Skip to content

Latest commit

 

History

History
167 lines (127 loc) · 5.85 KB

HACKING.md

File metadata and controls

167 lines (127 loc) · 5.85 KB
Raw

Build from repository

$ ./bootstrap
$ ./configure

The bootstrap script will use autotools to set up the build environment and create the configure script.

Use ./configure --help for options. Use --prefix to make an install in your home directory. This is necessary to test python scripts. The systemd user unit directory should be set to avoid writing to the system location.

Systemd will look for the unit files in ~/.config/systemd/user so this directory can be used as a target if the unit files will be used. Otherwise the location can be set to no to disable the systemd files.

Example:

$ ./configure --prefix=$HOME/redshift/root \
   --with-systemduserunitdir=$HOME/.config/systemd/user

Now, build the files:

$ make

The main redshift program can be run at this point. To install to the prefix directory run:

$ make install

You can now run the python script. Example:

$ $HOME/redshift/root/bin/redshift-gtk

Dependencies

  • autotools, gettext

  • libdrm (Optional, for DRM support)

  • libxcb, libxcb-randr (Optional, for RandR support)

  • libX11, libXxf86vm (Optional, for VidMode support)

  • geoclue (Optional, for geoclue support)

  • python3, pygobject, pyxdg (Optional, for GUI support)

  • appindicator (Optional, for Ubuntu-style GUI status icon)

Coding style

Redshift follows roughly the Linux coding style http://www.kernel.org/doc/Documentation/CodingStyle. Some specific rules to note are:

  • Lines should not be much longer than 80 chars but this is not strictly enforced. If lines are much longer than this the code could likely be improved by moving some parts to a smaller function.
  • All structures are typedef'ed.
  • Avoid Yoda conditions; they make the logic unnecessarily hard to comprehend.
  • Avoid multiline if-statements without braces; either use a single line or add the braces.
  • Use only C-style comments (/* */).

Creating a pull request

  1. Create a topic branch for your specific changes. You can base this off the master branch or a specific version tag if you prefer (git co -b topic master).
  2. Create a commit for each logical change on the topic branch. The commit log must contain a one line description (max 80 chars). If you cannot describe the commit in 80 characters you should probably split it up into multiple commits. The first line can be followed by a blank line and a longer description (split lines at 80 chars) for more complex commits. If the commit fixes a known issue, mention the issue number in the first line (Fix #11: ...).
  3. The topic branch itself should tackle one problem. Feel free to create many topic branches and pull requests if you have many different patches. Putting them into one branch makes it harder to review the code.
  4. Push the topic branch to Github, find it on github.com and create a pull request to the master branch. If you are making a bug fix for a specific release you can create a pull request to the release branch instead (e.g. release-1.9).
  5. Discussion will ensue. If you are not prepared to partake in the discussion or further improve your patch for inclusion, please say so and someone else may be able to take on responsibility for your patch. Otherwise we will assume that you will be open to critisism and suggestions for improvements and that you will take responsibility for further improving the patch. You can add further commits to your topic branch and they will automatically be added to the pull request when you push them to Github.
  6. You may be asked to rebase the patch on the master branch if your patch conflicts with recent changes to the master branch. However, if there is no conflict, there is no reason to rebase. Please do not merge the master back into your topic branch as that will convolute the history unnecessarily.
  7. Finally, when your patch has been refined, you may be asked to squash small commits into larger commits. This is simply so that the project history is clean and easy to follow. Remember that each commit should be able to stand on its own, be able to compile and function normally. Commits that fix a small error or adds a few comments to a previous commit should normally just be squashed into that larger commit.

If you want to learn more about the Git branching model that we use please see http://nvie.com/posts/a-successful-git-branching-model/ but note that we use the master branch as develop.

Creating a new release

  1. Select a commit in master to branch from, or if making a bugfix release use previous release tag as base (e.g. for 1.9.1 use 1.9 as base)
  2. Create release branch release-X.Y
  3. Apply any bugfixes for release
  4. Import updated translations from launchpad and commit. Remember to update po/LINGUAS if new languages were added
  5. Update version in configure.ac and create entry in NEWS
  6. Run make distcheck
  7. Commit and tag release (vX.Y or vX.Y.Z)
  8. Push tag to Github and also upload source dist file to Github

Also remember to check before release that

  • Windows build is ok
  • Build files for distributions are updated

Build Fedora RPMs

Run make dist-xz and copy the .tar.xz file to ~/rpmbuild/SOURCES. Then run

$ rpmbuild -ba contrib/redshift.spec

If successful this will place RPMs in ~/rpmbuild/RPMS.

Cross-compile for Windows

Install MinGW and run configure using the following command line. Use i686-w64-migw32 as host for 32-bit builds.

$ ./configure --disable-drm --disable-randr --disable-vidmode --enable-wingdi \
  --disable-geoclue --disable-gui --disable-ubuntu \
  --host=x86_64-w64-mingw32

Notes

  • verbose flag is (currently) only held in redshift.c; thus, write all verbose messages there.