Fix makefile, add workflow to build ASM and WASM versions, update dependencies (#14)

Author: Jaifroid

.github/workflows/build_libzim_wasm.yml

@@ -1,25 +1,71 @@
-name: Docker Image CI
+# Workflow to build and publish the libzim ASM and WASM arttefacts together with the JavaScript wrapper.
+# If this workflow is triggered by the creation of a draft release, then the artefacts are uploaded to the release assets.
+# If it is triggered by a push or pull request to master, or manually, then the artefacts are archived under the corresponding Action.
+
+name: Build and publish release artefacts (Docker)
 
 on:
-  # push:
-  #   branches: [ "master" ]
-  # pull_request:
-  #   branches: [ "master" ]
+  push:
+    branches: [ master ]
+    tags:
+    - 'v*' # Tag push events matching v*, i.e. v1.0, v20.15.10
+  pull_request:
+    branches: [ master ]
   workflow_dispatch:
+    inputs:
+      version:
+        description: |
+          If you wish to create a draft release, set the tag version, like v9.9.9 (must not be an exisitng tag).
+          If left blank or incorrect format, archives will be archived instead of being uploaded to Releases.
+        required: false
+        default: ''
 
+# Define top-level environment vars we can refer to below
+env:
+  VERSION: ${{ github.ref_name }}
+  DISPATCH_VERSION: ${{ github.event.inputs.version }}
+  GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      
 jobs:
-
   build:
-
+    name: Build and publish W/ASM artefacts 
     runs-on: ubuntu-latest
-
     steps:
     - uses: actions/checkout@v3
+    # Customizes the Emscripten docker container via the Dockerfile in this repo
     - name: Build the Docker image
       run: docker build -t "docker-emscripten-libzim:v3" ./docker
-    - name: Compile the libzim WASM
-      run: docker run --rm -v $(pwd):/src -u $(id -u):$(id -g) -it docker-emscripten-libzim:v3 make
-    - name: Publish the WASM to draft release
+    # Creates the ASM and WASM artefacts, and the JS wrappers, using the Makefile in this repo
+    - name: Compile the libzim WASM artefacts
+      run: docker run --rm -v $(pwd):/src -u $(id -u):$(id -g) docker-emscripten-libzim:v3 make
+    - name: List directories with updated archives
+      run: |
+        echo -e "\nList ./ :"
+        ls -l
+        echo -e "\nList ./tests/prototype/ :"
+        ls -l tests/prototype/
+        echo -e "\nList ./tests/test_large_file_access/ :"
+        ls -l tests/test_large_file_access/ 
+    # If we are not creating a release, archive the artefacts under this Action run
+    - name: Archive build artefacts
+      if: |
+        github.event_name == 'pull_request' || github.event_name == 'push' && ! startsWith(github.ref_name, 'v')
+        || ! startsWith(github.event.inputs.version, 'v')
+      uses: actions/upload-artifact@v3
+      with:
+        name: libzim-wasm-artefacts
+        path: |
+          libzim-wasm.*
+          libzim-asm.*
+          tests/test_large_file_access/large_file_access.*
+    # Otherwise, zip the artefacts into respective packages (asm and wasm), create and upload releases
+    - name: Zip the artefacts and create draft release
+      id: zip-release
+      if: github.event_name == 'push' && startsWith(github.ref_name, 'v') || startsWith(github.event.inputs.version, 'v')
       run: |
-        echo 'If the run was successful, we will build a publish action here, to publish:'
-        ls -l a.out.*
+        if [[ ! $VERSION =~ ^v?[0-9.]+ ]]; then
+          VERSION=$DISPATCH_VERSION
+        fi
+        # Create a draft release and upload zipped artefacts as release assets
+        chmod +x ./scripts/create_draft_release.sh
+        ./scripts/create_draft_release.sh

.github/workflows/upload_release_assets_to_kiwix.yml

@@ -18,7 +18,6 @@ on:
 # Define top-level environment vars we can refer to below
 env:
   REPOSITORY: ${{ github.repository }}
-  RELEASE: ${{ github.event.release }}
   VERSION: ${{ github.event.release.tag_name }}
   DISPATCH_VERSION: ${{ github.event.inputs.version }}
   GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
@@ -34,14 +33,13 @@ jobs:
     - name: Find the release, get assets, and upload
       run: |
         $repository = $Env:REPOSITORY
-        $release = $Env:RELEASE
         $version = $Env:VERSION
         if (! $version) { $version = $Env:DISPATCH_VERSION }
         if (! $version) {
           Write-Host "`n** You must set a tag version so that we can find and upload the assets! **`n" -ForegroundColor Red
           exit 1
         }
-        echo "`nUsing tag: $TAG_NAME for version upload`n"
+        echo "`nUsing tag: $version for version upload`n"
         $SSH_KEY = $Env:SSH_KEY
         if (! $SSH_KEY) {
           Write-Warning "The SSH secret is empty!"

