深海游弋的鱼 – 第 80 页 – 智障儿童欢乐多

macOS Mojave (10.14.3) Android Studio 3.3.1 NDK 19.1.5304403 导入并构建Vuh项目

以前在 Android Studio 3.2.1上vuh库使用的例子中实现了一个使用 vuh 库的例子。那个例子中的 vuh 库是我们编译好 libvuh.so 之后直接引用的，我们下面实现通过直接编译代码实现整合。

尝试过使用 ExternalProject_add 跟 include 的方式包含 vuh 库，但是都不是很成功。

其中 ExternalProject_add 导入的项目只能编译一次，即使指定 BUILD_ALWAYS 1 也没用，这个应该是 Ninja 导致的问题，导致当出现多个 ABI 或者 vuh 库代码变动之后，不能重新编译，出现各种编译错误。

使用 include 包含的项目会导致路径信息不正确，无法找到源代码文件。

最后使用 add_subdirectory实现。

修改之后的几个关键文件如下：

注意： VUH_ROOT_DIR 这个变量中指定 vuh 库代码的位置

# For more information about using CMake with Android Studio, read the
# documentation: https://d.android.com/studio/projects/add-native-code.html

# Sets the minimum version of CMake required to build the native library.

cmake_minimum_required(VERSION 3.8)

# for Vulkan
SET(Vulkan_INCLUDE_DIR ${ANDROID_NDK}/sources/third_party/vulkan/src/include/)
SET(Vulkan_LIBRARIES ${ANDROID_NDK}/platforms/${ANDROID_PLATFORM}/arch-${ANDROID_ARCH_NAME}/usr/lib)
if(X86_64)
    SET(Vulkan_LIBRARIES ${Vulkan_LIBRARIES}64)
endif()
SET(Vulkan_LIBRARIES ${Vulkan_LIBRARIES}/libvulkan.so)
add_library(vulkan SHARED IMPORTED)
set_target_properties(vulkan PROPERTIES IMPORTED_LOCATION ${Vulkan_LIBRARIES})

# for vuh
add_definitions(-DVK_USE_PLATFORM_ANDROID_KHR=1 -DVULKAN_HPP_TYPESAFE_CONVERSION=1)
SET(VUH_ROOT_DIR ${CMAKE_CURRENT_SOURCE_DIR}/../../vuh/)
SET(VUH_BUILD_TESTS OFF)
SET(VUH_BUILD_DOCS OFF)
SET(VUH_BUILD_EXAMPLES OFF)

add_subdirectory(${VUH_ROOT_DIR}src/ ${CMAKE_CURRENT_SOURCE_DIR}/build/)

# for example
SET(Vuh_INCLUDE_PATH ${VUH_ROOT_DIR}src/include)
include_directories(${Vulkan_INCLUDE_DIR})
include_directories(${Vuh_INCLUDE_PATH})

# Creates and names a library, sets it as either STATIC
# or SHARED, and provides the relative paths to its source code.
# You can define multiple libraries, and CMake builds them for you.
# Gradle automatically packages shared libraries with your APK.

add_library( # Sets the name of the library.
        native-lib

# Sets the library as a shared library.
        SHARED

# Provides a relative path to your source file(s).
        src/main/cpp/native-lib.cpp)

add_dependencies(native-lib vuh)

# Searches for a specified prebuilt library and stores the path as a
# variable. Because CMake includes system libraries in the search path by
# default, you only need to specify the name of the public NDK library
# you want to add. CMake verifies that the library exists before
# completing its build.

find_library( # Sets the name of the path variable.
        log-lib

# Specifies the name of the NDK library that
        # you want CMake to locate.
        log

android)

# Specifies libraries CMake should link to your target library. You
# can link multiple libraries, such as libraries you define in this
# build script, prebuilt third-party libraries, or system libraries.

target_link_libraries( # Specifies the target library.
        native-lib

vulkan

vuh

${log-lib})

# For more information about using CMake with Android Studio, read the

# documentation: https://d.android.com/studio/projects/add-native-code.html

# Sets the minimum version of CMake required to build the native library.

cmake_minimum_required(VERSION 3.8)

# for Vulkan

SET(Vulkan_INCLUDE_DIR ${ANDROID_NDK}/sources/third_party/vulkan/src/include/)

SET(Vulkan_LIBRARIES ${ANDROID_NDK}/platforms/${ANDROID_PLATFORM}/arch-${ANDROID_ARCH_NAME}/usr/lib)

if(X86_64)

SET(Vulkan_LIBRARIES ${Vulkan_LIBRARIES}64)

endif()

SET(Vulkan_LIBRARIES ${Vulkan_LIBRARIES}/libvulkan.so)

add_library(vulkan SHARED IMPORTED)

set_target_properties(vulkan PROPERTIES IMPORTED_LOCATION ${Vulkan_LIBRARIES})

# for vuh

add_definitions(-DVK_USE_PLATFORM_ANDROID_KHR=1 -DVULKAN_HPP_TYPESAFE_CONVERSION=1)

SET(VUH_ROOT_DIR ${CMAKE_CURRENT_SOURCE_DIR}/../../vuh/)

SET(VUH_BUILD_TESTS OFF)

SET(VUH_BUILD_DOCS OFF)

SET(VUH_BUILD_EXAMPLES OFF)

add_subdirectory(${VUH_ROOT_DIR}src/ ${CMAKE_CURRENT_SOURCE_DIR}/build/)

# for example

SET(Vuh_INCLUDE_PATH ${VUH_ROOT_DIR}src/include)

include_directories(${Vulkan_INCLUDE_DIR})

include_directories(${Vuh_INCLUDE_PATH})

# Creates and names a library, sets it as either STATIC

# or SHARED, and provides the relative paths to its source code.

# You can define multiple libraries, and CMake builds them for you.

# Gradle automatically packages shared libraries with your APK.

add_library( # Sets the name of the library.

