tbe

> I also don't see why you think fetching ALL results is going to be faster than limit and offset. Some comment tables have hundreds of thousands of comments. Fetching...

> As to the wide join speed issue, @johanndt found a great way to optimize the existing queries: to do a subselect _before_ the rest of the heavy joins. This...

> It means that the database has to do the filtering for each individual user doing blocking of other users, and some people really like to ignore other human beings...

> I don't see why this is a problem? Literally every other forum software works this way. Even Reddit works this way. If you view the front page, and then...

> One of the major problems it that there is little to no caching at all, no second-teir caching that is routine in server applications. You can cache the contents...

> The Rust code is hiding the problems, and the application crash logs of lemmy_server is a platinum mine of scaling problems. Can we please limit this discussion to SQL/Database...

> If it isn't the SQL faulting, what is it? We have trouble with SQL, that's why i'm here after all. Yes. But that an error handler on the HTTP...

Last time i try this @RocketDerp : Yes, that the JSON API Errors are not JSON in many cases is an issue, but not this one. It also isn't about...

During load testing, i have seen, that there is one statement, that is executed far to often, and leads to duplicate entries on the database: ``` INSERT INTO "local_user_language" ("local_user_id",...

Thank you @L3v3L ! Not as bad as i thought, but also: not good. If a user signs up without a given language, all 184 ID's are inserted in this...