SimpleDOM support for XML (with RapidXml) and JSON (with rapidjson) (a…

…libaba#445)

* and change ```integration``` into ```ecosystem```
* Use fetchcontent instead of copy code

---------

Co-authored-by: Coldwings <coldwings@me.com>

CMakeLists.txt

@@ -37,6 +37,9 @@ set(PHOTON_FSTACK_SOURCE "" CACHE STRING "")
 set(PHOTON_E2FS_SOURCE "" CACHE STRING "")
 set(PHOTON_GFLAGS_SOURCE "https://github.com/gflags/gflags/archive/refs/tags/v2.2.2.tar.gz" CACHE STRING "")
 set(PHOTON_GOOGLETEST_SOURCE "https://github.com/google/googletest/archive/refs/tags/release-1.12.1.tar.gz" CACHE STRING "")
+set(PHOTON_RAPIDJSON_SOURCE "https://github.com/Tencent/rapidjson/archive/refs/tags/v1.1.0.zip" CACHE STRING "")
+set(PHOTON_RAPIDXML_SOURCE "https://sourceforge.net/projects/rapidxml/files/rapidxml/rapidxml%201.13/rapidxml-1.13.zip/download" CACHE STRING "")
+set(PHOTON_CPP_REDIS_SOURCE "https://github.com/cpp-redis/cpp_redis/archive/refs/tags/4.3.1.tar.gz" CACHE STRING "")
 
 # Get CPU arch and number
 execute_process(COMMAND uname -m OUTPUT_VARIABLE ARCH OUTPUT_STRIP_TRAILING_WHITESPACE)
@@ -156,6 +159,7 @@ endforeach ()
 ################################################################################
 
 add_subdirectory(third_party)
+add_subdirectory(ecosystem)
 
 # Compile photon objects
 file(GLOB PHOTON_SRC
@@ -183,7 +187,7 @@ file(GLOB PHOTON_SRC
         net/security-context/tls-stream.cpp
         rpc/*.cpp
         thread/*.cpp
-        integration/*.cpp
+        ecosystem/*.cpp
         )
 if (APPLE)
     list(APPEND PHOTON_SRC io/kqueue.cpp)
@@ -209,6 +213,7 @@ endif ()
 
 # An object library compiles source files but does not archive or link their object files.
 add_library(photon_obj OBJECT ${PHOTON_SRC})
+target_link_libraries(photon_obj PRIVATE ecosystem_deps)
 target_include_directories(photon_obj PRIVATE include ${OPENSSL_INCLUDE_DIRS} ${AIO_INCLUDE_DIRS}
         ${ZLIB_INCLUDE_DIRS} ${CURL_INCLUDE_DIRS}
 )
@@ -355,6 +360,7 @@ if (PHOTON_BUILD_TESTING)
     add_subdirectory(rpc/test)
     add_subdirectory(thread/test)
     add_subdirectory(net/security-context/test)
+    add_subdirectory(ecosystem/test)
     if (PHOTON_ENABLE_EXTFS)
         add_subdirectory(fs/extfs/test)
     endif ()

common/enumerable.h

@@ -38,11 +38,11 @@ struct Enumerable
     {
         T* obj;
         explicit iterator(T* obj) : obj(obj) {
-          if (obj && obj->next() < 0)
+          if (obj && !obj->valid())
             this->obj = nullptr;
         }
         using R = typename std::result_of<decltype(&T::get)(T)>::type;
-        R operator*() { return obj->get(); }
+        R operator*() { return obj ? obj->get() : nullptr; }
         bool operator==(const iterator& rhs) const { return obj == rhs.obj; }
         bool operator!=(const iterator& rhs) const { return !(*this == rhs); }
         iterator& operator++()
@@ -92,6 +92,7 @@ inline void __example_of_enumerable__()
 {
     struct exam
     {
+        bool valid() { return false; }
         int next() { return -1; }           // move to next, return 0 for success, -1 for failure
         double* get() { return nullptr; }   // get current result
     };

common/iovector.h

@@ -40,8 +40,10 @@ limitations under the License.
 #include <photon/common/io-alloc.h>
 