native-lib

# Sets the library as a shared library.

SHARED

# Provides a relative path to your source file(s).

src/main/cpp/native-lib.cpp)

add_dependencies(native-lib vuh)

# Searches for a specified prebuilt library and stores the path as a

# variable. Because CMake includes system libraries in the search path by

# default, you only need to specify the name of the public NDK library

# you want to add. CMake verifies that the library exists before

# completing its build.

find_library( # Sets the name of the path variable.

log-lib

# Specifies the name of the NDK library that

# you want CMake to locate.

log

android)

# Specifies libraries CMake should link to your target library. You

# can link multiple libraries, such as libraries you define in this

# build script, prebuilt third-party libraries, or system libraries.

target_link_libraries( # Specifies the target library.

native-lib

vulkan

vuh

${log-lib})

注意：由于 vuh 库需要 CMake 3.8 。因此，我们需要手工指定CMake版本为3.10.2 。

如下：

apply plugin: 'com.android.application'

android {
    compileSdkVersion 28
    defaultConfig {
        applicationId "com.mobibrw.vuhandroid"
        minSdkVersion 24
        targetSdkVersion 28
        versionCode 1
        versionName "1.0"
        testInstrumentationRunner "android.support.test.runner.AndroidJUnitRunner"
        externalNativeBuild {
            cmake {
                version "3.10.2"
                cppFlags "-std=c++14 -v -g"
            }
        }
    }

buildTypes {
        release {
            minifyEnabled false
            proguardFiles getDefaultProguardFile('proguard-android.txt'), 'proguard-rules.pro'
        }
    }
    externalNativeBuild {
        cmake {
            version "3.10.2"
            path "CMakeLists.txt"
        }
    }
}

dependencies {
    implementation fileTree(dir: 'libs', include: ['*.jar'])
    implementation 'com.android.support:appcompat-v7:28.0.0'
    implementation 'com.android.support.constraint:constraint-layout:1.1.3'
    testImplementation 'junit:junit:4.12'
    androidTestImplementation 'com.android.support.test:runner:1.0.2'
    androidTestImplementation 'com.android.support.test.espresso:espresso-core:3.0.2'
}

apply plugin: 'com.android.application'

android {

compileSdkVersion 28

defaultConfig {

applicationId "com.mobibrw.vuhandroid"

minSdkVersion 24

targetSdkVersion 28

versionCode 1

versionName "1.0"

testInstrumentationRunner "android.support.test.runner.AndroidJUnitRunner"

externalNativeBuild {

cmake {

version "3.10.2"

cppFlags "-std=c++14 -v -g"

}

buildTypes {

release {

minifyEnabled false

proguardFiles getDefaultProguardFile('proguard-android.txt'), 'proguard-rules.pro'

}

externalNativeBuild {

cmake {

version "3.10.2"

path "CMakeLists.txt"

}

dependencies {

implementation fileTree(dir: 'libs', include: ['*.jar'])

implementation 'com.android.support:appcompat-v7:28.0.0'

implementation 'com.android.support.constraint:constraint-layout:1.1.3'

testImplementation 'junit:junit:4.12'

androidTestImplementation 'com.android.support.test:runner:1.0.2'

androidTestImplementation 'com.android.support.test.espresso:espresso-core:3.0.2'

}

如果出现如下错误：

1 2	CMake Error: CMake was unable to find a build program corresponding to "Ninja". CMAKE_MAKE_PROGRAM is not set. You probably need to select a different build tool. -- Configuring incomplete, errors occurred!

则执行如下操作：

1	$ brew install ninja

如果出现如下错误：

* What went wrong:

Execution failed for task ':app:transformNativeLibsWithMergeJniLibsForDebug'.

> More than one file was found with OS independent path 'lib/armeabi-v7a/libvuh.so'

则删除代码中的 jniLibs/armeabi-v7a/libvuh.so 即可解决问题。

完整的例子点击此处下载 vuhAndroid

参考链接

macOS Mojave (10.14.3) Android Studio 3.3.1 指定使用CMake 3.10.2版本

目前最新版本的 Android Studio 3.3.1默认使用CMake 3.6版本，但是已经支持 CMake 3.10.2 版本了。

新版本的 CMake 3.10.2 新增了 FindVulkan.cmake 等一系列的改进，对于很多项目来说，会更友好。

目前默认依旧使用 CMake 3.6 版本，但是可以手工指定使用 CMake 3.10.2。

继续阅读macOS Mojave (10.14.3) Android Studio 3.3.1 指定使用CMake 3.10.2版本

错误程序包org.apache.http.conn.util不存在

开发的时候遇到了这个问题，记录下下次就知道了。这个是因为在 sdk22 以后的版本不在支持了！！！

程序包 org.apache.http.conn.util 不存在，然后搜索了下，在 app 的 build.gradle 文件中加上useLibrary 'org.apache.http.legacy' 就可以了。

如下：

android {

compileSdkVersion 23

buildToolsVersion '23.0.3'

useLibrary 'org.apache.http.legacy'

defaultConfig {

applicationId "com.ah.woyi"

minSdkVersion 14

targetSdkVersion 21

versionCode 1

versionName "1.0"

}

参考链接

错误程序包org.apache.http.conn.util不存在

Parsing C++ in Python with Clang

macOS Mojave 10.14.2 llvm 7.0.1

$ brew install llvm --with-clang --with-python --with-lld

$ export DYLD_LIBRARY_PATH=/usr/local/opt/llvm/lib

$ export LLVM_VER=$(brew list --versions | grep llvm | sed 's/$[0-9]$.*/\1/g' | tr -cd "[0-9]")

$ pip install clang

#!/usr/bin/env python

# -*- coding: utf-8 -*-

""" Usage: call with <filename> <typename>

"""

import sys

import clang.cindex

# 也可以使用如下命令指定libclang的加载位置,这个动态库可以直接拷贝到其他机器上使用，是个独立库

# clang.cindex.Config.set_library_path('/usr/local/opt/llvm/lib')

# 如下可以简化代码

# from clang.cindex import CursorKind

# from clang.cindex import AccessSpecifier

def find_typerefs(node, typename):

