Johannes Süllner <johannes.suellner@mailbox.org> writes:
On Fri Dec 26, 2025 at 9:37 PM CET, Demi Marie Obenour wrote:
Nice catch! I wonder if automation can reduce the chance of similar problems. Could a template engine be used to reduce repetition? Or at least check that all of the internal links are valid?
I wondered the same after finding these broken links recently (I just procrastinated sending a patch), and looked for tools to find broken links. I found a few, the one I liked most is lychee (https://github.com/lycheeverse/lychee).
Using Nix, it can be run on the homepage repository as follows: nix-shell -p lychee --run 'lychee --root-dir $PWD .'
Doing this, some broken external links can also be found. Some external links are being reported as broken erroneously for me, namely to gnu.org and doi.org. However, there are also two links which are broken: 1. On logo/index.html: https://hazelnot.xyz/ 2. On motivation.html: https://docs.saltstack.com/en/latest/
The website of Hazelnot seems to be offline since about June 2024, see https://web.archive.org/web/20240301000000*/https://hazelnot.xyz/ The documentation of Saltstack moved to a new domain. Qubes moved their docs too, although they have redirects in place. I have a patch for this already which will follow in a minute.
Regarding the broken internal links, I wonder if we would be better off using absolute links to avoid copy-paste errors in the future?
That would prevent navigating the website locally during development, so I'd rather not. It probably does make sense to move to a static site generator at some point though, especially if the website acquires more content.