This blog is now hosted via ipfs as well.
What does this mean?
This means that you can access this blog via this IPNS name directly. As it is hosted in a distributed manner, a node which has the content must be online, else the retrieval fails. I try to keep nodes online which host this content, though I cannot promise anything, of course.
Old versions (starting today) of this blog can be found here, also hosted via IPFS. So you can see all typos I correct now and then.
How can I help?
hash=$(ipfs name resolve QmX95tkM6em8MP1SDs9Qae1G9YscqwozQJbX5rWTTYcJea) ipfs pin add -r "$hash"
The newest version will be published under this name:
So you can poll the newest version of this blog by re-resolving this name and
pinning the contents via
ipfs pin add -r.
How is it done?
Well, that’s rather trivial. I have an ipfs daemon running on my machine and a bash script for adding the contents of this blog to ipfs. Each time I run this script, it adds the blog contents to ipfs and writes the last hash to the “old-versions” file. After that, it publishes the name.
So my workflow is like this:
hugoto update the blog sources locally
- Run my ipfs publish script to publish the site in ipfs
- Run my publish script to publish the site on beyermatthias.de
You really should host your things with IPFS as well!
What I had to do before this worked with “hugo”: I had to ensure hugo did not generate absolute URLs. Because if each link links to the full URL of the blog (beyermatthias.de/yaddayadda), each link in the IPFS-checked-in-version links to this URL as well. Instead, each URL has to be relative to the current location, then, the IPFS versions link correctly.