.gitignore

@@ -1,5 +1,6 @@
+.vscode/
 /nbproject/private/
-/build/
+/build*
 /xz-*/
 /zlib-*/
 /zstd-*/
@@ -11,6 +12,6 @@
 /zstd-*.tar.gz
 /icu4c-*.tgz
 /xapian-core-*.tar.xz
-/libzim-*.tar.gz
+/libzim-*
 ssh_key
 github_token

Makefile

@@ -1,58 +1,81 @@
-all: demo_file_api.js big_file_demo.js
+all: libzim-wasm.dev.js libzim-asm.dev.js libzim-wasm.js libzim-asm.js large_file_access.js
+
 build/lib/liblzma.so : 
-	wget -N https://tukaani.org/xz/xz-5.2.4.tar.gz
+	# Origin: https://tukaani.org/xz/xz-5.2.4.tar.gz
+	[ ! -f xz-*.tar.gz ] && wget -N https://dev.kiwix.org/kiwix-build/xz-5.2.4.tar.gz || true
 	tar xf xz-5.2.4.tar.gz
 	cd xz-5.2.4 ; ./autogen.sh
 	cd xz-5.2.4 ; emconfigure ./configure --prefix=`pwd`/../build
 	cd xz-5.2.4 ; emmake make 
 	cd xz-5.2.4 ; emmake make install
 
-build/lib/libz.a : 
+build/lib/libz.a :
+ 	# Version not yet available in dev.kiwix.org
 	wget -N https://zlib.net/zlib-1.2.13.tar.gz
 	tar xf zlib-1.2.13.tar.gz
 	cd zlib-1.2.13 ; emconfigure ./configure --prefix=`pwd`/../build
 	cd zlib-1.2.13 ; emmake make
 	cd zlib-1.2.13 ; emmake make install
 
-build/lib/libzstd.a : 
-	wget -N https://github.com/facebook/zstd/releases/download/v1.4.4/zstd-1.4.4.tar.gz
-	tar xf zstd-1.4.4.tar.gz
-	cd zstd-1.4.4/build/meson ; meson setup --cross-file=../../../emscripten-crosscompile.ini -Dbin_programs=false -Dbin_contrib=false -Dzlib=disabled -Dlzma=disabled -Dlz4=disabled --prefix=`pwd`/../../../build --libdir=lib builddir
-	cd zstd-1.4.4/build/meson/builddir ; ninja
-	cd zstd-1.4.4/build/meson/builddir ; ninja install
+build/lib/libzstd.a :
+	# Origin: https://github.com/facebook/zstd/releases/download/v1.4.4/zstd-1.4.4.tar.gz 
+	[ ! -f zstd-*.tar.gz ] && wget -N https://dev.kiwix.org/kiwix-build/zstd-1.5.2.tar.gz || true
+	tar xf zstd-1.5.2.tar.gz
+	cd zstd-1.5.2/build/meson ; meson setup --cross-file=../../../emscripten-crosscompile.ini -Dbin_programs=false -Dbin_contrib=false -Dzlib=disabled -Dlzma=disabled -Dlz4=disabled --prefix=`pwd`/../../../build --libdir=lib builddir
+	cd zstd-1.5.2/build/meson/builddir ; ninja
+	cd zstd-1.5.2/build/meson/builddir ; ninja install
 
 build/lib/libicudata.so : 
-	wget -N https://github.com/unicode-org/icu/releases/download/release-69-1/icu4c-69_1-src.tgz
-	tar xf icu4c-69_1-src.tgz
+	# Version not yet available in dev.kiwix.org
+	wget -N https://github.com/unicode-org/icu/releases/download/release-71-1/icu4c-71_1-src.tgz
+	tar xf icu4c-71_1-src.tgz
 	# It's no use trying to compile examples
 	sed -i -e 's/^SUBDIRS =\(.*\)$$(DATASUBDIR) $$(EXTRA) $$(SAMPLE) $$(TEST)\(.*\)/SUBDIRS =\1\2/' icu/source/Makefile.in
 	cd icu/source ; emconfigure ./configure --prefix=`pwd`/../../build
 	cd icu/source ; emmake make 
 	cd icu/source ; emmake make install
 
