Merge pull request #20 from GordonSmith/EXPAT_WASM

feat(xml): Add simple XML parser based on expat
hpcc-systems · Mar 22, 2020 · 41fbcc1 · 41fbcc1
2 parents dc893c6 + cafb0f6
commit 41fbcc1
Show file tree

Hide file tree

Showing 13 changed files with 1,260 additions and 4 deletions.
diff --git a/.vscode/tasks.json b/.vscode/tasks.json
@@ -0,0 +1,46 @@
+{
+    // See https://go.microsoft.com/fwlink/?LinkId=733558 
+    // for the documentation about the tasks.json format
+    "version": "2.0.0",
+    "tasks": [
+        {
+            "type": "npm",
+            "label": "Compile Watch",
+            "script": "compile-es6-watch",
+            "problemMatcher": [
+                "$tsc-watch"
+            ],
+            "presentation": {
+                "group": "group-build"
+            }
+        },
+        {
+            "type": "npm",
+            "label": "Bundle Watch",
+            "script": "bundle-watch",
+            "problemMatcher": [],
+            "presentation": {
+                "group": "group-build"
+            }
+        },
+        {
+            "type": "npm",
+            "label": "Web Server",
+            "script": "dev-start",
+            "problemMatcher": [],
+            "presentation": {
+                "group": "group-build"
+            }
+        },
+        {
+            "label": "build",
+            "dependsOn": [
+                "Compile Watch",
+                "Bundle Watch",
+                "Web Server"
+            ],
+            "group": "build",
+            "problemMatcher": []
+        }
+    ]
+}
diff --git a/README.md b/README.md
@@ -15,6 +15,7 @@ npm install --save @hpcc-js/wasm
 * `index.js` / `index.min.js` files:  Exposes _all_ the available APIs for all WASM files.
 * WASM Files:
     * `graphvizlib.wasm`
+    * `expatlib.wasm`
     * ...more to follow...
 
 **Important**:  WASM files are dynamically loaded at runtime (this is a browser / emscripten requirement), which has a few implications for the consumer:  
@@ -29,6 +30,7 @@ npm install --save @hpcc-js/wasm
 ## API Reference
 * [Common](#common)
 * [GraphViz](#graphviz)
+* [Expat](#expat)
 
 ### Common
 Utility functions relating to @hpcc-js/wasm as a package
@@ -201,6 +203,58 @@ Convenience function that performs **patchwork** layout, is equivalent to `layou
 
 Convenience function that performs **twopi** layout, is equivalent to `layout(dotSource, outputFormat, "twopi");`.
 
+### Expat (`expatlib.wasm`)
+Expat WASM library, provides a simplified wrapper around the Expat XML Parser library, see [libexpat.github.io](https://libexpat.github.io/) for c++ details.
+
+#### Hello World
+```html
+<!DOCTYPE html>
+<html>
+
+<head>
+    <meta charset="UTF-8">
+    <title>GraphViz WASM</title>
+    <script src="https://unpkg.com/@hpcc-js/wasm/dist/index.min.js"></script>
+    <script>
+        var hpccWasm = window["@hpcc-js/wasm"];
+    </script>
+</head>
+
+<body>
+    <script>
+        const xml = `
+            <root>
+                <child xxx="yyy">content</child>
+            </root>
+        `;
+
+        var callback = {
+            startElement(tag, attrs) { console.log("start", tag, attrs); },
+            endElement(tag) { console.log("end", tag); },
+            characterData(content) { console.log("characterData", content); }
+        };
+        hpccWasm.parse(xml, callback);
+    </script>
+
+</body>
+
+</html>
+```
+
+#### Expat API
+
+<a name="parse" href="#parse">#</a> **parse**(_xml_, _callback_) · [<>](https://github.com/hpcc-systems/hpcc-js-wasm/blob/master/src/expat.ts "Source")
+
+* **_xml_**:  XML String.
+* **_callback_**:  Callback Object with the following methods:
+    * **startElement**(_tag_: string, _attrs_: {[key: string]: string]): void;
+    * **endElement**(_tag_: string): void;
+    * **characterData**(_content_: string): void;
+
+Parses the XML with suitable callbacks.
+
+**Note:** _characterData_ may get called several times for a single tag element.
+
 ## Building @hpcc-js/wasm
 _Building is supported on both Linux (tested with Ubuntu 18.04) and Windows with WSL enabled (Ubuntu-18.04).  Building in other environments should work, but may be missing certain prerequisites._
 

