Live data from Hacker News

Performance and telemetry analysis of Trae IDE, ByteDance's VSCode fork

github.com

231–240 of 390 posts

Re: Performance and telemetry analysis of Trae IDE, ByteDance's VSCode fork

#231
post #206

Earlier quoted context omitted.

The F-91W has long been associated with terrorism for it's use in improvised time-bombs: https://en.wikipedia.org/wiki/Casio_F-91W#Usage_in_terrorism

Oh so now we target people for buying inexpensive things? "Oh no! This guy doesn't have a $500 watch".

Yeah, I think it's pretty stupid, especially considering the F91-W is one of the most common and best selling watches in the world.

Re: Performance and telemetry analysis of Trae IDE, ByteDance's VSCode fork

#232

Am i the only one that finds System Information: Hardware specs, OS details, architecture Usage Patterns: Active time, session duration, feature usage Performance Metrics: Response times, resource consumption Unique Identifiers: Machine ID, user ID, device fingerprints Workspace Details: Project information, file paths (obfuscated) Not to really bad that obtrusive? Like i don't really see anything there that i'd be o…

Imagine your text editor sending network data to unknown resources and there is no way to disable that

Re: Performance and telemetry analysis of Trae IDE, ByteDance's VSCode fork

#233

Am i the only one that finds System Information: Hardware specs, OS details, architecture Usage Patterns: Active time, session duration, feature usage Performance Metrics: Response times, resource consumption Unique Identifiers: Machine ID, user ID, device fingerprints Workspace Details: Project information, file paths (obfuscated) Not to really bad that obtrusive? Like i don't really see anything there that i'd be o…

Yes, you're the only one.

Re: Performance and telemetry analysis of Trae IDE, ByteDance's VSCode fork

#234

Earlier quoted context omitted.

Arguably it was Microsoft who started the whole trend of calling spyware "telemetry" to obfuscate what they were doing.

I can't prove it, but I think that's untrue. Anecdotally, I've only heard MS using it in the last 10 years or so, and it's been pretty common terminology for years before that.

Last 10 years is right - Windows 10 was when they went all-in, and that was released in 2015. Before that, "telemetry" usually referred to situations where the same entity owned both ends of the data collection, so "consent" wasn't even necessary.

Re: Performance and telemetry analysis of Trae IDE, ByteDance's VSCode fork

#235

Earlier quoted context omitted.

> debuggers Parent mentioned >> DAP Here's the docs[0] > popup context windows for docs These are called "balloon"s[1]. Plenty of people have setups for things like docs (press "K") or other things (By default "K" assumes a man page) > contextual buttons on a line of code I don't know what this means, can you explain? > minimap Do you mean something like this?[2] Personally, I use tagbar[3] as I like using ctags and…

> These are called "balloon" And are not interactive as far as I know. I've not seen a way to get a balloon on the type in another balloon and then go to the browser docs from a link in that. > Do you mean something like this? Yes, but that's still restricted to terminal characters (you could probably do something fancy with sixel, but still) - for larger files with big indents it's not useful anymore. > contextual b…

  > [balloons] are not interactive as far as I know
I mean you use completion, right? That's interaction? In insert mode or , same to scroll through options.

  > [tabbar is] still restricted to terminal characters (you could probably do something fancy with sixel,
Wait... you want it as an image? I mean... sure? You could, but I'm really curious why you would want that. I told you this was one option, but there are others. Are you referring to the one that was more visual and didn't show actual text? Idk, I'm not going to hunt down that plugin for you and I'm willing to bet you that it exists.

  > For example options to refactor based on the current location.
First off, when quoting it helps to add more >'s to clarify the depth. So ">>>" in this case. I was confused at first as I didn't say those words (Also, try adding two leading spaces ;)

Second, sure, I refactor all the time. There's 3 methods I know. The best way is probably with bufdo and having all the files opened in a buffer (tabs, windows, or panes are not required). But I'm not sure why this is surprising. Maybe you don't know what ctags are? If not, they are what makes all that possible and I'd check them out because I think it will answer a lot of your questions.

  > Basically where's the "extract this as named constant", "rename this type across the project"
Correct me if I'm wrong, but you are asking about "search and replace" right? I really do recommend reading about ctags and I think these two docs will give you answers to a lot more things that just this question[0,1]. Hell, there's even The Primeagen's refactoring plugin in case you wanted to do it another way that's not vim-native.

