Adding new content to your hugo website hosted on gh-pages
June 3rd, 2017
Note 5.21.19: This website is no longer hosted on Github gh-pages, and no longer created with Hugo, but with another, and much better, in my opinion, static site generator. It’s called Gatsbyjs (version 2). However, the information related to gh-pages is still pertinent to those who want to follow the steps and use or continue to use Hugo. Lastly, this site is now HTTPS and no longer HTTP.
As some of you might already know, I just recently switched
WordPress to the Hugo
static website generator. So I’m still fairly
new to it. And some of you might already know what a stickler I am for good
So what does being a Hugo newbie have anything to do with developer docs?
Plenty! And here I will focus specifically on publishing new content to your
Hugo website hosted on the Github
gh-pages branch. Because I have been
doing a lot of that lately, and I want it etched in stone!
I’m assuming you have already created your new Hugo site locally. Perhaps
you have even pushed it to your remote master branch on Github
origin master). But now you want to host the site via the (free)
branch as opposed to paying oodles for hosting services which you might never
need or use.
Navigate to Hosting on Github Pages to read the official Hugo documentation, but refer to this article for explicit clarification.
First off, I chose the
deployment via gh-pages branch route. The setup and
functionality is analagous to the
gulp-gh-pages npm package setup and
functionality. What I love about
gulp-gh-pages is that once I have the
Gulp task set up, I can publish to
gh-pages from my master branch. I don’t
have to worry about committing changes before checking out to the
branch so as not to add unwanted content to it from the master branch, or
merging/rebasing branches to keep them in sync. It can get very messy.
gulp-gh-pages ensures that you keep branch histories separate. Deploying your
Hugo site via the Hugo
gh-pages branch method ensures that you keep branch
histories separate as well.
First make sure that you add your
public folder, the folder which is created
when Hugo generates your site for publication, is added to your
file. You can do this through Terminal with the command
echo "public" >> .gitignore. To double check that the command was successful,
open up your
.gitignore file to make sure the
public directory was added.
You are adding the
public directory to your
.gitignore file so that it’s
ignored on the master branch.
Next you want to initialize your
gh-pages branch as an
empty orphan branch.
Why? You don’t want to simply create a
gh-pages branch and then checkout out
to it, because that would mean 1. an invalid
gh-pages directory structure
and 2. unneeded files and directories in
gh-pages you would have to end up
deleting anyway. In addition, creating an
orphan branch with the
git checkout --orphan gh-pages command means that the first commit made on
this new branch will have no parents and will be the
root of a
DISCONNECTED from all other branches and commits. And that’s exactly
what we want, right?
All the commands/steps necessary for successfully initializing the
branch as an
empty orphan branch are the following:
git checkout --orphan gh-pages git reset --hard git commit --allow-empty -m "Initializing gh-pages branch" git push origin gh-pages git checkout master
git reset --hard deletes any files that might have crept into the
git commit --allow-empty permits you to initialize an empty branch. By
default, you cannot.
git push origin gh-pages pushes the local
gh-pages branch to the remote
git checkout master returns you to the
Now you want to checkout your
gh-pages branch into your
because it is the contents of the
public folder which will populate
and render the contents of your site there. You will do this using git’s
worktree feature. Why? Because the
worktree feature allows you to checkout
more than one branch at a time from your main working tree initialized by the
git init or
git clone. Usually that means the
worktree feature, however, initializes a
linked working tree using the
git worktree add command. A repository has
one main working tree (if it is
not a bare repository) and 0 or more
linked working trees. A linked working
tree permits you to push content from your main working tree to the linked
working tree, which is what is needed when
deploying your Hugo site to
gh-pages from your
master branch. The complete set of commands to check out
gh-pages branch into your
public folder using git’s
rm -rf public git worktree add -B gh-pages public origin/gh-pages
rm -rf public removes the
public folder so you can add your
linked gh-pages working tree with a clean slate.
git worktree add -B gh-pages public origin/gh-pages makes the
branch (-B) a
linked working tree (git worktree add).
public recreates the
public folder, and
origin/gh-pages links it to the remote
Since you now have an empty
public folder, you have to regenerate your
Hugo site with the
hugo command in Terminal. The complete set of
commands to regenerate your Hugo site and commit the generated files
gh-pages branch is the following:
hugo cd public git add --all git commit -m "Publishing to gh-pages" cd ..
hugo command regenerates your site by creating a new public folder with
all necessary files and directories within it.
cd public takes you into the
git add --all adds all the files and directories in
public to your staging
area so that they are ready to be committed.
git commit -m "Publishing to gh-pages" commits the files you added to the
staging area with the
git add --all command.
cd .. takes you back up one level into your
Go into your
public folder via your text editor to make sure that your site
was generated as expected. If you like what you see, you are ready to push your
commit to the remote
git push origin gh-pages
Now comes the part you all have been waiting for. You may be asking yourself,
But what do I do when I add new content, i.e., a new page or post, or ANY other changes to the site, and I want to push them to the remote branch without encountering MERGE CONFLICTS?
The complete set of commands for doing just that is the following:
echo "Deleting old publication" rm -rf public mkdir public git worktree prune rm -rf .git/worktrees/public/
echo command and
"" message string is if you want to output
status text in Terminal, letting you know at what stage of the set of
comands you are on. I haven’t done that yet, but I think I will start
echo command as well.
rm -rf public should be pretty familiar by now! It removes your
mkdir public creates a new, empty public directory.
git worktree prune allows you to prune, or remove a
linked working tree.
This is so that you don’t have any merge conflicts between your local
gh-pages branch and your remote
rm -rf .git/worktrees/public/ removes the
public gh-pages working tree.
Next we repeat checking out the
gh-pages branch into the
git worktree feature:
echo "Checking out gh-pages branch into public" git worktree add -B gh-pages public origin/gh-pages echo "Removing existing files" rm -rf public/* echo "Generating site" hugo echo "Updating gh-pages branch" cd public git add --all git commit -m "Publishing to gh-pages (publish.sh)"
git worktree add -B gh-pages public origin/gh-pages repeats the
checking out gh-pages branch into the public directory using the
git worktree feature set of commands which you executed when you first
Recap: this step allows you to have multiple branches of the same local repo to be checked out into different directories.
rm -rf public/* removes absolutely everything from
Again, you have to regenerate your site with the
hugo command. This
recreates a new public directory with all updated production site files and
cd public takes you back into the
git add --all adds all public content to the staging area preparing for
git commit -m "Publishing to gh-pages (publish.sh)" commits your staged
Last, but not least (and not re-mentioned in the Hugo docs), is the
git push origin gh-pages command. Committing is not enough!
Pushing to the remote
gh-pages branch is what completes the
site update for the world to see.
And that’s it! Happy publishing!