Why is cursor started having issues again with Indentation. It happened a while back, then it was fixed, now it’s started again. Will this problem ever go away. And the number of mistakes the bot makes is getting worse. But i suppose it depends on model, but in Auto, its certainly seen a decline recently.
Hey, could you provide more details on what indentations and in which files? Are you using any extensions for the programming language you’re using?
doing this to me a lot when testing grok-4-code, using python, i always have to go back and fix the indentations as it orphans large sections of code , easy to fix but annoying.
Does this happen with other models, like Claude? Since the Grok Code model is relatively new, it might produce incorrect code.
Can’t answer that, I know it wasn’t doing this in Claude 4 or ChatGPT5 when i had the paid subscription. Now in auto, its horrid. in fact ChatGPT5 has massively improved from the first time I used it on initial integration. did you apply any fixes in the last hour or so, as the indentations have slowed. they were horrendous since i started using the Auto mode a couple of days ago and been pulling my hair out, tried restarting, it would work ok at first, then quickly return and cause problems, can’t give you much more context than that, I’m not using any extensions I am aware of. the other issue it has, is it hangs a lot, when thinking. for example its curl -X POST “http://localhost:5000/api/ml-trading/execute-cycle” I have to click shirt to get it to do anything otherwise I will be there for hours and it may never do anything. it just crashes instead if I leave it over night. frustrating, it does this every few minutes. i shifted it, now its stuck again on : Let me wait for it to start and then test again: L Start-Sleep 10; curl -s “http://localhost:5000/api/health” then once its finished it tells me its still go the same errors, lol its useless, i’m tired. The other annoying thing it does is when you finally get it to a point where it finally fixes some code, it will not suggest any next moves, it just says when it did and leave it there.
Did you guys just fix it?
Yes, the team is working on it. Is it working for you already?
1 Like
Yes, i can see a market improvement already, its still hanging every now and then, but so much more fluent. Less stalls. Might be stalling when they are working on it live. I can see accuracy is improving. it makes barely any indent errors now. Well done and thanks. Soo much better now.
I also noticed the Emoji problem is also gone, fantastic!! it just hangs occasionally, when copying or checking files moving or testing. But will work if I ask it if its finished yet, So its hanging for longer than it needs to, This i could just be the auto model used, hope this makes senses, but 95% improvement.
1 Like