Handle) Ref() {}
+
+// llgo:link (*Handle).Unref C.uv_unref
+func (handle *Handle) Unref() {}
+
+// llgo:link (*Handle).HasRef C.uv_has_ref
+func (handle *Handle) HasRef() c.Int {
+	return 0
+}
+
+//go:linkname HandleSize C.uv_handle_size
+func HandleSize(handleType HandleType) uintptr
+
+// llgo:link (*Handle).GetType C.uv_handle_get_type
+func (handle *Handle) GetType() HandleType {
+	return 0
+}
+
+//go:linkname HandleTypeName C.uv_handle_type_name
+func HandleTypeName(handleType HandleType) *c.Char
+
+// llgo:link (*Handle).GetData C.uv_handle_get_data
+func (handle *Handle) GetData() c.Pointer {
+	return nil
+}
+
+// llgo:link (*Handle).GetLoop C.uv_handle_get_loop
+func (handle *Handle) GetLoop() *Loop {
+	return nil
+}
+
+// llgo:link (*Handle).SetData C.uv_handle_set_data
+func (handle *Handle) SetData(data c.Pointer) {}
+
+// llgo:link (*Handle).IsActive C.uv_is_active
+func (handle *Handle) IsActive() c.Int {
+	return 0
+}
+
+// llgo:link (*Handle).Close C.uv_close
+func (handle *Handle) Close(closeCb CloseCb) {}
+
+// llgo:link (*Handle).SendBufferSize C.uv_send_buffer_size
+func (handle *Handle) SendBufferSize(value *c.Int) c.Int {
+	return 0
+}
+
+// llgo:link (*Handle).RecvBufferSize C.uv_recv_buffer_size
+func (handle *Handle) RecvBufferSize(value *c.Int) c.Int {
+	return 0
+}
+
+// llgo:link (*Handle).Fileno C.uv_fileno
+func (handle *Handle) Fileno(fd *OsFd) c.Int {
+	return 0
+}
+
+//go:linkname UvPipe C.uv_pipe
+func UvPipe(fds [2]UvFile, readFlags c.Int, writeFlags c.Int) c.Int {
+	return 0
+}
+
+//go:linkname Socketpair C.uv_socketpair
+func Socketpair(_type c.Int, protocol c.Int, socketVector [2]OsSock, flag0 c.Int, flag1 c.Int) c.Int {
+	return 0
+}
+
+// llgo:link (*Handle).IsClosing C.uv_is_closing
+func (handle *Handle) IsClosing() c.Int {
+	return 0
+}
+
+// ----------------------------------------------
+
+/* Req related function and method */
+
+//go:linkname ReqSize C.uv_req_size
+func ReqSize(reqType ReqType) uintptr
+
+// llgo:link (*Req).GetData C.uv_req_get_data
+func (req *Req) GetData() c.Pointer {
+	return nil
+}
+
+// llgo:link (*Req).SetData C.uv_handle_set_data
+func (req *Req) SetData(data c.Pointer) {}
+
+// llgo:link (*Req).GetType C.uv_req_get_type
+func (req *Req) GetType() ReqType {
+	return 0
+}
+
+//go:linkname TypeName C.uv_req_type_name
+func TypeName(reqType ReqType) *c.Char
+
+// ----------------------------------------------
+
+/* Stream related function and method */
+
+// llgo:link (*Stream).GetWriteQueueSize C.uv_stream_get_write_queue_size
+func (stream *Stream) GetWriteQueueSize() uintptr {
+	return 0
+}
+
+// llgo:link (*Stream).Listen C.uv_listen
+func (stream *Stream) Listen(backlog c.Int, connectionCb ConnectionCb) c.Int {
+	return 0
+}
+
+// llgo:link (*Stream).Accept C.uv_accept
+func (server *Stream) Accept(client *Stream) c.Int {
+	return 0
+}
+
+// llgo:link (*Stream).StartRead C.uv_read_start
+func (stream *Stream) StartRead(allocCb AllocCb, readCb ReadCb) c.Int {
+	return 0
+}
+
+// llgo:link (*Stream).StopRead C.uv_read_stop
+func (stream *Stream) StopRead() c.Int {
+	return 0
+}
+
+// llgo:link (*Write).Write C.uv_write
+func (req *Write) Write(stream *Stream, bufs *Buf, nbufs c.Uint, writeCb WriteCb) c.Int {
+	return 0
+}
+
+// llgo:link (*Write).Write2 C.uv_write2
+func (req *Write) Write2(stream *Stream, bufs *Buf, nbufs c.Uint, sendStream *Stream, writeCb WriteCb) c.Int {
+	return 0
+}
+
+// llgo:link (*Stream).TryWrite C.uv_try_write
+func (stream *Stream) TryWrite(bufs *Buf, nbufs c.Uint) c.Int {
+	return 0
+}
+
+// llgo:link (*Stream).TryWrite2 C.uv_try_write2
+func (stream *Stream) TryWrite2(bufs *Buf, nbufs c.Uint, sendStream *Stream) c.Int {
+	return 0
+}
+
+// llgo:link (*Stream).IsReadable C.uv_is_readable
+func (stream *Stream) IsReadable() c.Int {
+	return 0
+}
+
+// llgo:link (*Stream).IsWritable C.uv_is_writable
+func (stream *Stream) IsWritable() c.Int {
+	return 0
+}
+
+// llgo:link (*Stream).SetBlocking C.uv_stream_set_blocking
+func (stream *Stream) SetBlocking(blocking c.Int) c.Int {
+	return 0
+}
+
+// ----------------------------------------------
+
+/* Loop related functions and method. */
+
+//go:linkname LoopSize C.uv_loop_size
+func LoopSize() uintptr
+
+//go:linkname Run C.uv_run
+func Run(loop *Loop, mode RunMode) c.Int
+
+//go:linkname LoopAlive C.uv_loop_alive
+func LoopAlive(loop *Loop) c.Int
+
+//go:linkname LoopClose C.uv_loop_close
+func LoopClose(loop *Loop) c.Int
+
+//go:linkname LoopConfigure C.uv_loop_configure
+func LoopConfigure(loop *Loop, option LoopOption, arg c.Int) c.Int
+
+//go:linkname LoopDefault C.uv_default_loop
+func LoopDefault() *Loop
+
+//go:linkname LoopDelete C.uv_loop_delete
+func LoopDelete(loop *Loop) c.Int
+
+//go:linkname LoopFork C.uv_loop_fork
+func LoopFork(loop *Loop) c.Int
+
+//go:linkname LoopInit C.uv_loop_init
+func LoopInit(loop *Loop) c.Int
+
+//go:linkname LoopNew C.uv_loop_new
+func LoopNew() *Loop
+
+//go:linkname LoopNow C.uv_now
+func LoopNow(loop *Loop) c.UlongLong
+
+//go:linkname LoopUpdateTime C.uv_update_time
+func LoopUpdateTime(loop *Loop)
+
+//go:linkname LoopBackendFd C.uv_backend_fd
+func LoopBackendFd(loop *Loop) c.Int
+
+//go:linkname LoopBackendTimeout C.uv_backend_timeout
+func LoopBackendTimeout(loop *Loop) c.Int
+
+//go:linkname LoopWalk C.uv_walk
+func LoopWalk(loop *Loop, walkCb WalkCb, arg c.Pointer)
+
+// ----------------------------------------------
+
+/* Buf related functions and method. */
+
+//go:linkname InitBuf C.uv_buf_init
+func InitBuf(base *c.Char, len c.Uint) Buf
+
+// ----------------------------------------------
+
+/* Poll related function and method */
+
+//go:linkname PollInit C.uv_poll_init
+func PollInit(loop *Loop, handle *Poll, fd OsFd) c.Int
+
+//go:linkname PollStart C.uv_poll_start
+func PollStart(handle *Poll, events c.Int, cb PollCb) c.Int
+
+//go:linkname PollStop C.uv_poll_stop
+func PollStop(handle *Poll) c.Int
+
+//go:linkname PollInitSocket C.uv_poll_init_socket
+func PollInitSocket(loop *Loop, handle *Poll, socket c.Int) c.Int
+
+// ----------------------------------------------
+
+/* Getaddrinfo related function and method */
+
+//go:linkname Getaddrinfo C.uv_getaddrinfo
+func Getaddrinfo(loop *Loop, req *GetAddrInfo, getaddrinfoCb GetaddrinfoCb, node *c.Char, service *c.Char, hints *net.AddrInfo) c.Int
+
+//go:linkname Freeaddrinfo C.uv_freeaddrinfo
+func Freeaddrinfo(addrInfo *net.AddrInfo)
+
+// ----------------------------------------------
+
+/* Getnameinfo related function and method */
+
+//go:linkname Getnameinfo C.uv_getnameinfo
+func Getnameinfo(loop *Loop, req *GetNameInfo, getnameinfoCb GetnameinfoCb, addr *net.SockAddr, flags c.Int) c.Int

