Can't add atom.xml to Inkwelll

Hosting on Micro.blog
1

I’m trying to subscribe to https://danluu.com/ , either with the plain url, or with the direct url to the atom.xml: https://danluu.com/atom.xml

In the first case it simply fails to subscribe, in the second it adds a suscription entry to an empty https://danluu.com/atom/index.xml URL.

P.S. Where should I be reporting bugs? While in general I like the UX choices and find the implementation of POSSE in micro.blog very nice, I struggle to recommend it to ordinary people (i.e. not FOSS nerds) because the whole thing lacks QA.

2

I don’t have any behind-the-scenes insights, but I notice two things about that feed:

  1. It’s pretty big, 6.3 MB.
  2. It does not validate.

I know @manton has limits on feed size in other parts of the system. So it could be that, or maybe Inkwell just trips over the broken feed.

Send bug reports to help@micro.blog.

1 Like
3

I checked that feed and I think the main problem is that the posts are all older than a couple years. Inkwell focuses on recent posts. The UI will usually only show posts from the last week, although it can show older posts when selecting a specific blog.

Also it is a big feed and will perform better if it was a lot smaller.

That danluu.com/atom/index.xml URL is showing up because it’s referenced in the feed. I think Inkwell is discovering it by accident, so that is something I can fix, but the feed is also probably wrong.

1 Like
4

Thanks for looking into this.

To be clear, I haven’t faced a situation where a URL that points to an actual XML file is saved with no entries appearing, so this is not just a misapprehension on my part of how the UI works.

The only feed URL that is saved is the one that returns 404 danluu.com/atom/index.xml .

So I suppose the primary difference in behavior between the Liferea RSS client (that imports the feed successfully) and Inkwell, is that Liferea uses the URL provided directly (https://danluu.com/atom.xml) while Inkwell attempts to read the one from within the <atom:link> tag, as you noted.

I gather since there’s multiple things wrong with this feed, there’s no obvious fix.

I see that, if i try to import https://danluu.com/atom/index.xml directly, it refuses to be saved because apparently Inkwell tries to validate and gets a 404.

Perhaps it would be safer for Inkwell to save the same URL that it validates (i.e. https://danluu.com/atom.xml) or, alternatively, to also validate the URL found inside the <atom:link> tag prior to saving it.