• 4 Posts
  • 265 Comments
Joined 1 year ago
cake
Cake day: June 12th, 2023

help-circle
  • Facebook marketplace is literally a shit show full of scams and it is hard to find anything, a bad UI, and no good filtering. The only thing holding it up is the algorithm to show you products (based on your browsing and advertising data). It is a step up from Craigslist, but it is mediocre at its absolute best. There are just no used alternatives in many places

    We have an app here called 2dehands which has a similar scam problem, but is easier to filter out. However, the UX is 10x better with robust filtering, setting up pre-specced searches based on location, price, text strings, quality, categories, etc… with email and/or app notifications for new items added and everything surrounding bidding and favorite item status.


  • Nope. I type in SPAR and it gives me a bunch of random results 150km away.

    I type in Grocery and it gives me SPAR 1.5km away.

    I need to type in the exact address for it to find a place, with no errors. Otherwise I can type a category name and hope that it finds it and that I guessed right, that also works.

    If you type in a partial name (I.e. not the full legal name including company abbreviations) of a store it will break itself and show you completely random unrelated results from a random place in the country.

    I took a bunch of screenshots a while back as proof.

    It is fundamentally broken, and it is widely reported IIRC, but I don’t know if there are any issues open about it.











  • There is one neat trick: don’t expose SSH.

    There is still not a reason anyone has been able to give for 99% of self-hosters to expose SSH.

    If you need to access your machine via ssh while on the go. Wireguard to your local network, use SSH. Done. Unless you are running an always-up public facing site, the amount of times you have to access your machine that can’t wait until after work is very low anyway.

    Bots will scan all ports. That is just how it works. Less than 22, but you will still get spammed. Why force your computer to go through the fail2ban loop and take up resources when it is simply not needed at all and you can block it on another machine?






  • Platformio maybe?

    That thing just sometimes won’t work on some PCBs unless you explicitly specify a build option (I had it once that I had to specify a build option that was already specified in that board configuration)

    Installing pycom stuff too. All of their software is crap (hardware isn’t amazing either) so their released version simply didn’t work by default on a fresh installation and the fix was to roll back to the previous version manually.


  • Solid works does the same thing though. Not crashing but even opening a simple model takes ages in solid works and the vast majority of things are single threaded there.

    Whenever we screen share a part in solidworks, it is literally 5-10 minutes of the meeting taken up by waiting for it to complete visual operations, load things in, and assembly constrainy computing.

    And you pay a shit ton of money for solidworks. Freecad also has these problems, but it is surprisingly not extremely worse than some professional cad software outside of crashing. Topo naming problem, UI, and crashing was definitely the worst thing about it. Apparently 2 of those 3 are getting fixed now.



  • The problem I have with bug reports for things like docker containers (I will just use that example for now) is again documentation… I get that most people make really bad “it broke, fix it now” kind of reports with no detail, but unless you actually dig into and know the code, there is often almost no debugging documentation (github issue templates help with that when the devs make one)

    Where are the logs? If you refuse to use the docker standard of pushing logs through the log api, don’t respect the LOG_LEVEL environment variable, errors are non-descriptive, and don’t provide documentation to where your logs are stored, how the hell can you expect users to provide relevant logs??

    I have run into dozens of pretty big projects that fail with 0 log output and there is 0 log documentation and then the dev auto closes the issue because “not enough logs to help.” The only way to find the logs is to find an old issue where the dev has laid out where the relevant logs are, like WTF? No I am not going to spend an hour looking in every single part of the undocumented directory structure to see if there happen to be logs in there. Use the standard, document your log locations and what they log (no 3ch9qjV7.log is not descriptive enough), or don’t complain about not being able to help without logs.