c/libuv/net.go

@@ -0,0 +1,275 @@
+package libuv
+
+import (
+	_ "unsafe"
+
+	"github.com/goplus/llgo/c"
+	"github.com/goplus/llgo/c/net"
+)
+
+const (
+	/* Used with uv_tcp_bind, when an IPv6 address is used. */
+	TCP_IPV6ONLY TcpFlags = 1
+)
+
+/*
+ * UDP support.
+ */
+const (
+	/* Disables dual stack mode. */
+	UDP_IPV6ONLY UdpFlags = 1
+	/*
+	 * Indicates message was truncated because read buffer was too small. The
+	 * remainder was discarded by the OS. Used in uv_udp_recv_cb.
+	 */
+	UDP_PARTIAL UdpFlags = 2
+	/*
+	 * Indicates if SO_REUSEADDR will be set when binding the handle.
+	 * This sets the SO_REUSEPORT socket flag on the BSDs and OS X. On other
+	 * Unix platforms, it sets the SO_REUSEADDR flag.  What that means is that
+	 * multiple threads or processes can bind to the same address without error
+	 * (provided they all set the flag) but only the last one to bind will receive
+	 * any traffic, in effect "stealing" the port from the previous listener.
+	 */
+	UDP_REUSEADDR UdpFlags = 4
+	/*
+	 * Indicates that the message was received by recvmmsg, so the buffer provided
+	 * must not be freed by the recv_cb callback.
+	 */
+	UDP_MMSG_CHUNK UdpFlags = 8
+	/*
+	 * Indicates that the buffer provided has been fully utilized by recvmmsg and
+	 * that it should now be freed by the recv_cb callback. When this flag is set
+	 * in uv_udp_recv_cb, nread will always be 0 and addr will always be NULL.
+	 */
+	UDP_MMSG_FREE UdpFlags = 16
+	/*
+	 * Indicates if IP_RECVERR/IPV6_RECVERR will be set when binding the handle.
+	 * This sets IP_RECVERR for IPv4 and IPV6_RECVERR for IPv6 UDP sockets on
+	 * Linux. This stops the Linux kernel from suppressing some ICMP error
+	 * messages and enables full ICMP error reporting for faster failover.
+	 * This flag is no-op on platforms other than Linux.
+	 */
+	UDP_LINUX_RECVERR UdpFlags = 32
+	/*
+	 * Indicates that recvmmsg should be used, if available.
+	 */
+	UDP_RECVMMSG UdpFlags = 256
+)
+
+type TcpFlags c.Int
+
+type UdpFlags c.Int
+
+// ----------------------------------------------
+
+/* Handle types. */
+
+type Tcp struct {
+	Unused [264]byte
+}
+
+type Udp struct {
+	Unused [0]byte
+}
+
+/* Request types. */
+
+type UdpSend struct {
+	Unused [0]byte
+}
+
+// ----------------------------------------------
+
+/* Function type */
+
+// llgo:type C
+type CloseCb func(handle *Handle)
+
+// llgo:type C
+type ConnectCb func(req *Connect, status c.Int)
+
+// llgo:type C
+type UdpSendCb func(req *UdpSend, status c.Int)
+
+// llgo:type C
+type UdpRecvCb func(handle *Udp, nread c.Long, buf *Buf, addr *net.SockAddr, flags c.Uint)
+
+// ----------------------------------------------
+
+/* Tcp related function and method */
+
+//go:linkname InitTcp C.uv_tcp_init
+func InitTcp(loop *Loop, tcp *Tcp) c.Int
+
+//go:linkname InitTcpEx C.uv_tcp_init_ex
+func InitTcpEx(loop *Loop, tcp *Tcp, flags c.Uint) c.Int
+
+// llgo:link (*Tcp).Open C.uv_tcp_open
+func (tcp *Tcp) Open(sock OsSock) c.Int {
+	return 0
+}
+
+// llgo:link (*Tcp).Nodelay C.uv_tcp_nodelay
+func (tcp *Tcp) Nodelay(enable c.Int) c.Int {
+	return 0
+}
+
+// llgo:link (*Tcp).KeepAlive C.uv_tcp_keepalive
+func (tcp *Tcp) KeepAlive(enable c.Int, delay c.Uint) c.Int {
+	return 0
+}
+
+// llgo:link (*Tcp).SimultaneousAccepts C.uv_tcp_simultaneous_accepts
+func (tcp *Tcp) SimultaneousAccepts(enable c.Int) c.Int {
+	return 0
+}
+
+// llgo:link (*Tcp).Bind C.uv_tcp_bind
+func (tcp *Tcp) Bind(addr *net.SockAddr, flags c.Uint) c.Int {
+	return 0
+}
+
+// llgo:link (*Tcp).Getsockname C.uv_tcp_getsockname
+func (tcp *Tcp) Getsockname(name *net.SockAddr, nameLen *c.Int) c.Int {
+	return 0
+}
+
+// llgo:link (*Tcp).Getpeername C.uv_tcp_getpeername
+func (tcp *Tcp) Getpeername(name *net.SockAddr, nameLen *c.Int) c.Int {
+	return 0
+}
+
+// llgo:link (*Tcp).CloseReset C.uv_tcp_close_reset
+func (tcp *Tcp) CloseReset(closeCb CloseCb) c.Int {
+	return 0
+}
+
+//go:linkname TcpConnect C.uv_tcp_connect
+func TcpConnect(req *Connect, tcp *Tcp, addr *net.SockAddr, connectCb ConnectCb) c.Int
+
+// ----------------------------------------------
+
+/* Udp related function and method */
+
+//go:linkname InitUdp C.uv_udp_init
+func InitUdp(loop *Loop, udp *Udp) c.Int
+
+//go:linkname InitUdpEx C.uv_udp_init_ex
+func InitUdpEx(loop *Loop, udp *Udp, flags c.Uint) c.Int
+
+// llgo:link (*Udp).Open C.uv_udp_open
+func (udp *Udp) Open(sock OsSock) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).Bind C.uv_udp_bind
+func (udp *Udp) Bind(addr *net.SockAddr, flags c.Uint) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).Connect C.uv_udp_connect
+func (udp *Udp) Connect(addr *net.SockAddr) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).Getpeername C.uv_udp_getpeername
+func (udp *Udp) Getpeername(name *net.SockAddr, nameLen *c.Int) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).Getsockname C.uv_udp_getsockname
+func (udp *Udp) Getsockname(name *net.SockAddr, nameLen *c.Int) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).SetMembership C.uv_udp_set_membership
+func (udp *Udp) SetMembership(multicastAddr *c.Char, interfaceAddr *c.Char, membership Membership) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).SourceMembership C.uv_udp_set_source_membership
+func (udp *Udp) SourceMembership(multicastAddr *c.Char, interfaceAddr *c.Char, sourceAddr *c.Char, membership Membership) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).SetMulticastLoop C.uv_udp_set_multicast_loop
+func (udp *Udp) SetMulticastLoop(on c.Int) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).SetMulticastTTL C.uv_udp_set_multicast_ttl
+func (udp *Udp) SetMulticastTTL(ttl c.Int) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).SetMulticastInterface C.uv_udp_set_multicast_interface
+func (udp *Udp) SetMulticastInterface(interfaceAddr *c.Char) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).SetBroadcast C.uv_udp_set_broadcast
+func (udp *Udp) SetBroadcast(on c.Int) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).SetTTL C.uv_udp_set_ttl
+func (udp *Udp) SetTTL(ttl c.Int) c.Int {
+	return 0
+}
+
+//go:linkname Send C.uv_udp_send
+func Send(req *UdpSend, udp *Udp, bufs *Buf, nbufs c.Uint, addr *net.SockAddr, sendCb UdpSendCb) c.Int
+
+// llgo:link (*Udp).TrySend C.uv_udp_try_send
+func (udp *Udp) TrySend(bufs *Buf, nbufs c.Uint, addr *net.SockAddr) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).StartRecv C.uv_udp_recv_start
+func (udp *Udp) StartRecv(allocCb AllocCb, recvCb UdpRecvCb) c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).UsingRecvmmsg C.uv_udp_using_recvmmsg
+func (udp *Udp) UsingRecvmmsg() c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).StopRecv C.uv_udp_recv_stop
+func (udp *Udp) StopRecv() c.Int {
+	return 0
+}
+
+// llgo:link (*Udp).GetSendQueueSize C.uv_udp_get_send_queue_size
+func (udp *Udp) GetSendQueueSize() uintptr {
+	return 0
+}
+
+// llgo:link (*Udp).GetSendQueueCount C.uv_udp_get_send_queue_count
+func (udp *Udp) GetSendQueueCount() uintptr {
+	return 0
+}
+
+// ----------------------------------------------
+
+//go:linkname Ip4Addr C.uv_ip4_addr
+func Ip4Addr(ip *c.Char, port c.Int, addr *net.SockaddrIn) c.Int
+
+//go:linkname Ip6Addr C.uv_ip6_addr
+func Ip6Addr(ip *c.Char, port c.Int, addr *net.SockaddrIn6) c.Int
+
+//go:linkname Ip4Name C.uv_ip4_name
+func Ip4Name(src *net.SockaddrIn, dst *c.Char, size uintptr) c.Int
+
+//go:linkname Ip6Name C.uv_ip6_name
+func Ip6Name(src *net.SockaddrIn6, dst *c.Char, size uintptr) c.Int
+
+//go:linkname IpName C.uv_ip_name
+func IpName(src *net.SockAddr, dst *c.Char, size uintptr) c.Int
+
+//go:linkname InetNtop C.uv_inet_ntop
+func InetNtop(af c.Int, src c.Pointer, dst *c.Char, size uintptr) c.Int
+
+//go:linkname InetPton C.uv_inet_pton
+func InetPton(af c.Int, src *c.Char, dst c.Pointer) c.Int