But honestly, I really can't tell if you're just curious or trying to defend your earlier position. I mean if you're curious and want to learn more we can totally continue and I'm sure others would love to add more. And in that case I would avoid language like "vim doesn't" and instead phrase it as "can vim ___?", "how would I do ____ in vim?", or "I find ___ useful in VS code, how do people do this in vim?" Any of those will have the same result but not be aggressive. But if you're just trying to defend your position, well... Sun Tzu said you should know your enemy and I don't think you know your enemy.

[0] https://vim.fandom.com/wiki/Browsing_programs_with_tags

[1] https://vim.fandom.com/wiki/Search_and_replace_in_multiple_b...

[2] https://github.com/ThePrimeagen/refactoring.nvim

Re: Performance and telemetry analysis of Trae IDE, ByteDance's VSCode fork

#236

Earlier quoted context omitted.

> Telemetry arguably helps with deobfuscation. Can you please expand on that? I have trouble understanding how telemetry helps me, as a user of the product, understand how the product works.

You can capture the telemetry data with a HTTPS MITM and read it yourself. Or (if you're working lower level) you can see an obfuscated function is emitting telemetry, saying "User did X", then you can understand that the function is doing X.

> You can capture the telemetry data with a HTTPS MITM and read it yourself.

That's not helping me, the user.

That's helping me, the developer.

> Or (if you're working lower level) you can see an obfuscated function is emitting telemetry, saying "User did X", then you can understand that the function is doing X.

Again, it helps me, the developer.

Neither of these help me, the user.

Re: Performance and telemetry analysis of Trae IDE, ByteDance's VSCode fork

#237
post #136

Earlier quoted context omitted.

Every single piece of telemetry sent over the internet includes PII - the IP address of the sender - by virtue of how our internet protocols are designed.

This is like saying every physical business is collecting PII because employees can technically take a photo of a customer. It's hard to do business without the possibility of collecting PII.

No, it's like saying a business that has a CCTV camera recording customers, and sending that data off site to a central location, where they proceed to proceed to use the data for some non-PII-related purpose (maybe they're tracking where in stores people walk, on average), are in fact sending PII to that off site location.

Distinguishing factors from your example include

1. PII is actually encoded and handled by computer systems, not the mere capability for that to occur.

2. PII is actually sent off site, not merely able to be sent off site.

3. It doesn't assert that the PII is collected, which could imply storage, it merely asserts that it is sent as my original post does. We don't know whether or not it is stored after being received and processed.

Re: Performance and telemetry analysis of Trae IDE, ByteDance's VSCode fork

#238
post #99

Why do people use obvious spyware when free software exists?

Telemetry isn't the same thing as spying on the user. People use it because it's not actually spying on them.

Any monitoring of my system without my explicit permission is spying.

Re: Performance and telemetry analysis of Trae IDE, ByteDance's VSCode fork

#239
post #157

Earlier quoted context omitted.

eclipse still is alive holy shit

Installing the VSCode extension pack for Java runs a headless version of Eclipse JDT under the hood, which isn’t quite what I think of as lightweight.

What's wrong with that? If they re-implement the whole thing it would amount to the same code size. It's the JDT language SERVER not some sort of "headless" software with UI needlessly bundled.

https://marketplace.visualstudio.com/items?itemName=redhat.j...

Re: Performance and telemetry analysis of Trae IDE, ByteDance's VSCode fork

#240
post #185

Earlier quoted context omitted.

Mhhh it is not really about “nothing to hide”, it was more that if you use niche services targeted at privacy, it puts a big target on you. Like the Casio watches, travelling to Syria, using Tor, Protonmail, etc… When it is better in reality to have a regular watch, a Gmail with encrypted .zip files or whatever, etc. It does not mean you are a criminal if you have that Casio watch, but if you have this, plus encrypte…

1. I don't really think this is true generally. 2. Using a secure, but niche, service is still more secure than using a service with no privacy. Sure, you can argue using Signal puts a "target" on your back. But there's nothing to target, right? Because it's not being run by Google or Meta. What are they gonna take? There's no data to leak about you. If I were a criminal, which I'm not, I'd rather rob a bank with an…

>If I were a criminal, which I'm not, I'd rather rob a bank with an actual gun than with a squirt gun. Even though having an actual gun puts a bigger target on your back. Because the actual gun works - the squirt gun is just kinda... useless

Actually, there was a case... I can't recall but it might have been in Argentina, where the robbers did explicitly use fake guns when robbing the banks because doing so still actually worked for the purposes of the robbery, and it also reduced their legal liability.

Post reply on HN