 build/lib/libxapian.a : build/lib/libz.a
-	wget -N https://oligarchy.co.uk/xapian/1.4.18/xapian-core-1.4.18.tar.xz
+	# Origin: https://oligarchy.co.uk/xapian/1.4.18/xapian-core-1.4.18.tar.xz
+	[ ! -f xapian-*.tar.gz ] && wget -N https://dev.kiwix.org/kiwix-build/xapian-core-1.4.18.tar.xz || true
 	tar xf xapian-core-1.4.18.tar.xz
         # Some options coming from https://github.com/xapian/xapian/tree/master/xapian-core/emscripten
-	#cd xapian-core-1.4.18; emconfigure ./configure --prefix=`pwd`/../build "CFLAGS=-I`pwd`/../build/include -L`pwd`/../build/lib" "CXXFLAGS=-I`pwd`/../build/include -L`pwd`/../build/lib" CPPFLAGS='-DFLINTLOCK_USE_FLOCK' CXXFLAGS='-Oz -s USE_ZLIB=1 -fno-rtti' --disable-backend-honey --disable-backend-inmemory --disable-shared --disable-backend-remote
+	# cd xapian-core-1.4.18; emconfigure ./configure --prefix=`pwd`/../build "CFLAGS=-I`pwd`/../build/include -L`pwd`/../build/lib" "CXXFLAGS=-I`pwd`/../build/include -L`pwd`/../build/lib" CPPFLAGS='-DFLINTLOCK_USE_FLOCK' CXXFLAGS='-Oz -s USE_ZLIB=1 -fno-rtti' --disable-backend-honey --disable-backend-inmemory --disable-shared --disable-backend-remote
 	cd xapian-core-1.4.18; emconfigure ./configure --prefix=`pwd`/../build "CFLAGS=-I`pwd`/../build/include -L`pwd`/../build/lib" "CXXFLAGS=-I`pwd`/../build/include -L`pwd`/../build/lib" --disable-shared --disable-backend-remote
 	cd xapian-core-1.4.18; emmake make "CFLAGS=-I`pwd`/../build/include -L`pwd`/../build/lib -std=c++11" "CXXFLAGS=-I`pwd`/../build/include -L`pwd`/../build/lib -std=c++11"
 	cd xapian-core-1.4.18; emmake make install
 
 build/lib/libzim.a : build/lib/liblzma.so build/lib/libz.a build/lib/libzstd.a build/lib/libicudata.so build/lib/libxapian.a
-	wget -N --content-disposition https://github.com/openzim/libzim/archive/7.2.2.tar.gz
-	tar xf libzim-7.2.2.tar.gz
+	# Origin: wget -N --content-disposition https://github.com/openzim/libzim/archive/7.2.2.tar.gz
+	[ ! -f libzim-*.tar.xz ] && wget -N https://download.openzim.org/release/libzim/libzim-8.1.0.tar.xz || true
+	tar xf libzim-8.1.0.tar.xz
 	# It's no use trying to compile examples
-	sed -i -e "s/^subdir('examples')//" libzim-7.2.2/meson.build
-	cd libzim-7.2.2; PKG_CONFIG_PATH=/src/build/lib/pkgconfig meson --prefix=`pwd`/../build --cross-file=../emscripten-crosscompile.ini . build -DUSE_MMAP=false
-	cd libzim-7.2.2; ninja -C build
-	cd libzim-7.2.2; ninja -C build install
+	sed -i -e "s/^subdir('examples')//" libzim-8.1.0/meson.build
+	cd libzim-8.1.0; PKG_CONFIG_PATH=/src/build/lib/pkgconfig meson --prefix=`pwd`/../build --cross-file=../emscripten-crosscompile.ini . build -DUSE_MMAP=false
+	cd libzim-8.1.0; ninja -C build
+	cd libzim-8.1.0; ninja -C build install
+
+# Development WASM version for testing, completely unoptimized
+libzim-wasm.dev.js: build/lib/libzim.a libzim_bindings.cpp prejs_file_api.js postjs_file_api.js
+	em++ -o libzim-wasm.dev.js --bind libzim_bindings.cpp -I/src/build/include -L/src/build/lib -lzim -llzma -lzstd -lxapian -lz -licui18n -licuuc -licudata -lpthread -lm -fdiagnostics-color=always -pipe -Wall -Winvalid-pch -Wnon-virtual-dtor -Werror -std=c++11 -O0 -g --pre-js prejs_file_api.js --post-js postjs_file_api.js -s WASM=1 -s DISABLE_EXCEPTION_CATCHING=0 -s "EXPORTED_RUNTIME_METHODS=['ALLOC_NORMAL','printErr','ALLOC_STACK','print']" -s DEMANGLE_SUPPORT=1 -s INITIAL_MEMORY=83886080 -s ALLOW_MEMORY_GROWTH=1 -lworkerfs.js
+	cp libzim-wasm.dev.* tests/prototype/
+
+# Development ASM version for testing, completely unoptimized
+libzim-asm.dev.js: build/lib/libzim.a libzim_bindings.cpp prejs_file_api.js postjs_file_api.js
+	em++ -o libzim-asm.dev.js --bind libzim_bindings.cpp -I/src/build/include -L/src/build/lib -lzim -llzma -lzstd -lxapian -lz -licui18n -licuuc -licudata -lpthread -lm -fdiagnostics-color=always -pipe -Wall -Winvalid-pch -Wnon-virtual-dtor -Werror -std=c++11 -O0 -g --pre-js prejs_file_api.js --post-js postjs_file_api.js -s WASM=0 --memory-init-file 0 -s DISABLE_EXCEPTION_CATCHING=0 -s "EXPORTED_RUNTIME_METHODS=['ALLOC_NORMAL','printErr','ALLOC_STACK','print']" -s DEMANGLE_SUPPORT=1 -s INITIAL_MEMORY=83886080 -s ALLOW_MEMORY_GROWTH=1 -lworkerfs.js
+
+# Production WASM version, optimized and packed
+libzim-wasm.js: build/lib/libzim.a libzim_bindings.cpp prejs_file_api.js postjs_file_api.js
+	em++ -o libzim-wasm.js --bind libzim_bindings.cpp -I/src/build/include -L/src/build/lib -lzim -llzma -lzstd -lxapian -lz -licui18n -licuuc -licudata -O3 --pre-js prejs_file_api.js --post-js postjs_file_api.js -s WASM=1 -s "EXPORTED_RUNTIME_METHODS=['ALLOC_NORMAL','printErr','ALLOC_STACK','print']" -s INITIAL_MEMORY=83886080 -s ALLOW_MEMORY_GROWTH=1 -lworkerfs.js
 