c/libuv/signal.go

@@ -0,0 +1,33 @@
+package libuv
+
+import (
+	_ "unsafe"
+
+	"github.com/goplus/llgo/c"
+)
+
+/* Handle types. */
+
+type Signal struct {
+	Unused [0]byte
+}
+
+// ----------------------------------------------
+
+/* Function type */
+
+// llgo:type C
+type SignalCb func(handle *Signal, sigNum c.Int)
+
+// ----------------------------------------------
+
+/* Signal related functions and method. */
+
+//go:linkname SignalInit C.uv_signal_init
+func SignalInit(loop *Loop, handle *Signal) c.Int
+
+//go:linkname SignalStart C.uv_signal_start
+func SignalStart(handle *Signal, cb SignalCb, signum c.Int) c.Int
+
+//go:linkname SignalStartOneshot C.uv_signal_start_oneshot
+func SignalStartOneshot(handle *Signal, cb SignalCb, signum c.Int) c.Int

c/libuv/timer.go

@@ -0,0 +1,55 @@
+package libuv
+
+import (
+	_ "unsafe"
+
+	"github.com/goplus/llgo/c"
+)
+
+// ----------------------------------------------
+
+/* Handle types. */
+
+type Timer struct {
+	Unused [0]byte
+}
+
+// ----------------------------------------------
+
+// llgo:type Cgit
+type TimerCb func(timer *Timer)
+
+// ----------------------------------------------
+
+/* Timer related function and method */
+
+//go:linkname InitTimer C.uv_timer_init
+func InitTimer(loop *Loop, timer *Timer) c.Int
+
+// llgo:link (*Timer).Start C.uv_timer_start
+func (timer *Timer) Start(cb TimerCb, timeout uint64, repeat uint64) c.Int {
+	return 0
+}
+
+// llgo:link (*Timer).Stop C.uv_timer_stop
+func (timer *Timer) Stop() c.Int {
+	return 0
+}
+
+// llgo:link (*Timer).Again C.uv_timer_again
+func (timer *Timer) Again() c.Int {
+	return 0
+}
+
+// llgo:link (*Timer).SetRepeat C.uv_timer_set_repeat
+func (timer *Timer) SetRepeat(repeat uint64) {}
+
+// llgo:link (*Timer).GetRepeat C.uv_timer_get_repeat
+func (timer *Timer) GetRepeat() uint64 {
+	return 0
+}
+
+// llgo:link (*Timer).GetDueIn C.uv_timer_get_due_in
+func (timer *Timer) GetDueIn() uint64 {
+	return 0
+}

c/llama2/README.md

@@ -0,0 +1,10 @@
+LLGo wrapper of karpathy/llama2.c
+=====
+
+## How to update source to llgo
+
+```
+git submodule init
+git submodule update
+cp llama2.c/run.c llama2/run.c
+```

c/llama2/llama2.go

@@ -23,7 +23,8 @@ import (
 )
 
 const (
-	LLGoPackage = "link"
+	LLGoFiles   = "llama2/llama2.c"
+	LLGoPackage = "link: -lm"
 )
 
 // -----------------------------------------------------------------------------

c/llama2/llama2/llama2.c

@@ -0,0 +1,2 @@
+#define TESTING
+#include "./run.c"

