Textpattern CMS support forum
You are not logged in. Register | Login | Help
- Topics: Active | Unanswered
#16 2024-02-14 14:21:29
- Algaris
- Member
- From: England
- Registered: 2006-01-27
- Posts: 605
Re: Textpattern docs site has moved to a new server
I keep hearing about Eleventy as an alternative to Jekyll. Apparently it’s built using JavaScript (Node.js), so no Ruby.
Also thank you for the hard work you’ve put into this. I’m constantly amazed at how much you do for the Textpattern ecosystem.
Last edited by Algaris (2024-02-14 14:23:16)
Offline
Re: Textpattern docs site has moved to a new server
Algaris wrote #336650:
I keep hearing about Eleventy as an alternative to Jekyll. Apparently it’s built using JavaScript (Node.js), so no Ruby.
I had a poke around with Eleventy a while ago for another project, then looked at the NPM ecosystem that I’d need to grok and automate…and that gave me pause.
The lack of search is a pain, honestly. The irony of the docs being searchable text content for a CMS that has good search functionality (Textpattern) that’s using another entirely different CMS to display it without search functionality (my remix of Jekyll) is not lost on me.
In a future timeline, I’d love to see Textpattern be able to import a bunch of supplied, pre-formatted content and turn it into articles, images & files. We sort of have it a the moment with the setup article, but something beyond that to check for content in a given directory and then build the article scaffold…that’d be amazing.
I can bodge it with a scheduled SQL import, but that involves getting the articles from the repo, formatting them in a way that won’t snap a Textpattern instance in half, build the SQL, import it, and away we go. No Ruby. No NPM.
I’m not averse to hosting docs on a third party service that does it well. It frees up some of my time and energy to work on other things that don’t involve Ruby.
And thank you, I appreciate your kind words.
Offline
Re: Textpattern docs site has moved to a new server
gaekwad wrote #336626:
I can power off the old server, save us some $$ and move to the next task.
This is now done. tarzan.textpattern.net is no more.
Next task: move servers to Debian 12.
Edit: this is scheduled to follow the release of Textpattern 4.9.0, so we can roll with Debian 12 + PHP 8.3. I don’t think I will have fully figured out PHP on Nginx Unit (we’re using PHP-FPM at the moment) by then, but who knows what the year will bring!
Last edited by gaekwad (2024-02-14 15:05:15)
Offline
Re: Textpattern docs site has moved to a new server
Honestly, the only reason we chose GitHub pages in the first place was so the community could own the docs and we wouldn’t have to do it. Submit PR or request commit rights and away you go.
If that’s not on the table or uptake isn’t overwhelming there’s absolutely no reason not to up sticks to a Txp instance which can do it better.
The smd plugin menagerie — for when you need one more gribble of power from Textpattern. Bleeding-edge code available on GitHub.
Hire Txp Builders – finely-crafted code, design and Txp
Offline
Re: Textpattern docs site has moved to a new server
And if we do move, I can feel a shell script running all the pages through pandoc coming on…
The smd plugin menagerie — for when you need one more gribble of power from Textpattern. Bleeding-edge code available on GitHub.
Hire Txp Builders – finely-crafted code, design and Txp
Offline
Re: Textpattern docs site has moved to a new server
gaekwad wrote #336652:
In a future timeline, I’d love to see Textpattern be able to import a bunch of supplied, pre-formatted content and turn it into articles, images & files. …
I can bodge it with a scheduled SQL import, …
One could also send this content to some etc_post entry point, that would then publish it as txp articles (passed via a md filter). And etc_cache could save them as static html files, that would be served when the search is not involved.
Anyway, thanks for your work and patience.
Offline
Re: Textpattern docs site has moved to a new server
Bloke wrote #336656:
Honestly, the only reason we chose GitHub pages in the first place was so the community could own the docs and we wouldn’t have to do it. Submit PR or request commit rights and away you go.
Oh, I absolutely agree with the decision, Bloke – please don’t take my posts as gripes, far from it – GitHub Pages was basic and mostly worked until it didn’t, hence the move over to self-hosted Jekyll…we were 95% of the way there with GitHub Pages so it wasn’t a massive stretch to get over the line, it’s the underlying bullshit that’s a pain point.
Bloke wrote #336657:
And if we do move, I can feel a shell script running all the pages through pandoc coming on…
Totally! No need to reinvent the wheel. If and when I can wrap my noggin around GitHub Actions and the CI / CD tubes, there’s nothing stopping us having multiple versions spat out on each commit. EPUB docs coming to LeanPub in the not too distant future…
etc wrote #336658:
One could also send this content to some
etc_postentry point, that would then publish it as txp articles (passed via a md filter). Andetc_cachecould save them as static html files, that would be served when the search is not involved.
This sounds far more elegant than my sledgehammer approach.
Offline
Re: Textpattern docs site has moved to a new server
gaekwad wrote #336652:
In a future timeline, I’d love to see Textpattern be able to import a bunch of supplied, pre-formatted content and turn it into articles, images & files. We sort of have it a the moment with the setup article, but something beyond that to check for content in a given directory and then build the article scaffold…that’d be amazing.
👍👍👍 Just wanted to thumbs-up this – but also the other follow-up ideas. If there’s an automatable path from PR-able docs to Textpattern, that would be great!
TXP Builders – finely-crafted code, design and txp
Offline
Re: Textpattern docs site has moved to a new server
jakob wrote #336661:
👍👍👍 Just wanted to thumbs-up this – but also the other follow-up ideas. If there’s an automatable path from PR-able docs to Textpattern, that would be great!
Tagging for info & relevance:
Offline