""" Find all references to the type named 'typename'

"""

if node.kind.is_reference():

ref_node = node.get_definition()

if ref_node :

if ref_node.spelling == typename:

print 'Found %s [line=%s, col=%s]' % (

typename, node.location.line, node.location.column)

# Recurse for children of this node

for c in node.get_children():

find_typerefs(c, typename)

index = clang.cindex.Index.create()

tu = index.parse(sys.argv[1])

print 'Translation unit:', tu.spelling

find_typerefs(tu.cursor, sys.argv[2])

#!/usr/bin/env python
# -*- coding: utf-8 -*-
""" C++ linter using libclang. Call with [filenames] """
from __future__ import print_function

from clang.cindex import Config, TypeKind, CursorKind, Index
from pprint import pprint
import platform
import sys

if platform.system() == "Linux":
    Config.set_library_file("/usr/lib/llvm-3.4/lib/libclang.so")
else:
    Config.set_library_file("/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/lib/libclang.dylib")

def error(out_errors, node, msg):
    error_msg = "{}:{}:  {}".format(node.location.file, node.location.line, msg)
    out_errors.append(error_msg)
    print(error_msg, file=sys.stderr)

def is_mut_ref(arg_type):
    if arg_type.kind in [TypeKind.POINTER, TypeKind.LVALUEREFERENCE,
                         TypeKind.INCOMPLETEARRAY, TypeKind.CONSTANTARRAY]:

if arg_type.kind in [TypeKind.POINTER, TypeKind.LVALUEREFERENCE]:
            pointee_type = arg_type.get_pointee()
        else:
            pointee_type = arg_type.get_array_element_type()

# print("pointee_type.kind: {}".format(pointee_type.kind))
        # print("pointee_type.is_const_qualified(): {}".format(pointee_type.is_const_qualified()))

if not pointee_type.is_const_qualified():
            return True

return False

def check_argument(out_errors, function, node, function_parse_progress):
    assert node.kind == CursorKind.PARM_DECL

if function.kind == CursorKind.FUNCTION_DECL and function.spelling == "main":
        # Ignore main function
        return

# print("")
    # print("node.spelling: {}".format(node.spelling))
    # print("node.type.kind: {}".format(node.type.kind))
    # print("node.type.get_ref_qualifier(): {}".format(node.type.get_ref_qualifier()))
    # print("node.type.is_const_qualified(): {}".format(node.type.is_const_qualified()))
    # pprint(dir(node))

name = node.spelling
    if not name:
        # Ignore nameless arguments (e.g. Foo(Foo&))
        return

if is_mut_ref(node.type):
        if name.startswith("o_"):
            function_parse_progress["state"] = "out"
        elif name.startswith("io_"):
            if function_parse_progress["state"] != "io":
                error(out_errors, node, "io_ arguments should be first")
            function_parse_progress["state"] = "io"

else:
            error(out_errors, node,
                  "Non-const reference/pointer/array argument should be prefixed with " \
                  "either o_ (for out) or io_ (for in-out), e.g. 'o_{}'".format(name))
    else:
        if function_parse_progress["state"] == "out":
            error(out_errors, node, "input arguments should come before output arguments")
        function_parse_progress["state"] = "in"

def do_lint(out_errors, node, root_file):
    if node.location.file and node.location.file.name != root_file.name:
        # This is ugly, but works.
        return

# print("{}:{} node.kind: {}".format(node.location.file, node.location.line, node.kind))
    # print("node.translation_unit: {}".format(node.translation_unit))
    # # pprint(dir(node.translation_unit))
    # print("node.location.file: {}".format(node.location.file))
    # pprint(dir(node.location))
    # exit()

# CursorKind.CONSTRUCTOR excluded: references there are often stored, so not o_ or io_
    if node.kind in [CursorKind.FUNCTION_DECL, CursorKind.CXX_METHOD]:
        # print("Found a function!")
        # print("node.spelling: {}".format(node.spelling))
        # print("node.displayname: {}".format(node.displayname))

function_parse_progress = {
            "state": "io", # "io", "in" or "out"
        }

for arg in node.get_arguments():
            check_argument(out_errors, node, arg, function_parse_progress)

# Recurse for children of this node
    for c in node.get_children():
        do_lint(out_errors, c, root_file)

def lint_file(filepath):
    index = Index.create()
    tu = index.parse(filepath)
    root_file = tu.get_file(tu.spelling)

errors = []
    do_lint(errors, tu.cursor, root_file)
    return errors

def main():
    if len(sys.argv) == 1:
        print("Usage: {} [filenames]".format(sys.argv[0]))
        return

for filepath in sys.argv[1:]:
        lint_file(filepath)

if __name__ == "__main__":
    main()

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

#!/usr/bin/env python

# -*- coding: utf-8 -*-

""" C++ linter using libclang. Call with [filenames] """

from __future__ import print_function

from clang.cindex import Config, TypeKind, CursorKind, Index

from pprint import pprint

import platform

import sys

if platform.system() == "Linux":

Config.set_library_file("/usr/lib/llvm-3.4/lib/libclang.so")

else:

Config.set_library_file("/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/lib/libclang.dylib")

def error(out_errors, node, msg):

error_msg = "{}:{}: {}".format(node.location.file, node.location.line, msg)

out_errors.append(error_msg)

print(error_msg, file=sys.stderr)

def is_mut_ref(arg_type):

if arg_type.kind in [TypeKind.POINTER, TypeKind.LVALUEREFERENCE,

TypeKind.INCOMPLETEARRAY, TypeKind.CONSTANTARRAY]:

if arg_type.kind in [TypeKind.POINTER, TypeKind.LVALUEREFERENCE]:

pointee_type = arg_type.get_pointee()

else:

pointee_type = arg_type.get_array_element_type()

