I used reader.google.com in the past, but it was shut down and I got reminded that it was a free service someone else was hosting. Then I used feedly, but it was not optimal. Tiny Tiny RSS (ttrss) seems like the best solution for me. https://tt-rss.org/ - Tiny Tiny RSS

Until 2021, I simply used a clone of the ttrss git repo, and updated that clone every now and then. Since 2021, the official/documented way of running ttrss is via containers. In 2023, I created a podman based version, which I'm using since then.

Many pages are not as social as they should be, they are not offering feeds. ttrss can directly read feeds, but for example twitter needs extra care. I use 2 solutions for it:

• rss-bridge https://github.com/RSS-Bridge/rss-bridge : The bridge collection also provides interfaces for example for dilbert.com .
  • pro: provides pictures directly in the generated feed
  • con: can follow only one single person at a time. So I clutter my ttrss interface with 10 rss-bridge-twitter feeds, instead of one which includes all these single persons.
  • con: although I am following these persons, they can not see that
• changedetection.io Self hosted, watches pages. link
  • not using it right now, as the RSS feed does not contain the actual change. So when I monitor a big page, I get notified 'it changed', but not hinted what.
• own scripts: The German/Japanese association in Japan offers no feed, and the page where they announce new events is small - but the contents are broken and meant for visually be read, but not parsed by a machine. I gave up trying to write an own rss-brigde bridge and now fetch with 'lynx -dump https://www.jdg.or.jp/koushin/whatsnew_j.html' the visual interpretation daily, and monitor for changes with 'diff'. Also feed generator ( https://createfeed.fivefilters.org ) looks not to bad trying to generate a feed from the page.

Normally, I just update the ttrss code via 'git pull origin', and schema updates are then recognized. For a friend, the schema update takes so long that it runs into a timeout. How to execute the schema update manually?

The single incremental schema updates are in this directory for Postgres:
$ ls -rt ttrss/schema/versions/pgsql/
[..]
137.sql
138.sql
139.sql

So if one is already on schema 138, one can access 
the ttrss-db and then execute the code from 139.sql.
$ su - postgres
[postgres]$ psql ttrss
<code from 139.sql>

Alternatively, “php update.php –schema-update” should work.

# first, in the ttrss webinterface go to preferences/feeds and
# create an OPML export.  With that, you can restore your feeds
# later if something goes wrong.
cd /opt/soft/ttrss/ttrss-docker
git pull origin
docker-compose stop
docker-compose pull
docker-compose up -d
# login again

• Automatic logout: My browser session with ttrss works just one day, then I am logged out. Configure this in config.php to extend the session: define('SESSION_COOKIE_LIFETIME', 8640000);
• Can I fetch and display content directly from ttrss? FeedIron can do that.
• Mediathekviewweb rss checker. To search for “Midnight”: https://mediathekviewweb.de/feed?query=%23Midnight
• https://github.com/DIYgod/RSSHub
• Japanese Kanji and Hiragana appearing in different fonts? 'gucharmap' can for a given glyph show which font is used to render. The issue with ttrss though is that the theme is guiding to use the wrong font. I played with firefox font settings, but without good results. Now I use a custom theme with modified font settings to fix the issue.

podman exec -it rss-app sh
cd /var/www/html/tt-rss/themes
cp night.less night_cust.less
cp night.css night_cust.css
grep system-ui night_cust.css
sed -i 's/system-ui/"Noto Sans CJK JP"/' night_cust.css