diff --git a/cpp/expat/CMakeLists.txt b/cpp/expat/CMakeLists.txt
@@ -15,3 +15,5 @@ ADD_LIBRARY(expat STATIC
     ${EXPAT_LIB_DIR}/xmltok_ns.c
     ${EXPAT_LIB_DIR}/xmltok.c
 )
+
+ADD_SUBDIRECTORY(expatlib)
diff --git a/cpp/expat/expatlib/CMakeLists.txt b/cpp/expat/expatlib/CMakeLists.txt
@@ -0,0 +1,33 @@
+PROJECT(expatlib)
+
+SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -s WASM=1 -s INVOKE_RUN=0 -s ENVIRONMENT=web -s ALLOW_MEMORY_GROWTH=1")
+SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -s MODULARIZE=1 -s EXPORT_NAME='${CMAKE_PROJECT_NAME}'")
+
+#  Generate Glue from IDL file  ---
+ADD_CUSTOM_COMMAND(
+    MAIN_DEPENDENCY ${CMAKE_CURRENT_SOURCE_DIR}/main.idl
+    OUTPUT ${CMAKE_CURRENT_BINARY_DIR}/main_glue.js ${CMAKE_CURRENT_BINARY_DIR}/main_glue.cpp
+    COMMAND python ${CMAKE_BINARY_DIR}/../emsdk/upstream/emscripten/tools/webidl_binder.py ${CMAKE_CURRENT_SOURCE_DIR}/main.idl ${CMAKE_CURRENT_BINARY_DIR}/main_glue
+)
+SET_PROPERTY(SOURCE main.cpp APPEND PROPERTY OBJECT_DEPENDS ${CMAKE_CURRENT_BINARY_DIR}/main_glue.cpp)
+SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} --post-js ${CMAKE_CURRENT_BINARY_DIR}/main_glue.js")
+#  ---  ---  ---
+
+INCLUDE_DIRECTORIES(
+    ${CMAKE_CURRENT_BINARY_DIR}
+    ${EXPAT_LIB_DIR}
+)
+
+ADD_EXECUTABLE(expatlib
+    main.cpp
+)
+
+TARGET_LINK_LIBRARIES(expatlib
+    expat
+)
+
+INSTALL(FILES 
+    ${CMAKE_CURRENT_BINARY_DIR}/expatlib.wasm
+    DESTINATION dist
+    COMPONENT runtime
+)
diff --git a/cpp/expat/expatlib/main.cpp b/cpp/expat/expatlib/main.cpp
@@ -0,0 +1,100 @@
+#include "stack_parser.h"
+
+#include <string>
+
+class CExpat : public CExpatImpl<CExpat>
+{
+private:
+    typedef CExpatImpl<CExpat> BaseClass;
+
+protected:
+    std::string m_tag;
+    std::string m_attrs;
+    std::string m_content;
+
+public:
+    CExpat()
+    {
+    }
+
+    void OnPostCreate()
+    {
+        EnableStartElementHandler();
+        EnableEndElementHandler();
+        EnableCharacterDataHandler();
+    }
+
+    bool create()
+    {
+        return BaseClass::Create();
+    }
+
+    void destroy()
+    {
+        BaseClass::Destroy();
+    }
+
+    bool parse(const char *xml)
+    {
+        return BaseClass::Parse(xml, (int)strlen(xml), XML_TRUE);
+    }
+
+    const char *tag() const
+    {
+        return m_tag.c_str();
+    }
+
+    const char *attrs() const
+    {
+        return m_attrs.c_str();
+    }
+
+    const char *content() const
+    {
+        return m_content.c_str();
+    }
+
+    virtual void startElement()
+    {
+    }
+
+    virtual void endElement()
+    {
+    }
+
+    virtual void characterData()
+    {
+    }
+
+    virtual void OnStartElement(const XML_Char *pszName, const XML_Char **papszAttrs)
+    {
+        m_tag = pszName;
+        m_attrs = "";
+        for (XML_Char **itr = (XML_Char **)papszAttrs; *itr != NULL; itr += 2)
+        {
+            if (!m_attrs.empty())
+            {
+                m_attrs += "\1\1";
+            }
+            m_attrs += *itr;
+            m_attrs += "\1";
+            m_attrs += *(itr + 1);
+        }
+        startElement();
+    }
+
+    virtual void OnEndElement(const XML_Char *pszName)
+    {
+        m_tag = pszName;
+        endElement();
+    }
+
+    virtual void OnCharacterData(const XML_Char *pszData, int nLength)
+    {
+        m_content.assign(pszData, nLength);
+        characterData();
+    }
+};
+
+//  Include JS Glue  ---
+#include "main_glue.cpp"
diff --git a/cpp/expat/expatlib/main.idl b/cpp/expat/expatlib/main.idl
@@ -0,0 +1,21 @@
+interface CExpat
+{
+  void CExpat();
+  boolean create();
+  void destroy();
+  boolean parse([Const] DOMString xml);
+  [Const] DOMString tag();
+  [Const] DOMString attrs();
+  [Const] DOMString content();
+  void startElement();
+  void endElement();
+  void characterData();
+};
+
+[JSImplementation = "CExpat"] 
+interface CExpatJS {
+  void CExpatJS();
+  void startElement();
+  void endElement();
+  void characterData();
+};