[mostly solved] Glass Elevator indexes are stale

When viewing the glass elevator RFC links at https://www.haiku-os.org/glass_elevator/rfc/ there are no links listed. When viewing at https://www.haiku-os.org/glass_elevator the links are broken and out of date. What needs to be done?

Duck Duck Go finds the pages just fine at https://duckduckgo.com/?sites=https%3A%2F%2Fwww.haiku-os.org&q=glass+elevator&ia=web.

Probably some PR on https://github.com/haiku/website/ fixing “_index.html” in both content/glass_elevator/ and content/glass_elevator/rfc so they include links to all the .html files under /rfc.

Thanks! It appears that only the Blog posts ever got Hugo scripts set to generate different indexes in the migration from Drupal. After I’ve patched this in the short term, I’ll need to see what’s necessary to fix the rest of the Blog-o-Sphere using Hugo.


After looking around the sources for the web page, it seems to me that the 6-year-old posts from Glass Elevator should be treated as an author in the Blog-o-Sphere in the same way as the admin account is. That way all of the old posts get listed at the bottom of the blog posts from times of antiquity when the Drupal site was migrated.

Update 2:

Pull request https://github.com/haiku/website/pull/609 has been submitted. Please comment there.

1 Like

codename Glasseleator is teory of Haiku R2. These are wishful thinking, the so-called theory-description. If you know the IT programming market, you know that first you know the bare theory, only later the code is created and much later, for intellectual and economic reasons.
an example from a different backyard is Facebook, where the theory of social networks was created much earlier than cod from the meta company, which committed itself to be one of the first to program this miracle for millions of course.

Hmm, maybe the year should be added to the heading of at least the index page too, just to emphasize the relative age of the content therein? (these things tend to get a bit stale with time)

Glass Elevator (201x)

My pull request left the original date stamps intact so they would appear at the bottom of the blog feed if the pull request is accepted. I ran Hugo on my Linux box using the supplied scripts and was successful in recreating a local copy of the web page and verified it generated the index page correctly.

After some input by @waddlesplash , I’ve closed the pull request. Are there any other ideas of how to merge the orphaned pages into the website without popping a new header up that points to old posts from a 6-year-old migration from another website software?

There should be a page under development/ with links to all these articles, linked from the top development page.


Thanks! I’ll see if I can redo my pull request sometime soon.

…and DONE!


Not quite… :stuck_out_tongue:

The RFC link on the main Glass Elevator page should point to https://www.haiku-os.org/glass_elevator/rfc/, not as currently https://www.haiku-os.org/rfc/ . Also missing applicable “year” timestamps in headings etc as discussed above (stuff like this always need to be read in the right context with the right glasses (no pun intended :wink: ) on to avoid confusion).

I also wish I knew how to get Hugo to arrange to posts in the order of their age (newest first) instead of an arbitrary ordering in the Markdown document I used; alphabetical order of the URLs in this case.

The RFC link was supposed to be a relative link so I’ll look into that. That may need to be made absolute to satisfy the Hugo generator.


The blog-list document index type is part of the blog theme. It was automatic for blog posts but not for article lists. The theme would need to be modified to give article lists an index automation type.

I looked through all the RFC metadata date tags and they covered the 2001-2009 period, so maybe this and a simple # designator could be added to the headings on the (1) main and (2) RFC links pages?

(1) Glass Elevator #1 was a 2001-2009 project for discussing and documenting ideas for future Haiku versions, R2 and beyond.

(2) This is an index of all Request for Comments (RFC) documents submitted to the 2001-2009 Glass Elevator #1 project.

…or something like that :wink:

According to Pulkomandy’s input on the pull request, the project never officially ended. There could still be articles added to the list even today.

My point was more that the existing articles are all more than 10 years old, some of them more than 20 years old even, and it needs to be read with that context in mind. Some of these ponderings may be outdated/obsolete, some of them may have been implemented, and so forth. But it’s up to PulkoMandy and the other core maintainers to decide of course.

I, for one, would like to see the most recent articles at the beginning of the list automatically but I’d have to create an article index list type in the Hugo theme to be able to automate the process. That way the older articles would paginate down to the end of the list where they are less likely to be seen.


I’ve tried making a paginator that generates the links in reverse chronological order based on the blog paginator. I can’t get the article_index type to be recognized as a list.html template in the theme though.

The relative link from the Glasss Elevator directory to the RFC directory works on the latest version of Hugo so I can only assume it’s some sort of regression in the website’s version of Hugo.


The link appears to be working now. I don’t know what happened since then. Somebody else must have fixed it.


I’ve got pagination working but needed to add a “list” page type to the theme to get it to work. As such it doesn’t seem to be a clean update to the web page. Also, the last article posted to Glass Elevator has blatant spelling errors in the title. Furthermore, the importer added an unneeded “Body:” message before the start of each article.


Is it better to make a separate pull request to fix the title and import artifacts, then fix the pagination generator or should I lump them into one?