Date: April 13th, 2023 8:56 PM Author: Talking dark messiness
Original spam blocking tool here. This solution works in all browsers: Chrome, Firefox, Edge and Safari. It does NOT use a custom extension. It relies on a popular browser extension called "Tampermonkey" that runs javascript on specific pages to modify the content. Took a couple of hours to write.
4) This script will remove posts that contain content you specify, so you can modify the script to remove any type of spam post. Change or add a string on the line that reads "const spamPostsContent =" to whatever you want. This text is from the nullo spam post.
FYI, this will also remove the preview post of any spam.
Date: April 15th, 2023 3:20 PM Author: Talking dark messiness
Don't agree.
Well, maybe if you're talking about people who literally only write shitty scripts and stuff then maybe. But actual SWEs spend probably 20% of their time writing code. The other 80% is spent doing things like systems design, reviewing backlog, fixing issues/handling tickets, etc.
you also have to be very descriptive and specific about what you want tools like ChatGPT to generate and even then it has issues with "hallucination." Overall I see generative AI as a tool that will boost dev productivity a lot but not replace devs.
Date: April 14th, 2023 9:40 PM Author: Talking dark messiness
I would argue that this is easier to install. And you can block multiple types of spam post by just creating a new script that filter on different strings. And this works on all browsers.
But whatever. We’re all on the same team here in our pursuit of no spam on xoxohth. I encourage everyone to use either solution.
XO Java should work for pumo and moniker blocking, you can copy the code that removes the poast and its sub-poasts from both the thread tree and the thread
Date: April 13th, 2023 9:11 PM Author: Talking dark messiness
Yeah, it's super tedious because the post elements don't really have easily identified attributes or classes. I basically just look for certain HTML strings and/or the links to certain images on the server associated with posts.
added console and error logging (so you can see that script and template loaded); config to block strings yourself (saved in config.json with your presets loaded already); and also a more general feature in which a simhash of user-cached xo content is created and stored, and then the hashes are compared using the hamming distance measure, so that _any_ long repetitious spam will be removed auto-magically.
(the crypto-js thing is https://cdnjs.com/libraries/crypto-js -- from standard library but can't dl from cdn so included it as part of the package).
should be a general solution but if the trannies want to get kyoot we can go the ai route.
just download as zip, extract, load the file as chrome extension. Tested on brave only.
HTH.
EDIT: fixed test bug but it is now verified working + tests. also note, the way this works is that you store a local cache, not of plain text, but of hash values extracted with simhash from stuff you browse on the bort, then if there's a match that is too close it's removed; but you can also edit the config.json file to remove strings you find. Enjoy!