How to package a QT program

Under Windows system

Through the windeployqt tool

1. Run the program in release mode

insert image description here

2. Copy the executable file in the release folder to a path, desktop is recommended

3. Create a new folder on the desktop and put the executable file just now into it

4. Run the windeployqt.exe program

insert image description here

5. Go to the newly created folder path

6. Execute the command windeployqt test.exe, where test is the file name of the executable file

insert image description here

7. After the execution is completed, you can find that there are more dlls required to run the program in the folder

insert image description here

8. The folder is packaged and can be run on other computers

Under Linux system

Through the linuxdeployqt tool

1. Run in release mode

Run the QT program in release mode like Windows to generate an executable file, that is, you can directly double-click the opened file, or you can view it through the right-click property

shared library file

insert image description here

executable file

insert image description here

Note that the icon is also different. If it is an executable file, the next step is not required.

2. For the files compiled in release mode are shared library files, perform the following operations

Add the following code to the .pro file

TEMPLATE = app #这生成一个exe
QMAKE_LFLAGS += -no-pies

Remember to execute qmake after adding

insert image description here

3. Create a new folder and copy the generated executable file into it

insert image description here

The name of the folder here can be the same as the name of the executable file for easy identification. I will not demonstrate it here.

4. Install linuxdeployqt

For Qt, the tool linuxdeployqt can help us quickly package Qt projects. The corresponding software under windows is called windowsdeployqt.

The github URL of linuxdeployqt: linuxdeployoqt

insert image description here

There are two ways to install linuxdeployqt on github, which are directly downloading and installing the compiled installation package, or installing from source code. It's not difficult either way.
However, since linuxdeployqt has not released the corresponding version of Ubuntu18.04, it is better for Ubuntu18.04 users to compile the source code for installation, otherwise when using linuxdeployqt to package the Qt project, an error similar to the operating system version being too high will be reported.

4.1 Download the linuxdeployqt executable program, the first item in the above picture

It is recommended to modify the file name after downloading, and change the file name to linuxdeployqt.
To modify the permissions of linuxdeployqt, you can use the following command:

sudo chmod a+x linuxdeployqt

Then you can consider copying the file to the directory /usr/local/bin/, so that you can use the command linuxdeployqt directly later. The copy command is as follows:

sudo cp linuxdeployqt /usr/local/bin/

4.2 Compile the source code and install linuxdeployqt

The operating system is Ubuntu18.04 and above. At present, you need to compile and install linuxdeployqt through source code. Otherwise, some error messages will appear when executing linuxdeployqt: such as the host system is too high.

Reference URL for compiling source code installation: Reference URL
We have already installed g++, git and other tools by default, if not, you must install them first.
You also need to install the patchelf tool.

An example command is as follows:

sudo apt-get -y install git g++ libgl1-mesa-dev
git clone https://github.com/probonopd/linuxdeployqt.git
sudo apt install patchelf

Execute the above commands in turn, you can

Enter the linuxdeployqt folder, the command is as follows:

cd linuxdeployqt

Modify the source code of tools/linuxdeployqt/main.cpp, comment out the version judgment statement in the source code, and modify the source code by command gedit tools/linuxdeployqt/main.cpp:

gedit tools/linuxdeployqt/main.cpp

The code that needs to be commented is as follows:

// openSUSE Leap 15.0 uses glibc 2.26 and is used on OBS
        /*if (strverscmp (glcv, "2.27") >= 0) {
            qInfo() << "ERROR: The host system is too new.";
            qInfo() << "Please run on a system with a glibc version no newer than what comes with the oldest";
            qInfo() << "currently still-supported mainstream distribution (xenial), which is glibc 2.23.";
            qInfo() << "This is so that the resulting bundle will work on most still-supported Linux distributions.";
            qInfo() << "For more information, please see";
            qInfo() << "https://github.com/probonopd/linuxdeployqt/issues/340";
            return 1;
        }*/

