This page is some detail regarding the setup of this website.


I have shared hosting with

Reclaim Hosting provides institutions and educators with an easy way to offer their students domains and web hosting that they own and control.

I like their model, I’ve found them responsive and it’s reasonably priced. It’s let me play with various systems (my website was on Wordpress for a long time, but I also ran a wiki instance, I played with Drupal, etc.) through CPanel.

Website writing

Since I moved off Wordpress to quartz (more on Quartz), I’ve also been using Obsidian as the main place I edit and create content.

My Obsidian instance includes a set of plugins, that currently comprises:


Website building

I now build my website on my PC using Quartz, and end up with a set of static html+ files. Those need moving online to be hosted and viewable for the world. I was doing that through FTP for a bit, but it’s very inefficient.

What my workflow now looks like is:

#npx quartz update
npm i
# git pull content upstream
#build quartz site
npx quartz build
# or to preview/view
npx quartz serve --fastRebuild
#add new items to git
git add .
# commit new and edited
git commit -m "customise"
#connect to server to push (assuming SSH setup via putty, and saved with name "reclaim")
putty -load "reclaim"
plink -load "reclaim"
#this may need more faffing in windows to cache the key, you might need to add an environment variable GIT_SSH for plink.exe
#push updated site from local machine to server
git push live content:main
git push origin content:content

How to sync static built site to server

What that rests on is using a post-receive script in a git repository hosted in my reclaimhosting space.

This largely reflects the very helpful instructions but (a) I’m on Windows, and (b) I found some bits confusing/I did things in a different sequence, so it shows a bit more detail.

Step 1: ideally work out SSH for your reclaimhosting account

(see their instructions)

Step 2: Setup directories on reclaimhosting space

You can do this either via SSH (using the instructions ), or via the CPanel GUI file manager (how I did it).

You need to create three directories:

  1. Name: gitbare, Description: This’ll hold a ‘bare’ git repository, it’s where the post-receive script below will go.
  2. Name:, Description: This is where your entire git history for the repo will go
  3. Name: public_html or a subdomain (preferably), Description: This is your live website

You also need to know:

  1. What the name of the directory that holds your public built website is (mine is public)
  2. If you want to add another directory, what that’s called (and, possible security issues)

Step 3:

Either using SSH or the CPanel terminal, you need to:

  1. Run git init --bare inside the directory gitbare
  2. add the script below (edited) as your post-receive inside the hooks directory (you’ll see there’s a sample in it already)

I did ‘2’ using the browser editor and CPanel file manager.

Make sure to change the two variables at the top so DOMAIN is where your website will be served from (setup a subdomain in CPanel so it will load from this directory if you need to), and DOMAINTRACKING is where the git history is.

  # instructions from
  # Replace this value with the actual domain name (or subdomain) that you want to serve from
  echo "~~ Updating $DOMAIN ~~"
  # Clearing git env
  unset GIT_DIR
  # auxiliary domain storing the entire tree (all repo content). You can name this what you want
  # DONT FORGET below you need to create this dir
  git pull
  echo "~~ inside the auxiliary domain ~~"
  # get to the static build directory, for jekyl _site for other builders public (or whatever you chose)
  cd public
  # delete domain contents
  rm -rf ~/$DOMAIN/*
  # copy the current directory (public) to our domain path
  cp -R * ~/$DOMAIN
  # if you want to add files that arent in your public directory, you can do that (assuming the directories exist), e.g. I have a htaccess file that my SSG doesn't build, so I keep it separately 
  echo "~~ inside the supplement to public dir ~~"
  cd ../web
  shopt -s dotglob # include hidden files in the wildcard match
  if [ "$(ls -A .)" ]; then
    cp -R * ~/$DOMAIN
  shopt -u dotglob # disable dotglob
  echo "~~ Done ~~"
# On server you also need to then
	# git clone you-tracking.git
	# or
# On your local machine assuming you have a repository already hook it up to the remote, live is the remote name
	# git remote add live ssh://<polyrhet>@<servername>/~/<webdomain>.git
	# git remote add live ssh://
	# then push, to live, from main (or in my case content)
	# git push live content

Step 4:

On your local computer, if you don’t have a repo, follow the instructions at

If you do, then you can run:

git remote add live ssh://<polyrhet>@<servername>/~/<webdomain>.git (changing it for the address for your barerepo).

You may need to fix branch names or connections.

You’ll see above that I have two (at least) remotes, and I push

To my live website on reclaimhosting, using: git push live content:main (i.e., my local branch is called content and the remote server branch is called main)

And to github using: git push origin content:content