# print("pointee_type.kind: {}".format(pointee_type.kind))

# print("pointee_type.is_const_qualified(): {}".format(pointee_type.is_const_qualified()))

if not pointee_type.is_const_qualified():

return True

return False

def check_argument(out_errors, function, node, function_parse_progress):

assert node.kind == CursorKind.PARM_DECL

if function.kind == CursorKind.FUNCTION_DECL and function.spelling == "main":

# Ignore main function

return

# print("")

# print("node.spelling: {}".format(node.spelling))

# print("node.type.kind: {}".format(node.type.kind))

# print("node.type.get_ref_qualifier(): {}".format(node.type.get_ref_qualifier()))

# print("node.type.is_const_qualified(): {}".format(node.type.is_const_qualified()))

# pprint(dir(node))

name = node.spelling

if not name:

# Ignore nameless arguments (e.g. Foo(Foo&))

return

if is_mut_ref(node.type):

if name.startswith("o_"):

function_parse_progress["state"] = "out"

elif name.startswith("io_"):

if function_parse_progress["state"] != "io":

error(out_errors, node, "io_ arguments should be first")

function_parse_progress["state"] = "io"

else:

error(out_errors, node,

"Non-const reference/pointer/array argument should be prefixed with " \

"either o_ (for out) or io_ (for in-out), e.g. 'o_{}'".format(name))

else:

if function_parse_progress["state"] == "out":

error(out_errors, node, "input arguments should come before output arguments")

function_parse_progress["state"] = "in"

def do_lint(out_errors, node, root_file):

if node.location.file and node.location.file.name != root_file.name:

# This is ugly, but works.

return

# print("{}:{} node.kind: {}".format(node.location.file, node.location.line, node.kind))

# print("node.translation_unit: {}".format(node.translation_unit))

# # pprint(dir(node.translation_unit))

# print("node.location.file: {}".format(node.location.file))

# pprint(dir(node.location))

# exit()

# CursorKind.CONSTRUCTOR excluded: references there are often stored, so not o_ or io_

if node.kind in [CursorKind.FUNCTION_DECL, CursorKind.CXX_METHOD]:

# print("Found a function!")

# print("node.spelling: {}".format(node.spelling))

# print("node.displayname: {}".format(node.displayname))

function_parse_progress = {

"state": "io", # "io", "in" or "out"

}

for arg in node.get_arguments():

check_argument(out_errors, node, arg, function_parse_progress)

# Recurse for children of this node

for c in node.get_children():

do_lint(out_errors, c, root_file)

def lint_file(filepath):

index = Index.create()

tu = index.parse(filepath)

root_file = tu.get_file(tu.spelling)

errors = []

do_lint(errors, tu.cursor, root_file)

return errors

def main():

if len(sys.argv) == 1:

print("Usage: {} [filenames]".format(sys.argv[0]))

return

for filepath in sys.argv[1:]:

lint_file(filepath)

if __name__ == "__main__":

main()

#!/usr/bin/env python

import clang.cindex
import sys
import os
import logging
import argparse
import fnmatch

from DotGenerator import *

index = clang.cindex.Index.create()
dotGenerator = DotGenerator()

def findFilesInDir(rootDir, patterns):
    """ Searches for files in rootDir which file names mathes the given pattern. Returns
    a list of file paths of found files"""
    foundFiles = []
    for root, dirs, files in os.walk(rootDir):
        for p in patterns:
            for filename in fnmatch.filter(files, p):
                foundFiles.append(os.path.join(root, filename))
    return foundFiles

def processClassField(cursor):
    """ Returns the name and the type of the given class field.
    The cursor must be of kind CursorKind.FIELD_DECL"""
    type = None
    fieldChilds = list(cursor.get_children())
    if len(fieldChilds) == 0:  # if there are not cursorchildren, the type is some primitive datatype
        type = cursor.type.spelling
    else:  # if there are cursorchildren, the type is some non-primitive datatype (a class or class template)
        for cc in fieldChilds:
            if cc.kind == clang.cindex.CursorKind.TEMPLATE_REF:
                type = cc.spelling
            elif cc.kind == clang.cindex.CursorKind.TYPE_REF:
                type = cursor.type.spelling
    name = cursor.spelling
    return name, type

def processClassMemberDeclaration(umlClass, cursor):
    """ Processes a cursor corresponding to a class member declaration and
    appends the extracted information to the given umlClass """
    if cursor.kind == clang.cindex.CursorKind.CXX_BASE_SPECIFIER:
        for baseClass in cursor.get_children():
            if baseClass.kind == clang.cindex.CursorKind.TEMPLATE_REF:
                umlClass.parents.append(baseClass.spelling)
            elif baseClass.kind == clang.cindex.CursorKind.TYPE_REF:
                umlClass.parents.append(baseClass.type.spelling)
    elif cursor.kind == clang.cindex.CursorKind.FIELD_DECL:  # non static data member
        name, type = processClassField(cursor)
        if name is not None and type is not None:
            # clang < 3.5: needs patched cindex.py to have
            # clang.cindex.AccessSpecifier available:
            # https://gitorious.org/clang-mirror/clang-mirror/commit/e3d4e7c9a45ed9ad4645e4dc9f4d3b4109389cb7
            if cursor.access_specifier == clang.cindex.AccessSpecifier.PUBLIC:
                umlClass.publicFields.append((name, type))
            elif cursor.access_specifier == clang.cindex.AccessSpecifier.PRIVATE:
                umlClass.privateFields.append((name, type))
            elif cursor.access_specifier == clang.cindex.AccessSpecifier.PROTECTED:
                umlClass.protectedFields.append((name, type))
    elif cursor.kind == clang.cindex.CursorKind.CXX_METHOD:
        try:
            returnType, argumentTypes = cursor.type.spelling.split(' ', 1)
            if cursor.access_specifier == clang.cindex.AccessSpecifier.PUBLIC:
                umlClass.publicMethods.append((returnType, cursor.spelling, argumentTypes))
            elif cursor.access_specifier == clang.cindex.AccessSpecifier.PRIVATE:
                umlClass.privateMethods.append((returnType, cursor.spelling, argumentTypes))
            elif cursor.access_specifier == clang.cindex.AccessSpecifier.PROTECTED:
                umlClass.protectedMethods.append((returnType, cursor.spelling, argumentTypes))
        except:
            logging.error("Invalid CXX_METHOD declaration! " + str(cursor.type.spelling))
    elif cursor.kind == clang.cindex.CursorKind.FUNCTION_TEMPLATE:
        returnType, argumentTypes = cursor.type.spelling.split(' ', 1)
        if cursor.access_specifier == clang.cindex.AccessSpecifier.PUBLIC:
            umlClass.publicMethods.append((returnType, cursor.spelling, argumentTypes))
        elif cursor.access_specifier == clang.cindex.AccessSpecifier.PRIVATE:
            umlClass.privateMethods.append((returnType, cursor.spelling, argumentTypes))
        elif cursor.access_specifier == clang.cindex.AccessSpecifier.PROTECTED:
            umlClass.protectedMethods.append((returnType, cursor.spelling, argumentTypes))

