Skip to content

Latest commit

 

History

History
242 lines (171 loc) · 13.4 KB

README.md

File metadata and controls

242 lines (171 loc) · 13.4 KB
Raw

linuxdeployqt Build Status Codacy Badge discourse Gitter irc

This Linux Deployment Tool for Qt, linuxdeployqt, takes an application as input and makes it self-contained by copying in the Qt libraries and plugins that the application uses into a bundle. This can optionally be put into an AppImage, and, using fpm, into cross-distro deb and rpm packages.

Differences to macdeployqt

This tool is conceptually based on the Mac Deployment Tool, macdeployqt in the tools applications of the Qt Toolkit, but has been changed to a slightly different logic and other tools needed for Linux.

  • Instead of an .app bundle for macOS, this produces an AppDir for Linux
  • Instead of a .dmg disk image for macOS, this produces an AppImage for Linux which is quite similar to a dmg but executes the contained application rather than just opening a window on the desktop from where the application can be launched

Known issues

This may not be fully working yet. See GitHub Issues for known issues. Use with care, run with maximum verbosity, submit issues and pull requests. Help is appreciated.

Installation

Please download linuxdeployqt-x86_64.AppImage from the Releases page and chmod a+x it. If you would like to build linuxdeployqt from source instead, see BUILDING.md.

Usage

Usage: linuxdeployqt app-binary [options]

Options:
   -verbose=<0-3>      : 0 = no output, 1 = error/warning (default), 2 = normal, 3 = debug
   -no-plugins         : Skip plugin deployment
   -appimage           : Create an AppImage
   -no-strip           : Don't run 'strip' on the binaries
   -bundle-non-qt-libs : Also bundle non-core, non-Qt libraries
   -executable=<path>  : Let the given executable use the deployed libraries too
   -qmldir=<path>      : Scan for QML imports to bundle from the given directory, determined by Qt's qmlimportscanner
   -always-overwrite   : Copy files even if the target file exists
   -qmake=<path>       : The qmake executable to use
   -no-translations    : Skip deployment of translations

linuxdeployqt takes an application as input and makes it
self-contained by copying in the Qt libraries and plugins that
the application uses.

Simplest example

Given that a desktop file should be provided with an AppImage, linuxdeployqt can use that to determine the parameters of the build.

linuxdeployqt ./path/to/appdir/usr/share/application_name.desktop

Where the desktop file specifies the executable to be run (with EXEC=), the name of the applications and an icon. See desktop file specification.

For a more detailed example, see "Using linuxdeployqt with Travis CI" below.

Checking library inclusion

Open in Qt Creator and build your application. Run it from the command line and inspect it with ldd to make sure the correct libraries from the correct locations are getting loaded, as linuxdeployqt will use ldd internally to determine from where to copy libraries into the bundle.

QMake configuration

Important: By default, linuxdeployqt deploys the Qt instance that qmake on the $PATH points to, so make sure that it is the correct one. Verify that qmake finds the correct Qt instance like this before running the linuxdeployqt tool:

qmake -v

QMake version 3.0
Using Qt version 5.7.0 in /tmp/.mount_QtCreator-5.7.0-x86_64/5.7/gcc_64/lib

If this does not show the correct path to your Qt instance that you want to be bundled, then adjust your $PATH to find the correct qmake.

Alternatively, use the -qmake command line option to point the tool directly to the qmake executable to be used.

Remove unecessary files

Before running linuxdeployqt it may be wise to delete unneeded files that you do not wish to distribute from the build directory. These may be autogenerated during the build. You can delete them like so:

find $HOME/build-*-*_Qt_* \( -name "moc_*" -or -name "*.o" -or -name "qrc_*" -or -name "Makefile*" -or -name "*.a" \) -exec rm {} \;

Alternatively, you could use $DESTDIR.

Using linuxdeployqt with Travis CI

A common use case for linuxdeployqt is to use it on Travis CI after the make command. The following example illustrates how to use linuxdeployqt with Travis CI. Create a .travis.yml file similar to this one (be sure to customize it, e.g., change APPNAME to the name of your application as it is spelled in the Name= entry of the .desktop file):

language: cpp
compiler: gcc
sudo: require
dist: trusty

before_install:
    - sudo add-apt-repository ppa:beineri/opt-qt59-trusty -y
    - sudo apt-get update -qq

install:
    - sudo apt-get -y install qt59base
    - source /opt/qt*/bin/qt*-env.sh

script:
  - qmake CONFIG+=release PREFIX=/usr
  - make -j$(nproc)
  - make INSTALL_ROOT=appdir -j$(nproc) install ; find appdir/
  - wget -c "https://github.com/probonopd/linuxdeployqt/releases/download/continuous/linuxdeployqt-continuous-x86_64.AppImage"
  - chmod a+x linuxdeployqt*.AppImage
  - unset QTDIR; unset QT_PLUGIN_PATH ; unset LD_LIBRARY_PATH
  - ./linuxdeployqt*.AppImage ./appdir/usr/share/applications/*.desktop -bundle-non-qt-libs
  - ./linuxdeployqt*.AppImage ./appdir/usr/share/applications/*.desktop -appimage

after_success:
  - find ./appdir -executable -type f -exec ldd {} \; | grep " => /usr" | cut -d " " -f 2-3 | sort | uniq
  - curl --upload-file ./APPNAME*.AppImage https://transfer.sh/APPNAME-git.$(git rev-parse --short HEAD)-x86_64.AppImage

When you save your change, then Travis CI should build and upload an AppImage for you. More likely than not, some fine-tuning will still be required.