There is a CMAKELIST file in the current directory, execute the cmake command:

cmake CMakeLists.txt

Execute the make command to compile the source code:

make

After the compilation is complete, an executable file linuxdeployqt will be generated under the directory tools/linuxdeployqt/linuxdeployqt, modify its permissions:

cd ./tools/linuxdeployqt
sudo chmod a+x linuxdeployqt

For later use, you can copy the executable file to the /usr/local/bin directory:

sudo cp linuxdeployqt /usr/local/bin/

The linuxdeployqt command may use the option -appImage (actually I used it), so you need to install appimagetool. The download and installation commands are as follows:

sudo wget -c "https://github.com/AppImage/AppImageKit/releases/download/continuous/appimagetool-x86_64.AppImage" -O /usr/local/bin/appimagetool
sudo chmod a+x /usr/local/bin/appimagetool

5. Add environment variables

At this time, if you execute qmake -vthe command operation, it means that the path is not found, and you need to add environment variables.

Enter gedit ~/.bashrcthe command in the terminal, modify the .bashrc file, and append the following content at the end of the file, where /opt/Qt5.9.9/5.9.9 is my Qt installation path, and you should use your own path instead:

#add QT ENV
export PATH=/opt/Qt5.9.9/5.9.9/gcc_64/bin:$PATH
export LD_LIBRARY_PATH=/opt/Qt5.9.9/5.9.9/gcc_64/lib:$LD_LIBRARY_PATH
export QT_PLUGIN_PATH=/opt/Qt5.9.9/5.9.9/gcc_64/plugins:$QT_PLUGIN_PATH
export QML2_IMPORT_PATH=/opt/Qt5.9.9/5.9.9/gcc_64/qml:$QML2_IMPORT_PATH

It is also possible to add only the first line of commands after testing

Double quotes can be added or not

Finally, source the ~/.bashrc shell file to take effect immediately without having to log out and log in again.

$ source ~/.bashrc

Then execute qmake -v

insert image description here

6. How to use

Order

linuxdeployqt ~/path/to/project/application -appimage

Example: Go to the newly created folder previously used to store executable files

cd /home/li/Desktop.qttest
linuxdeployqt untitled -appimage

Among them, untitled is an executable file, and the program can be automatically packaged after execution.

insert image description here

AppRun is equivalent to a shortcut.

When executing this command, the following error may be reported:

Categories entry not found in desktop file
.desktop file is missing a Categories= key
12

It is said that one attribute, Categories, is missing in the .desktop file, so we need to manually add a line to the desktop file:

Categories=Application;

7. Add qt application icon in ubuntu

Modify the desktop file in the qt directory. You can modify it according to the official ubuntu prompt.
ubuntu desktop file use

#-- 全局安装（所有用户可用),将xxx.desktop 复制到/usr/share/applications  
#-- 当前用户可用, 将xxx.desktop 复制到 ~/.local/share/applications 目录即可
#--appName.desktop
[Desktop Entry]
Version=1.0 #app的版本
Name=myQtApp #app的名字
Comment= this app use for xxx #说明信息 
Exec=/path/to/your/QtApp/myQtApp #app的执行路径，绝对路径
Icon=/path/to/your/app_icon/myQtApp.png #icon 路径，绝对路径
Terminal=false #是否在终端启动，效果自己试一下就知道了
Type=Application
Categories=Utility;Application;

screenplay

1. Create a new file

Create two new files in the newly created folder

pack.sh
untitled.sh The file name of this file must be the same as the file name of the executable file

2. Add script

Add the following to the pack.sh file:

#!/bin/sh  
exe="untitled" #你需要发布的程序名称
des="/home/li/Desktop/qttest2" #创建文件夹的位置
deplist=$(ldd $exe | awk  '{
     
     if (match($3,"/")){
    
     printf("%s "),$3 } }')  
cp $deplist $des

Then add the following content to the untitled.sh file (the content does not need to be changed):

