mono.git/hostnix/elmo/rss.nix, branch testUnnamed repository; edit this file 'description' to name the repository.http://git.mou.fo/2026/mono.git/atom/hostnix/elmo/rss.nix?h=test2026-08-26T17:38:22ZConvert miniflux to systemd socket activation2026-08-26T17:38:22ZJoe Moudev@mou.fo2026-08-11T17:26:08Zurn:sha1:3300c063ac1eeeaa8d37bd1ccd468fdc75c3ccf5
Fixes socket permission issue. Drop the RuntimeDirectoryMode workaround:
nginx no longer traverses /run/miniflux.
Also drop the now-redundant explicit socket dependencies from bjj-booker
and clippersnip; systemd adds those implicitly.
Remove danb/rss and host Miniflux on https://mf.mou.fo2025-04-16T04:13:55ZJoe Moudev@mou.fo2025-04-09T03:48:55Zurn:sha1:99ac59f5d550339252b2730245075d1945b264c5
We still want SSO. It seems most promising to register Miniflux as an
OIDC client, but this is pending switching to a different identity
provider.
Alternatively we could use oauth2-proxy and configure AUTH_PROXY_HEADER.
We would want to bypass for API endpoints:
- /accounts/ClientLogin
- /reader/api/
Try minimal danb/rss aggregator2025-04-09T04:04:36ZJoe Moudev@mou.fo2025-04-04T23:46:03Zurn:sha1:a453a7a062bcc6677a7a527032c3309dff2195a5
Probably won't keep this, but playing with OCI containers and XSLT were
interesting.
Converting Feedly OPML to feeds.txt:
$ nomad @xmlstarlet select -T -t -m '//outline[@type="rss"]' -v ./@xmlUrl -o $'\t' -v ./@title -o $'\t#' -v ../@text -o $'\n' Downloads/feedly-093988c0-b9a0-4bb7-97dc-6946128b509e-2025-03-30-d63a70cb-archive/subscriptions.opml | awk -F'\t' -v OFS=' ' '{gsub(/ /, "_", $2); gsub(/ /, "-", $3); $1=$1; print}'
Miniflux RSS aggregator2025-04-09T04:04:36ZJoe Moudev@mou.fo2025-04-04T22:40:49Zurn:sha1:c8feb7e24e2241d79fb17aa3fb8ee587b03c6973
Have not allocated the mf.mou.fo subdomain yet while testing the app.
BASE_URL seems nonessential, but at least fixes the API endpoint given.
To create the initial admin user:
$ sudo -u miniflux env DATABASE_URL='user=miniflux host=/run/postgresql dbname=miniflux' miniflux -create-admin
The UI is quite clunky. In particular it is very easy to get lost
navigating between article list and detail views. It would also be nice
to force opening articles on the external site (for Phoronix). Fetching
original content (instead of using the RSS content) is nice though.
To integrate with oauth2-proxy, probably need to set AUTH_PROXY_HEADER.