c/llama2/llama2/run.c

@@ -1,2 +1,973 @@
-#define TESTING
-#include "../llama2.c/run.c"
+/* Inference for Llama-2 Transformer model in pure C */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <ctype.h>
+#include <time.h>
+#include <math.h>
+#include <string.h>
+#include <fcntl.h>
+#if defined _WIN32
+    #include "win.h"
+#else
+    #include <unistd.h>
+    #include <sys/mman.h>
+#endif
+// ----------------------------------------------------------------------------
+// Transformer model
+
+typedef struct {
+    int dim; // transformer dimension
+    int hidden_dim; // for ffn layers
+    int n_layers; // number of layers
+    int n_heads; // number of query heads
+    int n_kv_heads; // number of key/value heads (can be < query heads because of multiquery)
+    int vocab_size; // vocabulary size, usually 256 (byte-level)
+    int seq_len; // max sequence length
+} Config;
+
+typedef struct {
+    // token embedding table
+    float* token_embedding_table;    // (vocab_size, dim)
+    // weights for rmsnorms
+    float* rms_att_weight; // (layer, dim) rmsnorm weights
+    float* rms_ffn_weight; // (layer, dim)
+    // weights for matmuls. note dim == n_heads * head_size
+    float* wq; // (layer, dim, n_heads * head_size)
+    float* wk; // (layer, dim, n_kv_heads * head_size)
+    float* wv; // (layer, dim, n_kv_heads * head_size)
+    float* wo; // (layer, n_heads * head_size, dim)
+    // weights for ffn
+    float* w1; // (layer, hidden_dim, dim)
+    float* w2; // (layer, dim, hidden_dim)
+    float* w3; // (layer, hidden_dim, dim)
+    // final rmsnorm
+    float* rms_final_weight; // (dim,)
+    // (optional) classifier weights for the logits, on the last layer
+    float* wcls;
+} TransformerWeights;
+
+typedef struct {
+    // current wave of activations
+    float *x; // activation at current time stamp (dim,)
+    float *xb; // same, but inside a residual branch (dim,)
+    float *xb2; // an additional buffer just for convenience (dim,)
+    float *hb; // buffer for hidden dimension in the ffn (hidden_dim,)
+    float *hb2; // buffer for hidden dimension in the ffn (hidden_dim,)
+    float *q; // query (dim,)
+    float *k; // key (dim,)
+    float *v; // value (dim,)
+    float *att; // buffer for scores/attention values (n_heads, seq_len)
+    float *logits; // output logits
+    // kv cache
+    float* key_cache;   // (layer, seq_len, dim)
+    float* value_cache; // (layer, seq_len, dim)
+} RunState;
+
+typedef struct {
+    Config config; // the hyperparameters of the architecture (the blueprint)
+    TransformerWeights weights; // the weights of the model
+    RunState state; // buffers for the "wave" of activations in the forward pass
+    // some more state needed to properly clean up the memory mapping (sigh)
+    int fd; // file descriptor for memory mapping
+    float* data; // memory mapped data pointer
+    ssize_t file_size; // size of the checkpoint file in bytes
+} Transformer;
+
+void malloc_run_state(RunState* s, Config* p) {
+    // we calloc instead of malloc to keep valgrind happy
+    int kv_dim = (p->dim * p->n_kv_heads) / p->n_heads;
+    s->x = calloc(p->dim, sizeof(float));
+    s->xb = calloc(p->dim, sizeof(float));
+    s->xb2 = calloc(p->dim, sizeof(float));
+    s->hb = calloc(p->hidden_dim, sizeof(float));
+    s->hb2 = calloc(p->hidden_dim, sizeof(float));
+    s->q = calloc(p->dim, sizeof(float));
+    s->key_cache = calloc(p->n_layers * p->seq_len * kv_dim, sizeof(float));
+    s->value_cache = calloc(p->n_layers * p->seq_len * kv_dim, sizeof(float));
+    s->att = calloc(p->n_heads * p->seq_len, sizeof(float));
+    s->logits = calloc(p->vocab_size, sizeof(float));
+    // ensure all mallocs went fine
+    if (!s->x || !s->xb || !s->xb2 || !s->hb || !s->hb2 || !s->q
+     || !s->key_cache || !s->value_cache || !s->att || !s->logits) {
+        fprintf(stderr, "malloc failed!\n");
+        exit(EXIT_FAILURE);
+    }
+}
+
+void free_run_state(RunState* s) {
+    free(s->x);
+    free(s->xb);
+    free(s->xb2);
+    free(s->hb);
+    free(s->hb2);
+    free(s->q);
+    free(s->att);
+    free(s->logits);
+    free(s->key_cache);
+    free(s->value_cache);
+}
+
+void memory_map_weights(TransformerWeights *w, Config* p, float* ptr, int shared_weights) {
+    int head_size = p->dim / p->n_heads;
+    // make sure the multiplications below are done in 64bit to fit the parameter counts of 13B+ models
+    unsigned long long n_layers = p->n_layers;
+    w->token_embedding_table = ptr;
+    ptr += p->vocab_size * p->dim;
+    w->rms_att_weight = ptr;
+    ptr += n_layers * p->dim;
+    w->wq = ptr;
+    ptr += n_layers * p->dim * (p->n_heads * head_size);
+    w->wk = ptr;
+    ptr += n_layers * p->dim * (p->n_kv_heads * head_size);
+    w->wv = ptr;
+    ptr += n_layers * p->dim * (p->n_kv_heads * head_size);
+    w->wo = ptr;
+    ptr += n_layers * (p->n_heads * head_size) * p->dim;
+    w->rms_ffn_weight = ptr;
+    ptr += n_layers * p->dim;
+    w->w1 = ptr;
+    ptr += n_layers * p->dim * p->hidden_dim;
+    w->w2 = ptr;
+    ptr += n_layers * p->hidden_dim * p->dim;
+    w->w3 = ptr;
+    ptr += n_layers * p->dim * p->hidden_dim;
+    w->rms_final_weight = ptr;
+    ptr += p->dim;
+    ptr += p->seq_len * head_size / 2; // skip what used to be freq_cis_real (for RoPE)
+    ptr += p->seq_len * head_size / 2; // skip what used to be freq_cis_imag (for RoPE)
+    w->wcls = shared_weights ? w->token_embedding_table : ptr;
+}
+
+void read_checkpoint(char* checkpoint, Config* config, TransformerWeights* weights,
+                     int* fd, float** data, ssize_t* file_size) {
+    FILE *file = fopen(checkpoint, "rb");
+    if (!file) { fprintf(stderr, "Couldn't open file %s\n", checkpoint); exit(EXIT_FAILURE); }
+    // read in the config header
+    if (fread(config, sizeof(Config), 1, file) != 1) { exit(EXIT_FAILURE); }
+    // negative vocab size is hacky way of signaling unshared weights. bit yikes.
+    int shared_weights = config->vocab_size > 0 ? 1 : 0;
+    config->vocab_size = abs(config->vocab_size);
+    // figure out the file size
+    fseek(file, 0, SEEK_END); // move file pointer to end of file
+    *file_size = ftell(file); // get the file size, in bytes
+    fclose(file);
+    // memory map the Transformer weights into the data pointer
+    *fd = open(checkpoint, O_RDONLY); // open in read only mode
+    if (*fd == -1) { fprintf(stderr, "open failed!\n"); exit(EXIT_FAILURE); }
+    *data = mmap(NULL, *file_size, PROT_READ, MAP_PRIVATE, *fd, 0);
+    if (*data == MAP_FAILED) { fprintf(stderr, "mmap failed!\n"); exit(EXIT_FAILURE); }
+    float* weights_ptr = *data + sizeof(Config)/sizeof(float);
+    memory_map_weights(weights, config, weights_ptr, shared_weights);
+}
+
+void build_transformer(Transformer *t, char* checkpoint_path) {
+    // read in the Config and the Weights from the checkpoint
+    read_checkpoint(checkpoint_path, &t->config, &t->weights, &t->fd, &t->data, &t->file_size);
+    // allocate the RunState buffers
+    malloc_run_state(&t->state, &t->config);
+}
+
+void free_transformer(Transformer* t) {
+    // close the memory mapping
+    if (t->data != MAP_FAILED) { munmap(t->data, t->file_size); }
+    if (t->fd != -1) { close(t->fd); }
+    // free the RunState buffers
+    free_run_state(&t->state);
+}
+
+// ----------------------------------------------------------------------------
+// neural net blocks; the dynamics of the Transformer
+
+void rmsnorm(float* o, float* x, float* weight, int size) {
+    // calculate sum of squares
+    float ss = 0.0f;
+    for (int j = 0; j < size; j++) {
+        ss += x[j] * x[j];
+    }
+    ss /= size;
+    ss += 1e-5f;
+    ss = 1.0f / sqrtf(ss);
+    // normalize and scale
+    for (int j = 0; j < size; j++) {
+        o[j] = weight[j] * (ss * x[j]);
+    }
+}
+
+void softmax(float* x, int size) {
+    // find max value (for numerical stability)
