Skip Navigation

Posts
130
Comments
162
Joined
3 yr. ago

  • That may be due to "the Hacker News hug". This link had a high score in HN a few hours ago. High traffic may have caused scalability problems.

  • To be honest, I cannot be sure that session invalidation actually worked. I could use the session from the day before as well. But the vulnerability was in lemmy-ui, and people not using the web site directly should be fine, I guess. If you want to be on the safe side, you can log out and log back in. It takes only a few seconds.

  • Where are you seeing this happen?

    It's at the top of the page when you visit programming.dev in a browser. I tried to explain what I know about it in a sibling comment.

  • Here is my understanding:

    Recently, a security vulnerability of Lemmy has been exploited by some malicious actors. This lead to some instances going down. The vulnerability has been fixed with version 0.18.2-rc.1 of lemmy-ui. But due to the way Lemmy issues and uses access tokens, the sessions has been invalidated in the database. So, the admins are recommending the users to log out and log back in if they haven't done so after the upgrade to version 0.18.2-rc.1 of lemmy-ui.

    But I may be wrong. Perhaps others can provide a more accurate description.

  • Ok, maybe I misunderstood your question. I though you were proposing # instead of $ sudo and I meant to say that being explicit is better.

  • That sounds cool. Thanks for the recommendation.

  • What about the packages that are not available in flatpak? I assume there must be some packages that are only available in certain corners of the internet?

  • Thanks

  • Are there any other distros that are flatpak-only?

  • I don't work much with Linux systems these days, but I would vote for $ sudo over #. Two reasons:

    1. It's easy to overlook the prompt. That part is basically "some characters before the actual command", so I don't normally pay attention to it.
    2. # is also used for comments. I think it would be confusing to use the same character for two wildly different things.
  • It doesn't look very good, no. It would be good to bring Lemmy to OAut2.1 where the self-contained token with a sensible lifetime is passed in the Authentication header. Currently it's either passed in the URL (GET) or in the model (PUT/POST).

    I have some OAuth experience, but I'm not a Rust developer. So, I thought of offering some help regarding design and testing of an OAuth mechanism, but since I cannot really contribute to implementation, that may not be that much of a help. Also, this kind of a change will break at least some of the existing clients. I don't know if the core team would be willing to make such a change.

  • Would love to see a browser based implementation of this.

  • Unflushable Cache

    [...] Other implementations, such as hand cranks in memory caches or even caches provided by mainstream frameworks will not expose any cache management tools. This leaves ops with the only option, to restart the service to flush the memory. (Or worse, know enough about the cache implementation to find it’s location on file system and clear it out manually.)

    This is a mistake I have made. It's easy to overlook during development but difficult to handle afterwards if restart is not trivial.

  • Fixed, thanks

  • Thanks. I didn't know that when one adds an image it would override the URL. Even though the post contains the URL, clicking on the title only shows the image. I included the URL in the post body, but it's not as visible, unfortunately.