In one of my feed I have links to images and resources like
// site.name / image.jpg
(i cant post links, so just delete spaces)
so when tt-rss tries to get images, it tries to get it from https, but on this site there isn’t any secure protocol at all or any “https” links in rss feed itself. So for all links I have errors. Is there any way to fix this?
tt-rss uses https for schema-less URLs with no fallbacks to plain-text.
i’m not sure how technically correct this is (and i’m not about to start digging in RFCs) but it’s not going to change, because the alternative i.e. tracking all schema-less URLs and requesting them again if SSL connection failed is highly inefficient and overall sounds like a bad idea.
as far as i’m concerned schema-less URLs are a stupid meme invention with no reason to exist. using them when you only serve plain-text http is doubly stupid and pointless, so there’s not going to be any core workarounds for sites like that, especially because the vast majority of websites use SSL anyway and the rest can either get on with the program, go back to fully-qualified URLs as God intended anyway, or simply fuck off.
that said, nobody is going to stop you from adding all sorts of per-feed workarounds using plugins. good luck.
OK. I understand. Thanx for explanation.
Relative urls is fine for a webpage but suck in a rss feed. You can no longer control the base url. “In one of my feed”, if you’re the publisher, then you need to fix it. If not, then write a plugin to rewrite the relative url using the feed schema like fox said.
I have seen feeds with relative links and images, I just shake my head at an amateurish attempt at publishing. :-/
And then there are those “hand-crafted” RSS feeds which always use HTTP as the web site at some time switched to HTTPS but forgot about their ancient code which generates the feed and still has the http:// prefix hardcoded.
A lot of my custom plugins contain fixes for this bullshit.
Relative links should be fine if the feed includes the xml:base attribute to define a base URL. I’ve seen it included in the element itself but at the moment I’m not finding an example of it being used to make relative links work in any of the feeds I’m subscribed to. I have found examples of it being used (e.g. kottke.org) as an attribute on individual elements within entries of an Atom feed.
this is cancerous atom-specific autism and personally i severely dislike it. it adds complexity to the parser for no good reason.
is it really that hard to just put the goddamn fully qualified URL instead of inventing all this garbage to save a few bytes.