-demo_file_api.js: build/lib/libzim.a demo_file_api.cpp prejs_file_api.js postjs_file_api.js
-	em++ --bind demo_file_api.cpp -I/src/build/include -L/src/build/lib -lzim -llzma -lzstd -lxapian -lz -licui18n -licuuc -licudata -lpthread -lm -fdiagnostics-color=always -pipe -Wall -Winvalid-pch -Wnon-virtual-dtor -Werror -std=c++11 -O0 -g --pre-js prejs_file_api.js --post-js postjs_file_api.js -s DISABLE_EXCEPTION_CATCHING=0 -s "EXPORTED_RUNTIME_METHODS=['ALLOC_NORMAL','printErr','ALLOC_STACK','print']" -s DEMANGLE_SUPPORT=1 -s TOTAL_MEMORY=83886080 -s ALLOW_MEMORY_GROWTH=1 -lworkerfs.js
+# Production ASM version, optimized and packed
+libzim-asm.js: build/lib/libzim.a libzim_bindings.cpp prejs_file_api.js postjs_file_api.js
+	em++ -o libzim-asm.js --bind libzim_bindings.cpp -I/src/build/include -L/src/build/lib -lzim -llzma -lzstd -lxapian -lz -licui18n -licuuc -licudata -O3 --pre-js prejs_file_api.js --post-js postjs_file_api.js -s WASM=0 --memory-init-file 0 -s MIN_EDGE_VERSION=40 -s "EXPORTED_RUNTIME_METHODS=['ALLOC_NORMAL','printErr','ALLOC_STACK','print']" -s INITIAL_MEMORY=83886080 -s ALLOW_MEMORY_GROWTH=1 -lworkerfs.js
 
-big_file_demo.js:
-	em++ --bind -std=c++11 -O0 --pre-js prejs_file_api_testbigfile.js --post-js postjs_file_api_testbigfile.js big_file_test.cpp -lworkerfs.js -o bigfile.js
+# Test case: for testing large files
+large_file_access.js: test_file_bindings.cpp prejs_test_file_access.js postjs_test_file_access.js
+	em++ -o large_file_access.js --bind test_file_bindings.cpp -std=c++11 -O0 --pre-js prejs_test_file_access.js --post-js postjs_test_file_access.js -lworkerfs.js
+	cp large_file_access.* tests/test_large_file_access/
 
 clean :
 	rm -rf xz-*
@@ -62,6 +85,5 @@ clean :
 	rm -rf icu*
 	rm -rf libzim-*
 	rm -rf build
-	rm a.out.*
 