def processClass(cursor, inclusionConfig):
    """ Processes an ast node that is a class. """
    umlClass = UmlClass()  # umlClass is the datastructure for the DotGenerator
                           # that stores the necessary information about a single class.
                           # We extract this information from the clang ast hereafter ...
    if cursor.kind == clang.cindex.CursorKind.CLASS_TEMPLATE:
        # process declarations like:
        #   template <typename T> class MyClass
        umlClass.fqn = cursor.spelling
    else:
        # process declarations like:
        #   class MyClass ...
        #   struct MyStruct ...
        umlClass.fqn = cursor.type.spelling  # the fully qualified name

import re
    if (inclusionConfig['excludeClasses'] and
            re.match(inclusionConfig['excludeClasses'], umlClass.fqn)):
        return

if (inclusionConfig['includeClasses'] and not
            re.match(inclusionConfig['includeClasses'], umlClass.fqn)):
        return

for c in cursor.get_children():
        # process member variables and methods declarations
        processClassMemberDeclaration(umlClass, c)

dotGenerator.addClass(umlClass)

def traverseAst(cursor, inclusionConfig):
    if (cursor.kind == clang.cindex.CursorKind.CLASS_DECL
            or cursor.kind == clang.cindex.CursorKind.STRUCT_DECL
            or cursor.kind == clang.cindex.CursorKind.CLASS_TEMPLATE):
        # if the current cursor is a class, class template or struct declaration,
        # we process it further ...
        processClass(cursor, inclusionConfig)
    for child_node in cursor.get_children():
        traverseAst(child_node, inclusionConfig)

def parseTranslationUnit(filePath, includeDirs, inclusionConfig):
    clangArgs = ['-x', 'c++'] + ['-I' + includeDir for includeDir in includeDirs]
    tu = index.parse(filePath, args=clangArgs, options=clang.cindex.TranslationUnit.PARSE_SKIP_FUNCTION_BODIES)
    for diagnostic in tu.diagnostics:
        logging.debug(diagnostic)
    logging.info('Translation unit:' + tu.spelling + "\n")
    traverseAst(tu.cursor, inclusionConfig)

if __name__ == "__main__":
    parser = argparse.ArgumentParser(description="CodeDependencyVisualizer (CDV)")
    parser.add_argument('-d', required=True, help="directory with source files to parse (searches recusively)")
    parser.add_argument('-o', '--outFile', default='uml.dot', help="output file name / name of generated dot file")
    parser.add_argument('-u', '--withUnusedHeaders', help="parse unused header files (slow)")
    parser.add_argument('-a', '--associations', action="store_true", help="draw class member assiciations")
    parser.add_argument('-i', '--inheritances', action="store_true", help="draw class inheritances")
    parser.add_argument('-p', '--privMembers', action="store_true", help="show private members")
    parser.add_argument('-t', '--protMembers', action="store_true", help="show protected members")
    parser.add_argument('-P', '--pubMembers', action="store_true", help="show public members")
    parser.add_argument('-I', '--includeDirs', help="additional search path(s) for include files (seperated by space)", nargs='+')
    parser.add_argument('-v', '--verbose', action="store_true", help="print verbose information for debugging purposes")
    parser.add_argument('--excludeClasses', help="classes matching this pattern will be excluded")
    parser.add_argument('--includeClasses', help="only classes matching this pattern will be included")

args = vars(parser.parse_args(sys.argv[1:]))

filesToParsePatterns = ['*.cpp', '*.cxx', '*.c', '*.cc']
    if args['withUnusedHeaders']:
        filesToParsePatterns += ['*.h', '*.hxx', '*.hpp']
    filesToParse = findFilesInDir(args['d'], filesToParsePatterns)
    subdirectories = [x[0] for x in os.walk(args['d'])]

loggingFormat = "%(levelname)s - %(module)s: %(message)s"
    logging.basicConfig(format=loggingFormat, level=logging.INFO)
    if args['verbose']:
        logging.basicConfig(format=loggingFormat, level=logging.DEBUG)

logging.info("found " + str(len(filesToParse)) + " source files.")

for sourceFile in filesToParse:
        logging.info("parsing file " + sourceFile)
        parseTranslationUnit(sourceFile, args['includeDirs'], {
            'excludeClasses': args['excludeClasses'],
            'includeClasses': args['includeClasses']})

dotGenerator.setDrawAssociations(args['associations'])
    dotGenerator.setDrawInheritances(args['inheritances'])
    dotGenerator.setShowPrivMethods(args['privMembers'])
    dotGenerator.setShowProtMethods(args['protMembers'])
    dotGenerator.setShowPubMethods(args['pubMembers'])

dotfileName = args['outFile']
    logging.info("generating dotfile " + dotfileName)
    with open(dotfileName, 'w') as dotfile:
        dotfile.write(dotGenerator.generate())

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

#!/usr/bin/env python

import clang.cindex

import sys

import os

import logging

import argparse

import fnmatch

from DotGenerator import *

index = clang.cindex.Index.create()

dotGenerator = DotGenerator()

def findFilesInDir(rootDir, patterns):

""" Searches for files in rootDir which file names mathes the given pattern. Returns

a list of file paths of found files"""

foundFiles = []

for root, dirs, files in os.walk(rootDir):

for p in patterns:

for filename in fnmatch.filter(files, p):

foundFiles.append(os.path.join(root, filename))

return foundFiles

def processClassField(cursor):

""" Returns the name and the type of the given class field.