For this to work, you need to enable Travis CI for your repository as described here prior to merging this, if you haven't already done so.

By default, qmake .pro files generated by Qt Creator unfortunately don't support make install out of the box. In this case you will get

make: Nothing to be done for `install'.
find: `appdir/': No such file or directory

If qmake does not allow for make install or does not install the desktop file and icon, then you need to change your .pro file it similar to https://github.com/probonopd/FeedTheMonkey/blob/master/FeedTheMonkey.pro.

  - make INSTALL_ROOT=appdir install ; find appdir/

CMake wants DESTDIR instead:

  - cmake . -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=/usr
  - make -j$(nproc)
  - make DESTDIR=appdir -j$(nproc) install ; find appdir/

autotools (the dinosaur that spends precious minutes "checking...") wants DESTDIR too but insists on an absolute link which we can feed it using readlink:

    - ./configure --prefix=/usr
    - make -j$(nproc)
    - make install DESTDIR=$(readlink -f appdir) ; find appdir/

Caution if you encounter

qmake PREFIX=/usr CONFIG+=use_qt_paths

Here, CONFIG+=use_qt_paths needs to be removed, otherwise it will install everything under the Qt installation paths in /opt/qt58 when using the beineri ppa.

The exception is that you are building Qt libraries that should be installed to the same location where Qt resides on your system, from where it will be picked up by linuxdeployqt.

Sending Pull Requests on GitHub

linuxdeployqt is great for upstream application projects that want to release their software in binary form to Linux users quickly and without much overhead. If you would like to see a particular application use linuxdeployqt, then sending a Pull Request may be an option to get the upstream application project to consider it. You can use the following template text for Pull Requests but make sure to customize it to the project in question.

This PR, when merged, will compile this application on [Travis CI](https://travis-ci.org/) upon each `git push`, and upload an [AppImage](http://appimage.org/) to a temporary download URL on transfer.sh (available for 14 days). The download URL is toward the end of each Travis CI build log of each build (see below for how to set up automatic uploading to your GitHub Releases page).

For this to work, you need to enable Travis CI for your repository as [described here](https://travis-ci.org/getting_started) __prior to merging this__, if you haven't already done so.

Providing an [AppImage](http://appimage.org/) would have, among others, these advantages:
- Applications packaged as an AppImage can run on many distributions (including Ubuntu, Fedora, openSUSE, CentOS, elementaryOS, Linux Mint, and others)
- One app = one file = super simple for users: just download one AppImage file, [make it executable](http://discourse.appimage.org/t/how-to-make-an-appimage-executable/80), and run
- No unpacking or installation necessary
- No root needed
- No system libraries changed
- Works out of the box, no installation of runtimes needed
- Optional desktop integration with `appimaged`
- Optional binary delta updates, e.g., for continuous builds (only download the binary diff) using AppImageUpdate
- Can optionally GPG2-sign your AppImages (inside the file)
- Works on Live ISOs
- Can use the same AppImages when dual-booting multiple distributions
- Can be listed in the [AppImageHub](https://appimage.github.io/) central directory of available AppImages

[Here is an overview](https://github.com/probonopd/AppImageKit/wiki/AppImages) of projects that are already distributing upstream-provided, official AppImages.

__Please note:__ Instead of storing AppImage builds temporarily for 14 days each on transfer.sh, you could use GitHub Releases to store the binaries permanently. This way, they would be visible on the Releases page of your project. This is what I recommend. See https://docs.travis-ci.com/user/deployment/releases/. If you want to do this for continuous builds, also see https://github.com/probonopd/uploadtool.

If you would like to see only one entry for the Pull Request in your project's history, then please enable [this GitHub functionality](https://help.github.com/articles/configuring-commit-squashing-for-pull-requests/) on your repo. It allows you to squash (combine) the commits when merging.

If you have questions, AppImage developers are on #AppImage on irc.freenode.net.

Projects using linuxdeployqt

These projects are already using Travis CI and linuxdeployqt to provide AppImages of their builds:

This project is already using linuxdeployqt in a custom Jenkins workflow:

These projects are already using linuxdeployqt:

This project on GitLab uses linuxdeployqt:

These can be bundled successfully using linuxdeployqt:

Contributing

One great way to contribute is to send Pull Requests to the application projects you'd like to see use linuxdeployqt, as described above. You are also welcome to contribute to linuxdeployqt development itself. Please discuss in the forum or using GitHub issues and Pull Requests.

Contact

The developers are in the channel #AppImage on irc.freenode.net