+    float max_val = x[0];
+    for (int i = 1; i < size; i++) {
+        if (x[i] > max_val) {
+            max_val = x[i];
+        }
+    }
+    // exp and sum
+    float sum = 0.0f;
+    for (int i = 0; i < size; i++) {
+        x[i] = expf(x[i] - max_val);
+        sum += x[i];
+    }
+    // normalize
+    for (int i = 0; i < size; i++) {
+        x[i] /= sum;
+    }
+}
+
+void matmul(float* xout, float* x, float* w, int n, int d) {
+    // W (d,n) @ x (n,) -> xout (d,)
+    // by far the most amount of time is spent inside this little function
+    int i;
+    #pragma omp parallel for private(i)
+    for (i = 0; i < d; i++) {
+        float val = 0.0f;
+        for (int j = 0; j < n; j++) {
+            val += w[i * n + j] * x[j];
+        }
+        xout[i] = val;
+    }
+}
+
+float* forward(Transformer* transformer, int token, int pos) {
+
+    // a few convenience variables
+    Config* p = &transformer->config;
+    TransformerWeights* w = &transformer->weights;
+    RunState* s = &transformer->state;
+    float *x = s->x;
+    int dim = p->dim;
+    int kv_dim = (p->dim * p->n_kv_heads) / p->n_heads;
+    int kv_mul = p->n_heads / p->n_kv_heads; // integer multiplier of the kv sharing in multiquery
+    int hidden_dim =  p->hidden_dim;
+    int head_size = dim / p->n_heads;
+
+    // copy the token embedding into x
+    float* content_row = w->token_embedding_table + token * dim;
+    memcpy(x, content_row, dim*sizeof(*x));
+
+    // forward all the layers
+    for(unsigned long long l = 0; l < p->n_layers; l++) {
+
+        // attention rmsnorm
+        rmsnorm(s->xb, x, w->rms_att_weight + l*dim, dim);
+
+        // key and value point to the kv cache
+        int loff = l * p->seq_len * kv_dim; // kv cache layer offset for convenience
+        s->k = s->key_cache + loff + pos * kv_dim;
+        s->v = s->value_cache + loff + pos * kv_dim;
+
+        // qkv matmuls for this position
+        matmul(s->q, s->xb, w->wq + l*dim*dim, dim, dim);
+        matmul(s->k, s->xb, w->wk + l*dim*kv_dim, dim, kv_dim);
+        matmul(s->v, s->xb, w->wv + l*dim*kv_dim, dim, kv_dim);
+
+        // RoPE relative positional encoding: complex-valued rotate q and k in each head
+        for (int i = 0; i < dim; i+=2) {
+            int head_dim = i % head_size;
+            float freq = 1.0f / powf(10000.0f, head_dim / (float)head_size);
+            float val = pos * freq;
+            float fcr = cosf(val);
+            float fci = sinf(val);
+            int rotn = i < kv_dim ? 2 : 1; // how many vectors? 2 = q & k, 1 = q only
+            for (int v = 0; v < rotn; v++) {
+                float* vec = v == 0 ? s->q : s->k; // the vector to rotate (query or key)
+                float v0 = vec[i];
+                float v1 = vec[i+1];
+                vec[i]   = v0 * fcr - v1 * fci;
+                vec[i+1] = v0 * fci + v1 * fcr;
+            }
+        }
+
+        // multihead attention. iterate over all heads
+        int h;
+        #pragma omp parallel for private(h)
+        for (h = 0; h < p->n_heads; h++) {
+            // get the query vector for this head
+            float* q = s->q + h * head_size;
+            // attention scores for this head
+            float* att = s->att + h * p->seq_len;
+            // iterate over all timesteps, including the current one
+            for (int t = 0; t <= pos; t++) {
+                // get the key vector for this head and at this timestep
+                float* k = s->key_cache + loff + t * kv_dim + (h / kv_mul) * head_size;
+                // calculate the attention score as the dot product of q and k
+                float score = 0.0f;
+                for (int i = 0; i < head_size; i++) {
+                    score += q[i] * k[i];
+                }
+                score /= sqrtf(head_size);
+                // save the score to the attention buffer
+                att[t] = score;
+            }
+
+            // softmax the scores to get attention weights, from 0..pos inclusively
+            softmax(att, pos + 1);
+
+            // weighted sum of the values, store back into xb
+            float* xb = s->xb + h * head_size;
+            memset(xb, 0, head_size * sizeof(float));
+            for (int t = 0; t <= pos; t++) {
+                // get the value vector for this head and at this timestep
+                float* v = s->value_cache + loff + t * kv_dim + (h / kv_mul) * head_size;
+                // get the attention weight for this timestep
+                float a = att[t];
+                // accumulate the weighted value into xb
+                for (int i = 0; i < head_size; i++) {
+                    xb[i] += a * v[i];
+                }
+            }
+        }
+
+        // final matmul to get the output of the attention
+        matmul(s->xb2, s->xb, w->wo + l*dim*dim, dim, dim);
+
+        // residual connection back into x
+        for (int i = 0; i < dim; i++) {
+            x[i] += s->xb2[i];
+        }
+
+        // ffn rmsnorm
+        rmsnorm(s->xb, x, w->rms_ffn_weight + l*dim, dim);
+
+        // Now for FFN in PyTorch we have: self.w2(F.silu(self.w1(x)) * self.w3(x))
+        // first calculate self.w1(x) and self.w3(x)
+        matmul(s->hb, s->xb, w->w1 + l*dim*hidden_dim, dim, hidden_dim);
+        matmul(s->hb2, s->xb, w->w3 + l*dim*hidden_dim, dim, hidden_dim);
+
+        // SwiGLU non-linearity
+        for (int i = 0; i < hidden_dim; i++) {
+            float val = s->hb[i];
+            // silu(x)=x*σ(x), where σ(x) is the logistic sigmoid
+            val *= (1.0f / (1.0f + expf(-val)));
+            // elementwise multiply with w3(x)
+            val *= s->hb2[i];
+            s->hb[i] = val;
+        }
+
+        // final matmul to get the output of the ffn
+        matmul(s->xb, s->hb, w->w2 + l*dim*hidden_dim, hidden_dim, dim);
+
+        // residual connection
+        for (int i = 0; i < dim; i++) {
+            x[i] += s->xb[i];
+        }
+    }
+
+    // final rmsnorm
+    rmsnorm(x, x, w->rms_final_weight, dim);
+
+    // classifier into logits
+    matmul(s->logits, x, w->wcls, p->dim, p->vocab_size);
+    return s->logits;
+}
+
+// ----------------------------------------------------------------------------
+// The Byte Pair Encoding (BPE) Tokenizer that translates strings <-> tokens
+
+typedef struct {
+    char *str;
+    int id;
+} TokenIndex;
+
+typedef struct {
+    char** vocab;
+    float* vocab_scores;
+    TokenIndex *sorted_vocab;
+    int vocab_size;
+    unsigned int max_token_length;
+    unsigned char byte_pieces[512]; // stores all single-byte strings
+} Tokenizer;
+
+int compare_tokens(const void *a, const void *b) {
+    return strcmp(((TokenIndex*)a)->str, ((TokenIndex*)b)->str);
+}
+
+void build_tokenizer(Tokenizer* t, char* tokenizer_path, int vocab_size) {
+    // i should have written the vocab_size into the tokenizer file... sigh
+    t->vocab_size = vocab_size;
+    // malloc space to hold the scores and the strings
+    t->vocab = (char**)malloc(vocab_size * sizeof(char*));
+    t->vocab_scores = (float*)malloc(vocab_size * sizeof(float));
+    t->sorted_vocab = NULL; // initialized lazily
+    for (int i = 0; i < 256; i++) {
+        t->byte_pieces[i * 2] = (unsigned char)i;
+        t->byte_pieces[i * 2 + 1] = '\0';
+    }
+    // read in the file
+    FILE *file = fopen(tokenizer_path, "rb");
+    if (!file) { fprintf(stderr, "couldn't load %s\n", tokenizer_path); exit(EXIT_FAILURE); }
+    if (fread(&t->max_token_length, sizeof(int), 1, file) != 1) { fprintf(stderr, "failed read\n"); exit(EXIT_FAILURE); }
+    int len;
+    for (int i = 0; i < vocab_size; i++) {
+        if (fread(t->vocab_scores + i, sizeof(float), 1, file) != 1) { fprintf(stderr, "failed read\n"); exit(EXIT_FAILURE);}
+        if (fread(&len, sizeof(int), 1, file) != 1) { fprintf(stderr, "failed read\n"); exit(EXIT_FAILURE); }
+        t->vocab[i] = (char *)malloc(len + 1);
