Why SQLite does not use Git (2018) • Hacker News Board

Re: Why SQLite does not use Git (2018)

#451

Post by tempodox » Thu, Jul 27, 2023, 4:05 PM UTC

willdr wrote: post #198

NikkiA wrote: post #109

Earlier quoted context omitted.

Actually it does, because with the windows git it includes it's own copy of ssh and bash, both of which will clash and fight with msys and/or other ssh installs - including the copy of ssh that microsoft themselves tuck away in \windows\system32 it's quite 'normal' for git-for-windows' ssh-agent to completely disable ssh-agent from working properly system-wide because it ends up pointing things at the wrong ssh-agent…

But a developer isn't using Windows.

Let's narrow that down and say they wouldn't voluntarily. But you obviously never heard of that adversarial entity named IT Department.

Re: Why SQLite does not use Git (2018)

#452

Post by Sprocklem » Thu, Jul 27, 2023, 6:22 PM UTC

tzs wrote: post #363

IshKebab wrote: post #342

Earlier quoted context omitted.

The Unix philosophy was a very rough guideline from the 70s, not a divine commandment that must be followed unwaveringly. There are a gazillion examples of excellent pieces of software that don't do "one thing" that can be piped as unstructured text into another program. Imagine editing a spreadsheet like `cat foo.xls | select-cell B3 | replace '=B2+1' > foo.xls`. Laughable. But how many people decry Excel for "not m…

> Imagine editing a spreadsheet like `cat foo.xls | select-cell B3 | replace '=B2+1' > foo.xls`. It would be even more cumbersome than that. After that command you'd have to restore foo.xls from a backup, and then do the edit again this time remembering that the "> foo.xls" executes before the pipe executes. :-) I wonder if anyone has written something to make pipes like that work? E.g., write two programs, "replace"…

`sponge` from moreutils more or less does this.

Re: Why SQLite does not use Git (2018)

#453

Post by Guvante » Fri, Jul 28, 2023, 3:20 PM UTC

tialaramex wrote: post #440

Guvante wrote: post #403

Earlier quoted context omitted.

I feel like you are being a bit unfair to Bjarne to to my knowledge isn't so much saying "we didn't make mistakes" as much as "we didn't make mistakes we should fix". C++ is the juggernaut it is because it just works. Your antient C++ codebase doesn't need many changes to get up to date on a language level. That doesn't mean it is painless to update toolchains but it does mean that the language isn't making that pain…

Bjarne is quite happy to admit that we made mistakes, but that wasn't my criticism was it? I'm saying that Bjarne doesn't like to admit that Bjarne made mistakes.

You can't be vague and make your point here. I explained a category of "mistakes" and you simply repeated your point.

To be clear my point wasn't about who made the mistake but that some categories of mistake aren't fixable. Oftentimes people conflate "don't change it" with "it wasn't a mistake".

Found an "I made a mistake" that is very explicitly self owned. Finding counterfactuals is nearly impossible so going to leave it at that without more specificity.

> Unified function call: The notational distinction between x.f(y) and f(x,y) comes from the flawed OO notion that there always is a single most important object for an operation. I made a mistake adopting that. It was a shallow understanding at the time (but extremely fashionable). Even then, I pointed to sqrt(2) and x+y as examples of problems caused by that view. With generic programming, the x.f(y) vs. f(x,y) distinction becomes a library design and usage issue (an inflexibility). With concepts, such problems get formalized. Again, the issues and solutions go back decades. Allowing virtual arguments for f(x,y,z) gives us multimethods.

Re: Why SQLite does not use Git (2018)

#454

Post by tialaramex » Sat, Jul 29, 2023, 4:29 PM UTC

Guvante wrote: post #453

tialaramex wrote: post #440

Earlier quoted context omitted.

Bjarne is quite happy to admit that we made mistakes, but that wasn't my criticism was it? I'm saying that Bjarne doesn't like to admit that Bjarne made mistakes.

You can't be vague and make your point here. I explained a category of "mistakes" and you simply repeated your point. To be clear my point wasn't about who made the mistake but that some categories of mistake aren't fixable. Oftentimes people conflate "don't change it" with "it wasn't a mistake". Found an "I made a mistake" that is very explicitly self owned. Finding counterfactuals is nearly impossible so going to l…

Yes, Bjarne wants Unified Function Call Syntax, he has wanted UFCS since at least the 1990s, years prior to C++ 98 but was not able to get WG21 to standardize it, not least because there are a bunch of tricky questions you need to answer to add UFCS and Bjarne's answers are... not the ones everybody else prefers.

The paragraph you're citing is from C++ proposal P1962, in 2019 - it's Bjarne basically arguing, as usual, that he should get his way and blaming the committee for the resulting problems. In that section Bjarne says that OK, he has just spent an entire paper telling people that they shouldn't have the certainty they do, but why defer to him, surely he can be wrong too? This is the sole example from that text where he admits to any mistake, over 25 years ago, and he blames everybody else for the mistake anyway saying it was "fashionable" to do this even though he pointed out the problems.

We should distinguish between problems which actually aren't fixable and problems which it was decided not to fix, I don't see a lot of the former in C++, but I do see a lot of the latter. And decisions can be changed.