@blakeyrat said:
I don't think typing in passwords is this horrible awful thing we needed this stupidly complex software (and the various "helper apps" it requires like Pageant) to solve it.
You don't seem to understand how public key cryptography works or why people use it. It's ok, you don't need to.
@blakeyrat said:
In any case, Git, if its developers weren't lazy and incompetent, could have simply exposed their username/password login over SSH. They'd already implemented one for HTTP(S). The code's already there.
And they have? Git doesn't care how you authenticate over SSH. Now you called people incompetent and lazy again because you don't know shit about anything.
@blakeyrat said:
You're beyond "bugs" into "your method of developing software is inherently broken".
Right, because nothing good can come out of writing your software in the public. Even your beloved Microsoft does it now. But I guess you're ok with it because they mostly have everything closed.
@blakeyrat said:
I didn't say it didn't work, I said it was shitty software. I have no interest in dancing bears.
Wat?
@blakeyrat said:
@hifi said:It's very much possible to write a good Git client,
Nope.
You'd think so, but then you think about things like: Git's error messages, which are impossible for a machine to parse. Or the features like pre-commit hooks, which basically require Git to run a ton of other software.
I don't think it is possible to write a good Git client. Maybe a passable one. But a good one? Not without changing Git itself. And that won't happen because Git's developers are incompetent idiots. (And they've coded themselves into a corner-- they can't change Git because since people use the CLI as an API, if they did they'd break software that uses Git.)
A good GUI Git client or integration does not use the Git CLI at all and client side hooks are useless and usually .
I don't think the Git developers care about other people using their CLI as an API, they rather like their CLI as it is. You can correct me on this if you find a reference which says they didn't improve something in the CLI because they are too scared of Atlassian.
@blakeyrat said:
@hifi said:Don't lie, you're really bad at using Git and you've proven it so many times.
Ok; so what?
Good.
@blakeyrat said:
The point I'm making is that Git requires an extremely high amount of cognitive load. So brainpower I could use on making my programming more efficient is instead being wasted memorizing all these various quirks about Git.
That was my point. Whether or not I am good at Git has nothing to do with that point. If you're going to debate me, please stick to the points I am making.
There's more wrong in the use cases I've heard you using it than Git itself.
@blakeyrat said:
Why is libgit2 even necessary? Why doesn't Git provide its own API? For a product like SOURCE CONTROL, which of ALL types of product is the MOST in need of being embedded in other applications.
Because an API wasn't necessary for the intended use case.
@blakeyrat said:
You don't know how the Recycle Bin works, do you? Because that's gibberish.
I think I know it's not a magic box that frees all the disk space the files were using before. Correct me if I'm wrong.
If my disk is full and I move 50 gigabytes of high-definition fetish porn to the Recycle Bin, how much free space does my disk have then? I'm pretty sure there's still 50 gigabytes of high-definition fetish porn somewhere and I can't download more.
@blakeyrat said:
@Zecc said:Besides, did you just call a safety feature a pointless dialog?
No. The safety feature is the Recycle Bin. Which you're bypassing by holding shift. When there's absolutely no reason to.
Although maybe you're also entirely ignorant of how it works, so I guess there's that.
So tell us how ignorant we are. Explain to us how the Recycle Bin makes the "deleted" data magically disappear from the disk until you want them back.