The cursor must be of kind CursorKind.FIELD_DECL"""

type = None

fieldChilds = list(cursor.get_children())

if len(fieldChilds) == 0: # if there are not cursorchildren, the type is some primitive datatype

type = cursor.type.spelling

else: # if there are cursorchildren, the type is some non-primitive datatype (a class or class template)

for cc in fieldChilds:

if cc.kind == clang.cindex.CursorKind.TEMPLATE_REF:

type = cc.spelling

elif cc.kind == clang.cindex.CursorKind.TYPE_REF:

type = cursor.type.spelling

name = cursor.spelling

return name, type

def processClassMemberDeclaration(umlClass, cursor):

""" Processes a cursor corresponding to a class member declaration and

appends the extracted information to the given umlClass """

if cursor.kind == clang.cindex.CursorKind.CXX_BASE_SPECIFIER:

for baseClass in cursor.get_children():

if baseClass.kind == clang.cindex.CursorKind.TEMPLATE_REF:

umlClass.parents.append(baseClass.spelling)

elif baseClass.kind == clang.cindex.CursorKind.TYPE_REF:

umlClass.parents.append(baseClass.type.spelling)

elif cursor.kind == clang.cindex.CursorKind.FIELD_DECL: # non static data member

name, type = processClassField(cursor)

if name is not None and type is not None:

# clang < 3.5: needs patched cindex.py to have

# clang.cindex.AccessSpecifier available:

# https://gitorious.org/clang-mirror/clang-mirror/commit/e3d4e7c9a45ed9ad4645e4dc9f4d3b4109389cb7

if cursor.access_specifier == clang.cindex.AccessSpecifier.PUBLIC:

umlClass.publicFields.append((name, type))

elif cursor.access_specifier == clang.cindex.AccessSpecifier.PRIVATE:

umlClass.privateFields.append((name, type))

elif cursor.access_specifier == clang.cindex.AccessSpecifier.PROTECTED:

umlClass.protectedFields.append((name, type))

elif cursor.kind == clang.cindex.CursorKind.CXX_METHOD:

try:

returnType, argumentTypes = cursor.type.spelling.split(' ', 1)

if cursor.access_specifier == clang.cindex.AccessSpecifier.PUBLIC:

umlClass.publicMethods.append((returnType, cursor.spelling, argumentTypes))

elif cursor.access_specifier == clang.cindex.AccessSpecifier.PRIVATE:

umlClass.privateMethods.append((returnType, cursor.spelling, argumentTypes))

elif cursor.access_specifier == clang.cindex.AccessSpecifier.PROTECTED:

umlClass.protectedMethods.append((returnType, cursor.spelling, argumentTypes))

except:

logging.error("Invalid CXX_METHOD declaration! " + str(cursor.type.spelling))

elif cursor.kind == clang.cindex.CursorKind.FUNCTION_TEMPLATE:

returnType, argumentTypes = cursor.type.spelling.split(' ', 1)

if cursor.access_specifier == clang.cindex.AccessSpecifier.PUBLIC:

umlClass.publicMethods.append((returnType, cursor.spelling, argumentTypes))

elif cursor.access_specifier == clang.cindex.AccessSpecifier.PRIVATE:

umlClass.privateMethods.append((returnType, cursor.spelling, argumentTypes))

elif cursor.access_specifier == clang.cindex.AccessSpecifier.PROTECTED:

umlClass.protectedMethods.append((returnType, cursor.spelling, argumentTypes))

def processClass(cursor, inclusionConfig):

""" Processes an ast node that is a class. """

umlClass = UmlClass() # umlClass is the datastructure for the DotGenerator

# that stores the necessary information about a single class.

# We extract this information from the clang ast hereafter ...

if cursor.kind == clang.cindex.CursorKind.CLASS_TEMPLATE:

# process declarations like:

# template <typename T> class MyClass

umlClass.fqn = cursor.spelling

else:

# process declarations like:

# class MyClass ...

# struct MyStruct ...

umlClass.fqn = cursor.type.spelling # the fully qualified name

import re

if (inclusionConfig['excludeClasses'] and

re.match(inclusionConfig['excludeClasses'], umlClass.fqn)):

return

if (inclusionConfig['includeClasses'] and not

re.match(inclusionConfig['includeClasses'], umlClass.fqn)):

return

for c in cursor.get_children():

# process member variables and methods declarations

processClassMemberDeclaration(umlClass, c)

dotGenerator.addClass(umlClass)

def traverseAst(cursor, inclusionConfig):

if (cursor.kind == clang.cindex.CursorKind.CLASS_DECL

or cursor.kind == clang.cindex.CursorKind.STRUCT_DECL

or cursor.kind == clang.cindex.CursorKind.CLASS_TEMPLATE):