+        if (fread(t->vocab[i], len, 1, file) != 1) { fprintf(stderr, "failed read\n"); exit(EXIT_FAILURE); }
+        t->vocab[i][len] = '\0'; // add the string terminating token
+    }
+    fclose(file);
+}
+
+void free_tokenizer(Tokenizer* t) {
+    for (int i = 0; i < t->vocab_size; i++) { free(t->vocab[i]); }
+    free(t->vocab);
+    free(t->vocab_scores);
+    free(t->sorted_vocab);
+}
+
+char* decode(Tokenizer* t, int prev_token, int token) {
+    char *piece = t->vocab[token];
+    // following BOS (1) token, sentencepiece decoder strips any leading whitespace (see PR #89)
+    if (prev_token == 1 && piece[0] == ' ') { piece++; }
+    // careful, some tokens designate raw bytes, and look like e.g. '<0x01>'
+    // parse this and convert and return the actual byte
+    unsigned char byte_val;
+    if (sscanf(piece, "<0x%02hhX>", &byte_val) == 1) {
+        piece = (char*)t->byte_pieces + byte_val * 2;
+    }
+    return piece;
+}
+
+void safe_printf(char *piece) {
+    // piece might be a raw byte token, and we only want to print printable chars or whitespace
+    // because some of the other bytes can be various control codes, backspace, etc.
+    if (piece == NULL) { return; }
+    if (piece[0] == '\0') { return; }
+    if (piece[1] == '\0') {
+        unsigned char byte_val = piece[0];
+        if (!(isprint(byte_val) || isspace(byte_val))) {
+            return; // bad byte, don't print it
+        }
+    }
+    printf("%s", piece);
+}
+
+int str_lookup(char *str, TokenIndex *sorted_vocab, int vocab_size) {
+    // efficiently find the perfect match for str in vocab, return its index or -1 if not found
+    TokenIndex tok = { .str = str }; // acts as the key to search for
+    TokenIndex *res = bsearch(&tok, sorted_vocab, vocab_size, sizeof(TokenIndex), compare_tokens);
+    return res != NULL ? res->id : -1;
+}
+
+void encode(Tokenizer* t, char *text, int8_t bos, int8_t eos, int *tokens, int *n_tokens) {
+    // encode the string text (input) into an upper-bound preallocated tokens[] array
+    // bos != 0 means prepend the BOS token (=1), eos != 0 means append the EOS token (=2)
+    if (text == NULL) { fprintf(stderr, "cannot encode NULL text\n"); exit(EXIT_FAILURE); }
+
+    if (t->sorted_vocab == NULL) {
+        // lazily malloc and sort the vocabulary
+        t->sorted_vocab = malloc(t->vocab_size * sizeof(TokenIndex));
+        for (int i = 0; i < t->vocab_size; i++) {
+            t->sorted_vocab[i].str = t->vocab[i];
+            t->sorted_vocab[i].id = i;
+        }
+        qsort(t->sorted_vocab, t->vocab_size, sizeof(TokenIndex), compare_tokens);
+    }
+
+    // create a temporary buffer that will store merge candidates of always two consecutive tokens
+    // *2 for concat, +1 for null terminator +2 for UTF8 (in case max_token_length is 1)
+    char* str_buffer = malloc((t->max_token_length*2 +1 +2) * sizeof(char));
+    size_t str_len = 0;
+
+    // start at 0 tokens
+    *n_tokens = 0;
+
+    // add optional BOS (=1) token, if desired
+    if (bos) tokens[(*n_tokens)++] = 1;
+
+    // add_dummy_prefix is true by default
+    // so prepend a dummy prefix token to the input string, but only if text != ""
+    // TODO: pretty sure this isn't correct in the general case but I don't have the
+    // energy to read more of the sentencepiece code to figure out what it's doing
+    if (text[0] != '\0') {
+        int dummy_prefix = str_lookup(" ", t->sorted_vocab, t->vocab_size);
+        tokens[(*n_tokens)++] = dummy_prefix;
+    }
+
+    // Okay UTF-8 time. This will get messy. Here is the reference from Wikipedia:
+    // Code point ↔ UTF-8 conversion
+    // First code point	Last code point	Byte 1	Byte 2	Byte 3	Byte 4
+    // U+0000	U+007F	    0xxxxxxx
+    // U+0080	U+07FF	    110xxxxx	10xxxxxx
+    // U+0800	U+FFFF	    1110xxxx	10xxxxxx	10xxxxxx
+    // U+10000	U+10FFFF    11110xxx	10xxxxxx	10xxxxxx	10xxxxxx
+
+    // process the raw (UTF-8) byte sequence of the input string
+    for (char *c = text; *c != '\0'; c++) {
+
+        // reset buffer if the current byte is ASCII or a leading byte
+        // 0xC0 is 11000000, so (*c & 0xC0) keeps the first 2 bits and zeros the rest
+        // 0x80 is 10000000
+        // in UTF-8, all continuation bytes start with "10" in first two bits
+        // so in English this is: "if this byte is not a continuation byte"
+        if ((*c & 0xC0) != 0x80) {
+            // this byte must be either a leading byte (11...) or an ASCII char (0x...)
+            // => reset our location, as we're starting a new UTF-8 codepoint
+            str_len = 0;
+        }
+
+        // append the current byte to the buffer
+        str_buffer[str_len++] = *c; // ++ is post-increment, incremented after this line
+        str_buffer[str_len] = '\0';
+
+        // while the next character is a continuation byte, continue appending
+        // but if there are too many of them, just stop to avoid overruning str_buffer size.
+        if ((*(c+1) & 0xC0) == 0x80 && str_len < 4) {
+            continue;
+        }
+
+        // ok c+1 is not a continuation byte, so we've read in a full codepoint
+        int id = str_lookup(str_buffer, t->sorted_vocab, t->vocab_size);
+
+        if (id != -1) {
+            // we found this codepoint in vocab, add it as a token
+            tokens[(*n_tokens)++] = id;
+        } else {
+            // byte_fallback encoding: just encode each byte as a token
+            // +3 is here because the first 3 vocab elements are <unk>, <s>, </s>
+            // so the individual bytes only start at index 3
+            for (int i=0; i < str_len; i++) {
+                tokens[(*n_tokens)++] = (unsigned char)str_buffer[i] + 3;
+            }
+        }
+        str_len = 0; // protect against a sequence of stray UTF8 continuation bytes
+    }
+
+    // merge the best consecutive pair each iteration, according the scores in vocab_scores
+    while (1) {
+        float best_score = -1e10;
+        int best_id = -1;
+        int best_idx = -1;
+
+        for (int i=0; i < (*n_tokens-1); i++) {
+            // check if we can merge the pair (tokens[i], tokens[i+1])
+            sprintf(str_buffer, "%s%s", t->vocab[tokens[i]], t->vocab[tokens[i+1]]);
+            int id = str_lookup(str_buffer, t->sorted_vocab, t->vocab_size);
+            if (id != -1 && t->vocab_scores[id] > best_score) {
+                // this merge pair exists in vocab! record its score and position
+                best_score = t->vocab_scores[id];
+                best_id = id;
+                best_idx = i;
+            }
+        }
+
+        if (best_idx == -1) {
+            break; // we couldn't find any more pairs to merge, so we're done
+        }
+
+        // merge the consecutive pair (best_idx, best_idx+1) into new token best_id
+        tokens[best_idx] = best_id;
+        // delete token at position best_idx+1, shift the entire sequence back 1
+        for (int i = best_idx+1; i < (*n_tokens-1); i++) {
+            tokens[i] = tokens[i+1];
+        }
+        (*n_tokens)--; // token length decreased
+    }
+
+    // add optional EOS (=2) token, if desired
+    if (eos) tokens[(*n_tokens)++] = 2;
+
+    free(str_buffer);
+}
+
+// ----------------------------------------------------------------------------
+// The Sampler, which takes logits and returns a sampled token
+// sampling can be done in a few ways: greedy argmax, sampling, top-p sampling
+
+typedef struct {
+    float prob;
+    int index;
+} ProbIndex; // struct used when sorting probabilities during top-p sampling
+
+typedef struct {
+    int vocab_size;
+    ProbIndex* probindex; // buffer used in top-p sampling
+    float temperature;
+    float topp;
+    unsigned long long rng_state;
+} Sampler;
+
+int sample_argmax(float* probabilities, int n) {
+    // return the index that has the highest probability
+    int max_i = 0;
+    float max_p = probabilities[0];