#!/bin/sh  
appname=`basename $0 | sed s,\.sh$,,`  
dirname=`dirname $0`  
tmp="${dirname#?}"  
if [ "${dirname%$tmp}" != "/" ]; then  
dirname=$PWD/$dirname  
fi  
LD_LIBRARY_PATH=$dirname  
export LD_LIBRARY_PATH  
$dirname/$appname "$@"

3. Execute the script

Enter the folder directory through the terminal and run:

sudo ./pack.sh

There will be an error message that the command is not found, you can execute the following command, and use the file as an executable file

chmod a+x pack.sh

Then execute sudo ./pack.shthe command again.

insert image description here

Then the dependent link library of the executable file will be copied to the directory where it is located.

insert image description here

at last

But one thing to note, it is better to use non-Linux FHS (File System Hierarchy Standard)

When creating a folder, the structure is as follows

.
├── build
├── CI
│   ├── build.sh
│   ├── exe
├── include
│   ├── bchartdir.h
├── lib
│   ├── libchartdir.so
│   ├── libchartdir.so.6.0
│   └── libchartdir.so.6.0.0
├── src
│   ├── ubuntuapp.pro
│   ├── main.cpp
│   ├── mainwindow.cpp
│   ├── mainwindow.h
└── tool
    └── deployqt

The content in build.sh is as follows

# /bin/bash
cd /home/wu/workspace/ubuntuapp/CI/exe
cp /home/wu/workspace/ubuntuapp/tool/deployqt ./
cd /home/wu/workspace/ubuntuapp/src
/home/wu/Qt5.10.1/5.10.1/gcc_64/bin/qmake /home/wu/workspace/ubuntuapp/src/ubuntuapp.pro -o /home/wu/workspace/ubuntuapp/build -spec linux-g++ CONFIG+=debug
make clean -C /home/wu/workspace/ubuntuapp/build
make -C /home/wu/workspace/ubuntuapp/build
cd /home/wu/workspace/ubuntuapp/CI/exe
cp /home/wu/workspace/ubuntuapp/build/UbuntuApp ./
./deployqt UbuntuApp -qmake="/home/wu/Qt5.10.1/5.10.1/gcc_64/bin/qmake"
rm deployqt

After writing, add permissions

sudo chmod 777 build.sh

The content in ubuntuapp.pro is as follows

QT += core gui
TARGET = UbuntuApp
TEMPLATE = app
CONFIG += C++11
DEFINES += QT_DEPRECATED_WARNINGS
SOURCES += main.cpp \
    mainwindow.cpp
HEADERS += mainwindow.h
INCLUDEPATH += ../include
LIBS += -L../lib -lchartdir
unix:MOC_DIR = ../build/moc
unix:OBJECTS_DIR = ../build/obj

In this way, after executing build.sh , you can get some files and folders in the exe folder

The hyperlink of the AppRun inside can be double-clicked to open the UbuntuApp program

Then, change the name of the exe folder, compress it into a zip format, and copy it to other machines for use

1. About qt.conf

This file specifies the running environment of the qt program, that is, we can use qt.conf to specify the running path and library path of the qt program. The content is as follows:

# Generated by linuxdeployqt
# https://github.com/probonopd/linuxdeployqt/
[Paths]
Prefix = ./		#程序的运行路径
Libraries =  ./lib  #程序的库路径
Plugins = plugins	#插件路径
Imports = qml
Qml2Imports = qml
ld.sh/)** 后，就能在 exe 文件夹里得到 一些文件和文件夹

里面的 AppRun 的超链接可以双击后 打开 UbuntuApp 程序

然后，把 exe 文件夹改个名字，压缩成 zip 格式，就可以拷贝到其他机器上用了

##### **1.关于qt.conf**

这个文件指定了qt 程序的运行环境，即我们可以使用qt.conf 指定qt程序的运行路径和库路径。内容如下所示：

```shell
# Generated by linuxdeployqt
# https://github.com/probonopd/linuxdeployqt/
[Paths]
Prefix = ./		#程序的运行路径
Libraries =  ./lib  #程序的库路径
Plugins = plugins	#插件路径
Imports = qml
Qml2Imports = qml