 #pragma GCC diagnostic push
+#if __GNUC__ >= 13
 #pragma GCC diagnostic ignored "-Wunknown-warning-option"
 #pragma GCC diagnostic ignored "-Wzero-length-bounds"
+#endif
 
 inline bool operator == (const iovec& a, const iovec& b)
 {

common/retval.h

@@ -46,6 +46,9 @@ struct retval : public retval_base {
     operator T() const {
         return get();
     }
+    T operator->() {
+        return get();
+    }
     T get() const {
         return _val;
     }

common/utility.h

@@ -66,6 +66,8 @@ struct ptr_array_t
     T* pend;
     T* begin() { return pbegin; }
     T* end()   { return pend; }
+    T& front() { return pbegin[0]; }
+    T& back()  { return pend[-1]; }
 };
 
 template<typename T>

ecosystem/CMakeLists.txt

@@ -0,0 +1,57 @@
+# Fetch depend code for ecosystem
+# Here for header-only parts.
+
+include(FetchContent)
+
+# Rapidjson
+FetchContent_Declare(
+    rapidjson
+    URL ${PHOTON_RAPIDJSON_SOURCE}
+    URL_HASH SHA256=8e00c38829d6785a2dfb951bb87c6974fa07dfe488aa5b25deec4b8bc0f6a3ab
+    PATCH_COMMAND patch -p1 < ${CMAKE_CURRENT_SOURCE_DIR}/patches/rapidjson.patch
+    DOWNLOAD_EXTRACT_TIMESTAMP ON
+    UPDATE_DISCONNECTED 1)
+# Pre-set rapidjson build option, prevent building docs, examples and tests
+set(RAPIDJSON_BUILD_DOC OFF CACHE BOOL "Build rapidjson documentation.")
+set(RAPIDJSON_BUILD_EXAMPLES OFF CACHE BOOL "Build rapidjson documentation.")
+set(RAPIDJSON_BUILD_TESTS OFF CACHE BOOL "Build rapidjson perftests and unittests.")
+FetchContent_MakeAvailable(rapidjson)
+message(STATUS "Rapidjson source dir: ${rapidjson_SOURCE_DIR}")
+
+# RapidXml
+FetchContent_Declare(
+    rapidxml
+    URL ${PHOTON_RAPIDXML_SOURCE}
+    URL_HASH
+        SHA256=c3f0b886374981bb20fabcf323d755db4be6dba42064599481da64a85f5b3571
+    DOWNLOAD_EXTRACT_TIMESTAMP ON
+    UPDATE_DISCONNECTED 1)
+FetchContent_MakeAvailable(rapidxml)
+message(STATUS "Rapidxml source dir: ${rapidxml_SOURCE_DIR}")
+
+# cpp-redis
+FetchContent_Declare(
+    cpp-redis
+    URL ${PHOTON_CPP_REDIS_SOURCE}
+    URL_HASH
+        SHA256=3859289d8254685fc775bda73de03dad27df923423b8ceb375b02d036c03b02f
+    DOWNLOAD_EXTRACT_TIMESTAMP ON
+    UPDATE_DISCONNECTED 1)
+# uses only a simple header, so do not add sub directory to avoid unnecessary build
+# do not use FetchContent_MakeAvailable, just populate it.
+FetchContent_GetProperties(cpp-redis)
+if(NOT cpp-redis_POPULATED)
+  FetchContent_Populate(cpp-redis)
+endif()
+message(STATUS "cpp-redis source dir: ${cpp-redis_SOURCE_DIR}")
+
+add_library(ecosystem_deps INTERFACE)
+target_include_directories(
+    ecosystem_deps
+    INTERFACE ${rapidjson_SOURCE_DIR}/include ${rapidxml_SOURCE_DIR}
+    ${cpp-redis_SOURCE_DIR}/includes)
+get_property(
+    incs
+    TARGET ecosystem_deps
+    PROPERTY INTERFACE_INCLUDE_DIRECTORIES)
+message(STATUS "ecosystem_deps incs: ${incs}")