+    for (int i = 1; i < n; i++) {
+        if (probabilities[i] > max_p) {
+            max_i = i;
+            max_p = probabilities[i];
+        }
+    }
+    return max_i;
+}
+
+int sample_mult(float* probabilities, int n, float coin) {
+    // sample index from probabilities (they must sum to 1!)
+    // coin is a random number in [0, 1), usually from random_f32()
+    float cdf = 0.0f;
+    for (int i = 0; i < n; i++) {
+        cdf += probabilities[i];
+        if (coin < cdf) {
+            return i;
+        }
+    }
+    return n - 1; // in case of rounding errors
+}
+
+int compare(const void* a, const void* b) {
+    ProbIndex* a_ = (ProbIndex*) a;
+    ProbIndex* b_ = (ProbIndex*) b;
+    if (a_->prob > b_->prob) return -1;
+    if (a_->prob < b_->prob) return 1;
+    return 0;
+}
+
+int sample_topp(float* probabilities, int n, float topp, ProbIndex* probindex, float coin) {
+    // top-p sampling (or "nucleus sampling") samples from the smallest set of
+    // tokens that exceed probability topp. This way we never sample tokens that
+    // have very low probabilities and are less likely to go "off the rails".
+    // coin is a random number in [0, 1), usually from random_f32()
+
+    int n0 = 0;
+    // quicksort indices in descending order of probabilities
+    // values smaller than (1 - topp) / (n - 1) cannot be part of the result
+    // so for efficiency we crop these out as candidates before sorting
+    const float cutoff = (1.0f - topp) / (n - 1);
+    for (int i = 0; i < n; i++) {
+        if (probabilities[i] >= cutoff) {
+            probindex[n0].index = i;
+            probindex[n0].prob = probabilities[i];
+            n0++;
+        }
+    }
+    qsort(probindex, n0, sizeof(ProbIndex), compare);
+
+    // truncate the list where cumulative probability exceeds topp
+    float cumulative_prob = 0.0f;
+    int last_idx = n0 - 1; // in case of rounding errors consider all elements
+    for (int i = 0; i < n0; i++) {
+        cumulative_prob += probindex[i].prob;
+        if (cumulative_prob > topp) {
+            last_idx = i;
+            break; // we've exceeded topp by including last_idx
+        }
+    }
+
+    // sample from the truncated list
+    float r = coin * cumulative_prob;
+    float cdf = 0.0f;
+    for (int i = 0; i <= last_idx; i++) {
+        cdf += probindex[i].prob;
+        if (r < cdf) {
+            return probindex[i].index;
+        }
+    }
+    return probindex[last_idx].index; // in case of rounding errors
+}
+
+void build_sampler(Sampler* sampler, int vocab_size, float temperature, float topp, unsigned long long rng_seed) {
+    sampler->vocab_size = vocab_size;
+    sampler->temperature = temperature;
+    sampler->topp = topp;
+    sampler->rng_state = rng_seed;
+    // buffer only used with nucleus sampling; may not need but it's ~small
+    sampler->probindex = malloc(sampler->vocab_size * sizeof(ProbIndex));
+}
+
+void free_sampler(Sampler* sampler) {
+    free(sampler->probindex);
+}
+
+unsigned int random_u32(unsigned long long *state) {
+    // xorshift rng: https://en.wikipedia.org/wiki/Xorshift#xorshift.2A
+    *state ^= *state >> 12;
+    *state ^= *state << 25;
+    *state ^= *state >> 27;
+    return (*state * 0x2545F4914F6CDD1Dull) >> 32;
+}
+float random_f32(unsigned long long *state) { // random float32 in [0,1)
+    return (random_u32(state) >> 8) / 16777216.0f;
+}
+
+int sample(Sampler* sampler, float* logits) {
+    // sample the token given the logits and some hyperparameters
+    int next;
+    if (sampler->temperature == 0.0f) {
+        // greedy argmax sampling: take the token with the highest probability
+        next = sample_argmax(logits, sampler->vocab_size);
+    } else {
+        // apply the temperature to the logits
+        for (int q=0; q<sampler->vocab_size; q++) { logits[q] /= sampler->temperature; }
+        // apply softmax to the logits to get the probabilities for next token
+        softmax(logits, sampler->vocab_size);
+        // flip a (float) coin (this is our source of entropy for sampling)
+        float coin = random_f32(&sampler->rng_state);
+        // we sample from this distribution to get the next token
+        if (sampler->topp <= 0 || sampler->topp >= 1) {
+            // simply sample from the predicted probability distribution
+            next = sample_mult(logits, sampler->vocab_size, coin);
+        } else {
+            // top-p (nucleus) sampling, clamping the least likely tokens to zero
+            next = sample_topp(logits, sampler->vocab_size, sampler->topp, sampler->probindex, coin);
+        }
+    }
+    return next;
+}
+
+// ----------------------------------------------------------------------------
+// utilities: time
+
+long time_in_ms() {
+    // return time in milliseconds, for benchmarking the model speed
+    struct timespec time;
+    clock_gettime(CLOCK_REALTIME, &time);
+    return time.tv_sec * 1000 + time.tv_nsec / 1000000;
+}
+
+// ----------------------------------------------------------------------------
+// generation loop
+
+void generate(Transformer *transformer, Tokenizer *tokenizer, Sampler *sampler, char *prompt, int steps) {
+    char *empty_prompt = "";
+    if (prompt == NULL) { prompt = empty_prompt; }
+
+    // encode the (string) prompt into tokens sequence
+    int num_prompt_tokens = 0;
+    int* prompt_tokens = (int*)malloc((strlen(prompt)+3) * sizeof(int)); // +3 for '\0', ?BOS, ?EOS
+    encode(tokenizer, prompt, 1, 0, prompt_tokens, &num_prompt_tokens);
+    if (num_prompt_tokens < 1) {
+        fprintf(stderr, "something is wrong, expected at least 1 prompt token\n");
+        exit(EXIT_FAILURE);
+    }
+
+    // start the main loop
+    long start = 0;  // used to time our code, only initialized after first iteration
+    int next;        // will store the next token in the sequence
+    int token = prompt_tokens[0]; // kick off with the first token in the prompt
+    int pos = 0;     // position in the sequence
+    while (pos < steps) {
+
+        // forward the transformer to get logits for the next token
+        float* logits = forward(transformer, token, pos);
+
+        // advance the state machine
+        if (pos < num_prompt_tokens - 1) {
+            // if we are still processing the input prompt, force the next prompt token
+            next = prompt_tokens[pos + 1];
+        } else {
+            // otherwise sample the next token from the logits
+            next = sample(sampler, logits);
+        }
+        pos++;
+
+        // data-dependent terminating condition: the BOS (=1) token delimits sequences
+        if (next == 1) { break; }
+
+        // print the token as string, decode it with the Tokenizer object
+        char* piece = decode(tokenizer, token, next);
+        safe_printf(piece); // same as printf("%s", piece), but skips "unsafe" bytes
+        fflush(stdout);
+        token = next;
+
+        // init the timer here because the first iteration can be slower
+        if (start == 0) { start = time_in_ms(); }
+    }
+    printf("\n");
+
+    // report achieved tok/s (pos-1 because the timer starts after first iteration)
+    if (pos > 1) {
+        long end = time_in_ms();
+        fprintf(stderr, "achieved tok/s: %f\n", (pos-1) / (double)(end-start)*1000);
+    }
+
+    free(prompt_tokens);
+}
+
+void read_stdin(const char* guide, char* buffer, size_t bufsize) {
+    // read a line from stdin, up to but not including \n
+    printf("%s", guide);
+    if (fgets(buffer, bufsize, stdin) != NULL) {
+        size_t len = strlen(buffer);
+        if (len > 0 && buffer[len - 1] == '\n') {
+            buffer[len - 1] = '\0'; // strip newline
+        }
+    }
+}
+
+// ----------------------------------------------------------------------------
+// chat loop
+// I manually inspected the tokens for a few chat conversations compared to
+// python reference and that seemed ok, but this was not thoroughly tested and