# if the current cursor is a class, class template or struct declaration,

# we process it further ...

processClass(cursor, inclusionConfig)

for child_node in cursor.get_children():

traverseAst(child_node, inclusionConfig)

def parseTranslationUnit(filePath, includeDirs, inclusionConfig):

clangArgs = ['-x', 'c++'] + ['-I' + includeDir for includeDir in includeDirs]

tu = index.parse(filePath, args=clangArgs, options=clang.cindex.TranslationUnit.PARSE_SKIP_FUNCTION_BODIES)

for diagnostic in tu.diagnostics:

logging.debug(diagnostic)

logging.info('Translation unit:' + tu.spelling + "\n")

traverseAst(tu.cursor, inclusionConfig)

if __name__ == "__main__":

parser = argparse.ArgumentParser(description="CodeDependencyVisualizer (CDV)")

parser.add_argument('-d', required=True, help="directory with source files to parse (searches recusively)")

parser.add_argument('-o', '--outFile', default='uml.dot', help="output file name / name of generated dot file")

parser.add_argument('-u', '--withUnusedHeaders', help="parse unused header files (slow)")

parser.add_argument('-a', '--associations', action="store_true", help="draw class member assiciations")

parser.add_argument('-i', '--inheritances', action="store_true", help="draw class inheritances")

parser.add_argument('-p', '--privMembers', action="store_true", help="show private members")

parser.add_argument('-t', '--protMembers', action="store_true", help="show protected members")

parser.add_argument('-P', '--pubMembers', action="store_true", help="show public members")

parser.add_argument('-I', '--includeDirs', help="additional search path(s) for include files (seperated by space)", nargs='+')

parser.add_argument('-v', '--verbose', action="store_true", help="print verbose information for debugging purposes")

parser.add_argument('--excludeClasses', help="classes matching this pattern will be excluded")

parser.add_argument('--includeClasses', help="only classes matching this pattern will be included")

args = vars(parser.parse_args(sys.argv[1:]))

filesToParsePatterns = ['*.cpp', '*.cxx', '*.c', '*.cc']

if args['withUnusedHeaders']:

filesToParsePatterns += ['*.h', '*.hxx', '*.hpp']

filesToParse = findFilesInDir(args['d'], filesToParsePatterns)

subdirectories = [x[0] for x in os.walk(args['d'])]

loggingFormat = "%(levelname)s - %(module)s: %(message)s"

logging.basicConfig(format=loggingFormat, level=logging.INFO)

if args['verbose']:

logging.basicConfig(format=loggingFormat, level=logging.DEBUG)

logging.info("found " + str(len(filesToParse)) + " source files.")

for sourceFile in filesToParse:

logging.info("parsing file " + sourceFile)

parseTranslationUnit(sourceFile, args['includeDirs'], {

'excludeClasses': args['excludeClasses'],

'includeClasses': args['includeClasses']})

dotGenerator.setDrawAssociations(args['associations'])

dotGenerator.setDrawInheritances(args['inheritances'])

dotGenerator.setShowPrivMethods(args['privMembers'])

dotGenerator.setShowProtMethods(args['protMembers'])

dotGenerator.setShowPubMethods(args['pubMembers'])

dotfileName = args['outFile']

logging.info("generating dotfile " + dotfileName)

with open(dotfileName, 'w') as dotfile:

dotfile.write(dotGenerator.generate())

参考链接

Android Studio 3.2.1上vuh库使用的例子

Android Studio 3.2.1 上 vuh 库使用的例子，首先使用基于Vulkan的GPGPU计算框架Vuh 编译出 Android 版本的动态库，然后依照如下步骤建立工程。

继续阅读Android Studio 3.2.1上vuh库使用的例子

Python - os.walk忽略隐藏文件

Python中进行文件夹遍历的时候，使用os.walk是非常方便的，但是os.walk会把隐藏文件一起遍历，我们有时候需要忽略隐藏文件，可以用如下方法忽略隐藏文件以及文件夹

import os

path = '.'

for root, dirs, files in os.walk(path):

files = [f for f in files if not f[0] == '.']

dirs[:] = [d for d in dirs if not d[0] == '.']

# use files and dirs

for file_name in files:

print(os.path.join(root, file_name))

参考链接

python – os.walk没有隐藏的文件夹

macOS Mojave(10.14.2)配置OpenVPN软件Tunnelblick 3.7.8(客户端）

我们需要在macOS Mojave (10.14.2)配置连接在OpenVPN服务器实现加密上网或者翻墙。这里我们使用在Tunnelblick 3.7.8实现我们需要的功能。
继续阅读

使用ccache加速编译Android NDK项目

如果使用cmake构建项目,配置如下：

# macOS `brew install ccache` ubuntu `apt-get install ccache`

$ export ANDROID_NDK_HOME=~/Library/Android/sdk/ndk-bundle

$ cmake .. \

-DNDK_CCACHE=ccache \

-DCMAKE_TOOLCHAIN_FILE=$ANDROID_NDK_HOME/build/cmake/android.toolchain.cmake \

-DANDROID_ABI="armeabi-v7a" \

-DCMAKE_BUILD_TYPE=Release \

-DANDROID_STL=c++_static \

-DANDROID_NATIVE_API_LEVEL=android-24 \

-DANDROID_TOOLCHAIN=clang

$ make -j4

也就是定义NDK_CCACHE=ccache。

如果使用ndk-build构建项目，配置如下：

# macOS `brew install ccache` ubuntu `apt-get install ccache`

$ export NDK_CCACHE=ccache

$ export USE_CCACHE=1

$ ndk-build

也就是增加两个环境变量。

不指定缓存目录的情况下，缓存文件的目录一般在当前用户名下的.ccache目录下,时间长了，可能会产生很多的缓存文件，需要定时清理，当然也可以限制缓存的最大大小，让ccache根据需要进行淘汰。

参考链接

macOS Mojave(10.14.2)系统上编译LunarG/VulkanTools工程的Android版本

$ cd ~

$ git clone https://github.com/LunarG/VulkanTools.git

$ cd VulkanTools

$ cd build-android

$ export ANDROID_SDK_HOME=~/Library/Android/sdk

$ export ANDROID_NDK_HOME=~/Library/Android/sdk/ndk-bundle

# 当前的版本需要Android SDK 23.0.0,但是需要Java 11以下版本的Java才能执行SDK里面的sdkmanager

# 显示可以安装的Java版本号信息，需要检出这个内容
$ brew tap caskroom/versions

$ brew search "java*"
==> Formulae
app-engine-java            javarepl                   libreadline-java
google-java-format         jslint4java

