kromonos

cross-posted from: https://fapsi.be/post/161886

A few days ago, a page was published in the Fediverse where you could check which Mastodon instance is blocking whom in the Fediverse. This was made possible by a programmer who used the new Mastodon API introduced in 4.0 and automatically read out the blacklists of the instances.
The information about who blocks who is in many ways more than just interesting for those who want to choose an instance. As it was already written in a German thread about the topic, it would be helpful to know whether you reach your target group at all, if you register on an instance.

Especially with the rush from Twitter, there are more and more bots and crawlers coming up from the Fediverse that just can't manage to send a correct UserAgent.
Many apps, which want to make usage of several platforms, just submit the framework they use for their connection. Like python/1.2.3, Dart/1.23 or http.rb/1.2.3.
The biggest problem with this behavior is, that I as the operator and administrator want to secure the system, but I would lock out legitimate software by targeting only the UserAgent. Especially the python/1.2.3 in particular, but has also attracted malicious attention many times in my log files.
I'm really grateful, that Lemmy developer decided to use a customized UserAgent, so that I as admin can directly see what request it is and where it comes from. That makes it so much easier to make the decision of harm or no harm.
On my servers, I started to block all requests with a “default” or empty UserAgent, but I wonder, how it's possible to fix this issue in general? Any ideas?

Cross posted from https://fapsi.be/post/111920

Please reply on the original post at https://fapsi.be/post/111920

cross-posted from: https://fapsi.be/post/64476

Found this little helper today, tried it, and I'm really happy. Finally, I can write with Markdown and still have an overview of plot, characters, notes, …

There is also a mobile app for iOS.
But in general I am always skeptical about software that advertises itself as “Based in Luxembourg (Germany, …), secure by design”. As if the country in which a software is developed has any influence on how secure a software is. 🤷

I see it very often, especially with Linux, Libre Culture and Open Source, that one and the same is discussed in 2 or more threads. In one a consent was found, in another not, and in the third are only posted links to one of the other discussions. So I asked myself, what if a cross post would only link to the original instead of creating an own entry?

I think I need this 😂

I've managed to include lemmy.ml and fapsi.be search results in my public searx instance on search.bka.li

To do the same, open the settings.yml file of your searx installation, search for the engines: part and add the following lines:

  - name : lemmy
    engine : json_engine
    search_url : https://lemmy.ml/api/v3/search?q={query}&page={pageno}&type_=Posts
    results_query : posts
    url_query : post/ap_id
    title_query : post/name
    content_query : post/body
    categories : general, social media
    shortcut : le
    paging : True
    disabled : False

After restart, you're able to see and select lemmy at the searx preferences page under the tabs general and social media. A search would look like this:
Source: search.bka.li

You can also add comments:

  - name : lemmy comments
    engine : json_engine
    search_url : https://lemmy.ml/api/v3/search?q={query}&page={pageno}&type_=Comments
    results_query : comments
    url_query : post/ap_id
    title_query : post/name
    content_query : post/body
    categories : general, social media
    shortcut : lec
    paging : True
    disabled : False

As well as communities:

  - name : lemmy communities
    engine : json_engine
    search_url : https://lemmy.ml/api/v3/search?q={query}&page={pageno}&type_=Communities
    results_query : communities
    url_query : post/ap_id
    title_query : post/name
    content_query : post/body
    categories : general, social media
    shortcut : leco
    paging : True
    disabled : False

User search is possible too, but adding this would only support bullying and stalking. For ethical reasons, I decline to post an example of this as well.
Be careful: Never use the same shortcut for different engines. When your searx will not load or time out, check the shortcut first.

Is it possible to get Lemmy's search result in JSON format? E.g., to include the search in search engines like searx. That would be really awesome.

🤦🤦