 .PHONY : all clean

README.md

@@ -1,26 +1,35 @@
-# Prototype of libzim in Webassembly
+# Prototype of libzim in WebAssembly (WASM)
 
-Demo at https://openzim.github.io/javascript-libzim/
+This Repository provides the source code and utilities for compiling the [ZIM File](https://wiki.openzim.org/wiki/ZIM_file_format) reader [lbizim](https://wiki.openzim.org/wiki/Libzim) from C++ to [WebAssembly](https://developer.mozilla.org/en-US/docs/WebAssembly) (and [ASM.js](https://developer.mozilla.org/en-US/docs/Games/Tools/asm.js)).
 
-It uses WORKERFS as FS with emscripten and run in a web worker, file object is mount before run, and name is passed as argument.
+A prototype in HTML/JS, for testing the WASM version, is provided at https://openzim.github.io/javascript-libzim/tests/prototyep/. This prototype uses WORKERFS as the Emscripten File System and runs in a Web Worker. The file object is mounted before run, and the name is passed as argument.
+
+There is also an HTML/JS utility for testing the ability of Emscripten File Systems to read large files (muliti-gigabyte) at https://openzim.github.io/javascript-libzim/tests/test_large_file_access/.
 
 ## Steps to recompile manually
+
 - Install emscripten : https://kripken.github.io/emscripten-site/docs/getting_started/downloads.html
-- Install dependencies necessary for compilation. On ubuntu 18.04, you need to activate universe repository and :
+- Install dependencies necessary for compilation. On ubuntu 18.04, you need to activate universe repository and:
+
 ```
 sudo apt install ninja-build meson pkg-config python3 autopoint libtool autoconf
 sudo apt install zlib1g-dev libicu-dev libxapian-dev liblzma-dev
 ```
+
 - activate emscripten environment variables with something like `source ./emsdk_env.sh`
 - run `make`
 
 ## Steps to recompile with Docker
-While being at the root of this repository :
- - Build the Docker image with the provided Dockerfile (based on https://hub.docker.com/r/emscripten/emsdk , which is based on Debian) :
+
+While being at the root of this repository:
+ - Build the Docker image with the provided Dockerfile (based on https://hub.docker.com/r/emscripten/emsdk, which is based on Debian):
+
 ```
 docker build -t "docker-emscripten-libzim:v3" ./docker
 ```
- - Run the build with :
+
+ - Run the build with:
+
 ```
 docker run --rm -v $(pwd):/src -v /tmp/emscripten_cache/:/home/emscripten/.emscripten_cache -u $(id -u):$(id -g) -it docker-emscripten-libzim:v3 make
 ```

emscripten-crosscompile.ini

@@ -4,6 +4,7 @@ cpp = 'em++'
 c_ld = 'wasm-ld'
 cpp_ld = 'wasm-ld'
 pkgconfig = 'pkg-config'
+ar = 'emar'
 
 [host_machine]
 system = 'emscripten'

bigfile.js renamed to large_file_access.js

@@ -1503,7 +1503,7 @@ function createExportWrapper(name, fixedasm) {
 }
 
 var wasmBinaryFile;
-  wasmBinaryFile = 'bigfile.wasm';
+  wasmBinaryFile = 'large_file_access.wasm';
   if (!isDataURI(wasmBinaryFile)) {
     wasmBinaryFile = locateFile(wasmBinaryFile);
   }

bigfile.wasm renamed to large_file_access.wasm

@@ -14,7 +14,7 @@ std::shared_ptr<zim::Archive> g_archive;
 
 int main(int argc, char* argv[])
 {
-    std::cout << "wasm initialized" << std::endl;
+    std::cout << "assembler initialized" << std::endl;
     return 0;
 }
 
@@ -95,11 +95,11 @@ std::unique_ptr<EntryWrapper> getEntryByPath(std::string url) {
 }
 
 // Search for a text, and returns the path of the first result
-std::vector<EntryWrapper> search(std::string text) {
+std::vector<EntryWrapper> search(std::string text, int numResults) {
     auto searcher = zim::Searcher(*g_archive);
     auto query = zim::Query(text);
     auto search = searcher.search(query);
-    auto searchResultSet = search.getResults(0,50);
+    auto searchResultSet = search.getResults(0, numResults);
     std::vector<EntryWrapper> ret;
     for(auto entry:searchResultSet) {
         ret.push_back(EntryWrapper(entry));