+// is not safely implemented, it's more a proof of concept atm.
+
+void chat(Transformer *transformer, Tokenizer *tokenizer, Sampler *sampler,
+          char *cli_user_prompt, char *cli_system_prompt, int steps) {
+
+    // buffers for reading the system prompt and user prompt from stdin
+    // you'll notice they are soomewhat haphazardly and unsafely set atm
+    char system_prompt[512];
+    char user_prompt[512];
+    char rendered_prompt[1152];
+    int num_prompt_tokens = 0;
+    int* prompt_tokens = (int*)malloc(1152 * sizeof(int));
+    int user_idx;
+
+    // start the main loop
+    int8_t user_turn = 1; // user starts
+    int next;        // will store the next token in the sequence
+    int token;       // stores the current token to feed into the transformer
+    int prev_token;
+    int pos = 0;     // position in the sequence
+    while (pos < steps) {
+
+        // when it is the user's turn to contribute tokens to the dialog...
+        if (user_turn) {
+            // get the (optional) system prompt at position 0
+            if (pos == 0) {
+                // at position 0, the user can also contribute a system prompt
+                if (cli_system_prompt == NULL) {
+                    // system prompt was not passed in, attempt to get it from stdin
+                    read_stdin("Enter system prompt (optional): ", system_prompt, sizeof(system_prompt));
+                } else {
+                    // system prompt was passed in, use it
+                    strcpy(system_prompt, cli_system_prompt);
+                }
+            }
+            // get the user prompt
+            if (pos == 0 && cli_user_prompt != NULL) {
+                // user prompt for position 0 was passed in, use it
+                strcpy(user_prompt, cli_user_prompt);
+            } else {
+                // otherwise get user prompt from stdin
+                read_stdin("User: ", user_prompt, sizeof(user_prompt));
+            }
+            // render user/system prompts into the Llama 2 Chat schema
+            if (pos == 0 && system_prompt[0] != '\0') {
+                char system_template[] = "[INST] <<SYS>>\n%s\n<</SYS>>\n\n%s [/INST]";
+                sprintf(rendered_prompt, system_template, system_prompt, user_prompt);
+            } else {
+                char user_template[] = "[INST] %s [/INST]";
+                sprintf(rendered_prompt, user_template, user_prompt);
+            }
+            // encode the rendered prompt into tokens
+            encode(tokenizer, rendered_prompt, 1, 0, prompt_tokens, &num_prompt_tokens);
+            user_idx = 0; // reset the user index
+            user_turn = 0;
+            printf("Assistant: ");
+        }
+
+        // determine the token to pass into the transformer next
+        if (user_idx < num_prompt_tokens) {
+            // if we are still processing the input prompt, force the next prompt token
+            token = prompt_tokens[user_idx++];
+        } else {
+            // otherwise use the next token sampled from previous turn
+            token = next;
+        }
+        // EOS (=2) token ends the Assistant turn
+        if (token == 2) { user_turn = 1; }
+
+        // forward the transformer to get logits for the next token
+        float* logits = forward(transformer, token, pos);
+        next = sample(sampler, logits);
+        pos++;
+
+        if (user_idx >= num_prompt_tokens && next != 2) {
+            // the Assistant is responding, so print its output
+            char* piece = decode(tokenizer, token, next);
+            safe_printf(piece); // same as printf("%s", piece), but skips "unsafe" bytes
+            fflush(stdout);
+        }
+        if (next == 2) { printf("\n"); }
+    }
+    printf("\n");
+    free(prompt_tokens);
+}
+
+
+// ----------------------------------------------------------------------------
+// CLI, include only if not testing
+#ifndef TESTING
+
+void error_usage() {
+    fprintf(stderr, "Usage:   run <checkpoint> [options]\n");
+    fprintf(stderr, "Example: run model.bin -n 256 -i \"Once upon a time\"\n");
+    fprintf(stderr, "Options:\n");
+    fprintf(stderr, "  -t <float>  temperature in [0,inf], default 1.0\n");
+    fprintf(stderr, "  -p <float>  p value in top-p (nucleus) sampling in [0,1] default 0.9\n");
+    fprintf(stderr, "  -s <int>    random seed, default time(NULL)\n");
+    fprintf(stderr, "  -n <int>    number of steps to run for, default 256. 0 = max_seq_len\n");
+    fprintf(stderr, "  -i <string> input prompt\n");
+    fprintf(stderr, "  -z <string> optional path to custom tokenizer\n");
+    fprintf(stderr, "  -m <string> mode: generate|chat, default: generate\n");
+    fprintf(stderr, "  -y <string> (optional) system prompt in chat mode\n");
+    exit(EXIT_FAILURE);
+}
+
+int main(int argc, char *argv[]) {
+
+    // default parameters
+    char *checkpoint_path = NULL;  // e.g. out/model.bin
+    char *tokenizer_path = "tokenizer.bin";
+    float temperature = 1.0f;   // 0.0 = greedy deterministic. 1.0 = original. don't set higher
+    float topp = 0.9f;          // top-p in nucleus sampling. 1.0 = off. 0.9 works well, but slower
+    int steps = 256;            // number of steps to run for
+    char *prompt = NULL;        // prompt string
+    unsigned long long rng_seed = 0; // seed rng with time by default
+    char *mode = "generate";    // generate|chat
+    char *system_prompt = NULL; // the (optional) system prompt to use in chat mode
+
+    // poor man's C argparse so we can override the defaults above from the command line
+    if (argc >= 2) { checkpoint_path = argv[1]; } else { error_usage(); }
+    for (int i = 2; i < argc; i+=2) {
+        // do some basic validation
+        if (i + 1 >= argc) { error_usage(); } // must have arg after flag
+        if (argv[i][0] != '-') { error_usage(); } // must start with dash
+        if (strlen(argv[i]) != 2) { error_usage(); } // must be -x (one dash, one letter)
+        // read in the args
+        if (argv[i][1] == 't') { temperature = atof(argv[i + 1]); }
+        else if (argv[i][1] == 'p') { topp = atof(argv[i + 1]); }
+        else if (argv[i][1] == 's') { rng_seed = atoi(argv[i + 1]); }
+        else if (argv[i][1] == 'n') { steps = atoi(argv[i + 1]); }
+        else if (argv[i][1] == 'i') { prompt = argv[i + 1]; }
+        else if (argv[i][1] == 'z') { tokenizer_path = argv[i + 1]; }
+        else if (argv[i][1] == 'm') { mode = argv[i + 1]; }
+        else if (argv[i][1] == 'y') { system_prompt = argv[i + 1]; }
+        else { error_usage(); }
+    }
+
+    // parameter validation/overrides
+    if (rng_seed <= 0) rng_seed = (unsigned int)time(NULL);
+    if (temperature < 0.0) temperature = 0.0;
+    if (topp < 0.0 || 1.0 < topp) topp = 0.9;
+    if (steps < 0) steps = 0;
+
+    // build the Transformer via the model .bin file
+    Transformer transformer;
+    build_transformer(&transformer, checkpoint_path);
+    if (steps == 0 || steps > transformer.config.seq_len) steps = transformer.config.seq_len; // override to ~max length
+
+    // build the Tokenizer via the tokenizer .bin file
+    Tokenizer tokenizer;
+    build_tokenizer(&tokenizer, tokenizer_path, transformer.config.vocab_size);
+
+    // build the Sampler
+    Sampler sampler;
+    build_sampler(&sampler, transformer.config.vocab_size, temperature, topp, rng_seed);
+
+    // run!
+    if (strcmp(mode, "generate") == 0) {
+        generate(&transformer, &tokenizer, &sampler, prompt, steps);
+    } else if (strcmp(mode, "chat") == 0) {
+        chat(&transformer, &tokenizer, &sampler, prompt, system_prompt, steps);
+    } else {
+        fprintf(stderr, "unknown mode: %s\n", mode);
+        error_usage();
+    }
+
+    // memory and file handles cleanup
+    free_sampler(&sampler);
+    free_tokenizer(&tokenizer);
+    free_transformer(&transformer);
+    return 0;
+}
+#endif