==> Casks
charles-applejava          java-beta                  netbeans-java-se
eclipse-java               java6                      oracle-jdk-javadoc
eclipse-javascript         java8                      yourkit-java-profiler
java                       netbeans-java-ee
==> Did you mean to perform a regular expression search?
Surround your query wit

# 可以看到上面存在Java8可以安装，我们安装Java8
$ brew cask install java8

# 指定默认的Java版本
$ export JAVA_HOME=`/usr/libexec/java_home -v 1.8`

$ $ANDROID_SDK_HOME/tools/bin/sdkmanager --update

# 目前的代码需要android-23的SDK
$ $ANDROID_SDK_HOME/tools/bin/sdkmanager "platforms;android-23"

$ $ANDROID_SDK_HOME/tools/bin/sdkmanager "platforms;android-24"

$ $ANDROID_SDK_HOME/tools/bin/sdkmanager "build-tools;24.0.3"

# 可选安装
# $ANDROID_SDK_HOME/tools/bin/sdkmanager "platform-tools" "platforms;24.0.3"

# 需要Android 7.0以上才支持Vulkan，因此，我们需要更高版本的编译工具
$ export PATH=$ANDROID_NDK_HOME:$ANDROID_SDK_HOME/build-tools/24.0.3/:$PATH

# 启用 ccache 减少二次编译的时间 
$ brew install ccache

$ export NDK_CCACHE=ccache
 
$ export USE_CCACHE=1

$ export VULKAN_HEADERS_INSTALL_DIR=./third_party/Vulkan-Headers

$ bash build_all.sh

# 编译完成后，执行 bash test_APK.sh 测试功能，不过话说，这个工具是做什么的，有点看不懂了
# 难不成是厂家测试驱动用的？
$ bash test_APK.sh

$ cd ~

$ git clone https://github.com/LunarG/VulkanTools.git

$ cd VulkanTools

$ cd build-android

$ export ANDROID_SDK_HOME=~/Library/Android/sdk

$ export ANDROID_NDK_HOME=~/Library/Android/sdk/ndk-bundle

# 当前的版本需要Android SDK 23.0.0,但是需要Java 11以下版本的Java才能执行SDK里面的sdkmanager

# 显示可以安装的Java版本号信息，需要检出这个内容

$ brew tap caskroom/versions

$ brew search "java*"

==> Formulae

app-engine-java javarepl libreadline-java

google-java-format jslint4java

==> Casks

charles-applejava java-beta netbeans-java-se

eclipse-java java6 oracle-jdk-javadoc

eclipse-javascript java8 yourkit-java-profiler

java netbeans-java-ee

==> Did you mean to perform a regular expression search?

Surround your query wit

# 可以看到上面存在Java8可以安装，我们安装Java8

$ brew cask install java8

# 指定默认的Java版本

$ export JAVA_HOME=`/usr/libexec/java_home -v 1.8`

$ $ANDROID_SDK_HOME/tools/bin/sdkmanager --update

# 目前的代码需要android-23的SDK

$ $ANDROID_SDK_HOME/tools/bin/sdkmanager "platforms;android-23"

$ $ANDROID_SDK_HOME/tools/bin/sdkmanager "platforms;android-24"

$ $ANDROID_SDK_HOME/tools/bin/sdkmanager "build-tools;24.0.3"

# 可选安装

# $ANDROID_SDK_HOME/tools/bin/sdkmanager "platform-tools" "platforms;24.0.3"

# 需要Android 7.0以上才支持Vulkan，因此，我们需要更高版本的编译工具

$ export PATH=$ANDROID_NDK_HOME:$ANDROID_SDK_HOME/build-tools/24.0.3/:$PATH

# 启用 ccache 减少二次编译的时间

$ brew install ccache

$ export NDK_CCACHE=ccache

$ export USE_CCACHE=1

$ export VULKAN_HEADERS_INSTALL_DIR=./third_party/Vulkan-Headers

$ bash build_all.sh

# 编译完成后，执行 bash test_APK.sh 测试功能，不过话说，这个工具是做什么的，有点看不懂了

# 难不成是厂家测试驱动用的？

$ bash test_APK.sh

参考链接

OpenCL代码编译成Vulkan代码的工具clspv

最近接到个任务，就是把OpenCL的.cl代码编译成Vulkan程序。

使用Google开源的工具clspv实现这个功能即可。

$ git clone https://github.com/google/clspv.git

$ cd clspv

$ python utils/fetch_sources.py

$ mkdir build

$ cd build

$ cmake ..

$ make -j8

编译完成后，在 bin 目录下生成 clspv 这个可执行程序。

使用的时候参考项目的文档 OpenCL C 1.2 Language on Vulkan 。

只是需要注意的是，如果使用了 image 类型，比如image2d_t 则需要提供 -samplermap 这个参数，这个参数指向一个文件，文件内容可能如下：

1 2	CLK_NORMALIZED_COORDS_FALSE \| CLK_ADDRESS_CLAMP_TO_EDGE \| CLK_FILTER_NEAREST, CLK_NORMALIZED_COORDS_TRUE \| CLK_ADDRESS_CLAMP_TO_EDGE \| CLK_FILTER_LINEAR

具体的定义跟在 OpenCL 中调用 read_image/write_image 函数时候指定的 sampler 参数一致即可。

参考链接

OpenCL C 1.2 Language on Vulkan

2025 年 6 月
一	二	三	四	五	六	日
						1
2	3	4	5	6	7	8
9	10	11	12	13	14	15
16	17	18	19	20	21	22
23	24	25	26	27	28	29
30