ecosystem/patches/rapidjson.patch

@@ -0,0 +1,61 @@
+diff --git a/include/rapidjson/reader.h b/include/rapidjson/reader.h
+index 19f8849b..618492a4 100644
+--- a/include/rapidjson/reader.h
++++ b/include/rapidjson/reader.h
+@@ -153,6 +153,7 @@ enum ParseFlag {
+     kParseNumbersAsStringsFlag = 64,    //!< Parse all numbers (ints/doubles) as strings.
+     kParseTrailingCommasFlag = 128, //!< Allow trailing commas at the end of objects and arrays.
+     kParseNanAndInfFlag = 256,      //!< Allow parsing NaN, Inf, Infinity, -Inf and -Infinity as doubles.
++    kParseBoolsAsStringFlag = 512,  //!< Parse all booleans (true/false) as strings.
+     kParseDefaultFlags = RAPIDJSON_PARSE_DEFAULT_FLAGS  //!< Default parse flags. Can be customized by defining RAPIDJSON_PARSE_DEFAULT_FLAGS
+ };
+
+@@ -201,6 +202,8 @@ struct BaseReaderHandler {
+     bool Default() { return true; }
+     bool Null() { return static_cast<Override&>(*this).Default(); }
+     bool Bool(bool) { return static_cast<Override&>(*this).Default(); }
++    // enabled via kParseBoolsAsStringsFlag, string is not null-terminated (use length)
++    bool RawBool(const Ch* str, SizeType len, bool copy) { return static_cast<Override&>(*this).Default(); }
+     bool Int(int) { return static_cast<Override&>(*this).Default(); }
+     bool Uint(unsigned) { return static_cast<Override&>(*this).Default(); }
+     bool Int64(int64_t) { return static_cast<Override&>(*this).Default(); }
+@@ -714,13 +717,22 @@ private:
+             RAPIDJSON_PARSE_ERROR(kParseErrorValueInvalid, is.Tell());
+     }
+
++    template<unsigned parseFlags, typename InputStream, typename Handler>
++    void ParseRawBools(InputStream& is, Handler& handler) {
++
++    }
++
+     template<unsigned parseFlags, typename InputStream, typename Handler>
+     void ParseTrue(InputStream& is, Handler& handler) {
+         RAPIDJSON_ASSERT(is.Peek() == 't');
++        auto begin = is.PutBegin();
+         is.Take();
+
+         if (RAPIDJSON_LIKELY(Consume(is, 'r') && Consume(is, 'u') && Consume(is, 'e'))) {
+-            if (RAPIDJSON_UNLIKELY(!handler.Bool(true)))
++            auto copy = !(parseFlags & kParseInsituFlag);
++            bool ret = (parseFlags & kParseBoolsAsStringFlag) ?
++                handler.RawBool(begin, 4, copy) : handler.Bool(true);
++            if (RAPIDJSON_UNLIKELY(!ret))
+                 RAPIDJSON_PARSE_ERROR(kParseErrorTermination, is.Tell());
+         }
+         else
+@@ -730,10 +742,14 @@ private:
+     template<unsigned parseFlags, typename InputStream, typename Handler>
+     void ParseFalse(InputStream& is, Handler& handler) {
+         RAPIDJSON_ASSERT(is.Peek() == 'f');
++        auto begin = is.PutBegin();
+         is.Take();
+
+         if (RAPIDJSON_LIKELY(Consume(is, 'a') && Consume(is, 'l') && Consume(is, 's') && Consume(is, 'e'))) {
+-            if (RAPIDJSON_UNLIKELY(!handler.Bool(false)))
++            auto copy = !(parseFlags & kParseInsituFlag);
++            bool ret = (parseFlags & kParseBoolsAsStringFlag) ?
++                handler.RawBool(begin, 5, copy) : handler.Bool(false);
++            if (RAPIDJSON_UNLIKELY(!ret))
+                 RAPIDJSON_PARSE_ERROR(kParseErrorTermination, is.Tell());
+         }
+         else

integration/redis.cpp → ecosystem/redis.cpp

@@ -14,12 +14,12 @@ See the License for the specific language governing permissions and
 limitations under the License.
 */
 
-#include <photon/integration/redis.h>
+#include <photon/ecosystem/redis.h>
 #include <inttypes.h>
 #include <memory>
 #include <photon/net/socket.h>
 #include <photon/common/alog.h>
-#include "cpp-redis/network/tcp_client_iface.hpp"
+#include <cpp_redis/network/tcp_client_iface.hpp>
 
 namespace photon {
 using namespace net;