Dockerfile.cache

# Emscripten Cache Component - Emscripten 4.0.16
# Builds Emscripten system libraries cache from source
#
# Emscripten revision: 00c5c7d93f6c898cb4634299164fc5f8b87e10bf
# From: emscripten-releases tag 4.0.16
#
# Build: docker build -f Dockerfile.cache \
#          --build-arg LLVM_IMAGE=ghcr.io/discere-os/wasm-builder-llvm:07ca4db1 \
#          --build-arg BINARYEN_IMAGE=ghcr.io/discere-os/wasm-builder-binaryen:c326e66b \
#          -t ghcr.io/discere-os/wasm-builder-cache:00c5c7d9 .
# Time: ~20-30 minutes
# Size: ~500MB-1GB
#
# This stage builds all the Emscripten system libraries and then discards
# the build artifacts, keeping only the final .a files.

# Global ARGs for component images (must be before any FROM to use in FROM)
ARG LLVM_IMAGE=ghcr.io/discere-os/wasm-builder-llvm:07ca4db1
ARG BINARYEN_IMAGE=ghcr.io/discere-os/wasm-builder-binaryen:c326e66b

#
# Stage 0: Node.js base
#
FROM ubuntu:24.04 AS nodejs-base

ENV DEBIAN_FRONTEND=noninteractive
ENV LANG=C.UTF-8
ENV LC_ALL=C.UTF-8
ENV TZ=Pacific/Auckland

# Install Node.js 22
RUN apt-get update && apt-get install -y --no-install-recommends \
    curl \
    ca-certificates \
    && curl -fsSL https://deb.nodesource.com/setup_22.x | bash - && \
    apt-get install -y nodejs && \
    rm -rf /var/lib/apt/lists/*

#
# Stage 1: Builder base
#
FROM nodejs-base AS builder-base

# Install build dependencies
RUN apt-get update && apt-get install -y --no-install-recommends \
    build-essential \
    cmake \
    ninja-build \
    git \
    python3 \
    python3-pip \
    curl \
    wget \
    ca-certificates \
    pkg-config \
    file \
    && rm -rf /var/lib/apt/lists/*

#
# Stage 2: Import pre-built components
#
# Docker doesn't support variable expansion in COPY --from, so we need
# to create intermediate FROM stages using the ARG values as a workaround
FROM ${LLVM_IMAGE} AS llvm-prebuilt
FROM ${BINARYEN_IMAGE} AS binaryen-prebuilt

#
# Stage 3: Build Emscripten cache
#
FROM builder-base AS cache-builder

# Get LLVM and Binaryen from pre-built image stages
COPY --from=llvm-prebuilt /opt/llvm /opt/llvm
COPY --from=binaryen-prebuilt /opt/binaryen /opt/binaryen

# Install Emscripten temporarily for cache building
ARG EMSCRIPTEN_REVISION=00c5c7d93f6c898cb4634299164fc5f8b87e10bf
WORKDIR /opt
RUN git clone https://github.com/emscripten-core/emscripten.git && \
    cd emscripten && \
    git checkout ${EMSCRIPTEN_REVISION} && \
    npm install --production && \
    npm cache clean --force && \
    rm -rf .git tests third_party/websockify third_party/lzma

# Set up environment for cache building
ENV PATH="/opt/emscripten:/opt/llvm/bin:/opt/binaryen/bin:${PATH}"
ENV EM_CONFIG=/opt/emscripten/.emscripten
ENV EM_CACHE=/opt/emscripten/cache

# Create minimal .emscripten config for cache building
RUN cat > /opt/emscripten/.emscripten << 'EOF'
import os
LLVM_ROOT = '/opt/llvm/bin'
BINARYEN_ROOT = '/opt/binaryen'
NODE_JS = '/usr/bin/node'
COMPILER_ENGINE = NODE_JS
JS_ENGINES = [NODE_JS]
CACHE = '/opt/emscripten/cache'
PORTS = None
FROZEN_CACHE = False
EOF

# Create wrapper scripts for Emscripten tools
RUN cd /opt/emscripten && \
    for tool in em++ emcc emar emranlib emsize emmake emcmake emconfigure emnm emdump emstrip; do \
        if [ -f "${tool}.py" ]; then \
            echo '#!/bin/bash' > ${tool} && \
            echo "exec python3 /opt/emscripten/${tool}.py \"\$@\"" >> ${tool} && \
            chmod +x ${tool}; \
        fi; \
    done

# Initialize Emscripten
RUN mkdir -p /opt/emscripten/cache && \
    echo 'int main() { return 42; }' > /tmp/test.c && \
    emcc /tmp/test.c -o /tmp/test.js && \
    rm -f /tmp/test.*

# Build all cache variants
# This creates the system libraries in all needed configurations
RUN cd /opt/emscripten && \
    echo "Building standard libraries..." && \
    python3 embuilder.py build ALL && \
    echo "Building LTO variants..." && \
    python3 embuilder.py build ALL --lto && \
    echo "Building PIC variants (for SIDE_MODULE)..." && \
    python3 embuilder.py build ALL --pic && \
    echo "Building PIC+LTO variants..." && \
    python3 embuilder.py build ALL --pic --lto && \
    echo "Cache build complete"

# CRITICAL OPTIMIZATION: Clean up build artifacts
# This removes ~3GB of temporary files while keeping the final .a libraries
RUN echo "Cleaning cache build artifacts..." && \
    # Remove build directory (contains .o files, no longer needed)
    rm -rf /opt/emscripten/cache/build && \
    # Remove ports downloads (we don't use ports system)
    rm -rf /opt/emscripten/cache/ports && \
    # Remove Python cache
    find /opt/emscripten -type d -name __pycache__ -exec rm -rf {} + 2>/dev/null || true && \
    find /opt/emscripten -type f -name "*.pyc" -delete && \
    # Report final cache size
    echo "Final cache size: $(du -sh /opt/emscripten/cache | cut -f1)" && \
    echo "Sysroot size: $(du -sh /opt/emscripten/cache/sysroot | cut -f1)"

# Stage 4: Final stage - Cache only (without LLVM/Binaryen to reduce duplication)
FROM ubuntu:24.04

COPY --from=cache-builder /opt/emscripten/cache /opt/emscripten/cache

# Verify cache structure
RUN test -d /opt/emscripten/cache/sysroot && \
    echo "Cache size: $(du -sh /opt/emscripten/cache | cut -f1)" && \
    echo "Emscripten Cache Component Ready"

WORKDIR /workspace