Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
Svetlitski
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
10 ms
·
61.
▲
by
Svetlitski
5y ago
I work at Facebook and am currently working on implementing the foundation to support exactly this. https://github.com/osandov/drgn/pull/129
62.
▲
by
Svetlitski
5y ago
That they claim as one of their benefits that Intel chips have better security than ARM is just laughable in light of the myriad speculative execution vulnerabilities present in x86 chips.
63.
▲
by
Svetlitski
5y ago
:h was new to me, I have always used ${path##*/} to achieve the same thing.
64.
▲
C++ Frequently Questioned Answers
(yosefk.com)
4 points
by
Svetlitski
5y ago
|
0 comments
65.
▲
by
Svetlitski
5y ago
His Master’s Voice by Stanisław Lem is mind-blowing, it’s a slow burn, but I’d highly recommend it.
66.
▲
by
Svetlitski
5y ago
On this topic, something that drives me nuts about Python packaging is that the name of the top-level import often doesn’t match the name of the package. For example if you see `import yaml` in a Python script, you might try to run `pip3 in
67.
▲
by
Svetlitski
5y ago
By far the most frustrating password-handling sin I've encountered is when a website silently truncates the password field to some fixed length. This is especially bad when you use a password-manager, as the password-manager will r
68.
▲
Missouri Digital News
(mdn.org)
1 points
by
Svetlitski
5y ago
|
0 comments
69.
▲
by
Svetlitski
5y ago
All of this is addressed in the FAQ about the story: https://www.ibiblio.org/harris/500milemail-faq.html
70.
▲
by
Svetlitski
5y ago
Here's my prediction for the future of the web: I think one of two possible scenarios will occur 1. Firefox fades into obscurity/irrelevance, Chromium becomes the platform of the web, and thus Google dictates its future. 2. Fire
71.
▲
by
Svetlitski
5y ago
Killer specs, better keyboard (no touchbar), more ports, and yet the most frequent comment here is about the notch. HN is a tough crowd to say the least, in this instance IMO to the point of it being comical.
72.
▲
by
Svetlitski
5y ago
The specs on the new MacBook Pro are incredible, especially considering people were already blown away by the performance of the first M1 MacBook. It’s a seriously powerful machine. And the additional ports are a very welcome change.
73.
▲
by
Svetlitski
5y ago
Upon seeing this, I can’t help but think of “So What’s Wrong with 1975 Programming” from the author of Varnish [1]. [1] https://varnish-cache.org/docs/trunk/phk/notes.html
74.
▲
by
Svetlitski
5y ago
Are you certain your benchmarks are correct? The GoAccess FAQ states that it parses over 100,000 lines/second [1]. While this figure depends on the hardware used, this still is massively faster than the figure quoted in the README. B
75.
▲
by
Svetlitski
5y ago
You should consider looking into Gradescope (Gradescope.com). As a former TA, I can attest to it making grading much more pleasant and streamlined than it would be otherwise.
76.
▲
Space City Weather
(spacecityweather.com)
2 points
by
Svetlitski
5y ago
|
0 comments
77.
▲
by
Svetlitski
5y ago
I’m surprised Stack Overflow doesn’t have a CSP that would prevent this.
78.
▲
by
Svetlitski
5y ago
For what it’s worth, you can sign-up through DMAChoice to stop receiving junk mail [1]. It takes a few months after you sign up to really take effect, but it has easily eliminated 90% of the mail I used to receive. Well worth it IMO for the
79.
▲
by
Svetlitski
5y ago
I'll kick things off. About a month-and-a-half ago I posted a "Show HN" for a nifty CLI tool I built called fcp [1], a significantly faster alternative to the classic Unix cp(1) command, copying large files and directories in
80.
▲
Ask HN: Share your side project
38 points
by
Svetlitski
5y ago
|
68 comments
81.
▲
by
Svetlitski
5y ago
This very much sounds like a typical Relational Database (to a degree), with the language thus being SQL.
82.
▲
A gentle introduction to jump threading optimizations
(developers.redhat.com)
2 points
by
Svetlitski
5y ago
|
0 comments
83.
▲
by
Svetlitski
5y ago
Genuinely curious, what didn’t you like Homebrew (assuming you tried it)?
84.
▲
by
Svetlitski
5y ago
Definitely agree. Another example of lost performance that comes to mind is (ignorance of) C struct padding. In terms of modern tools, Rust has definitely made some headway here: • Variables are immutable by default • The layout of a struct
85.
▲
A little-known Soviet mission to rescue a dead space station (2014)
(arstechnica.com)
73 points
by
Svetlitski
5y ago
|
21 comments
86.
▲
by
Svetlitski
5y ago
You can set the attribute autocomplete=“off” on the corresponding <input> in order to prevent this from happening. See the relevant MDN page for more details: https://developer.mozilla.org/en-US/docs/Web/
87.
▲
by
Svetlitski
5y ago
Sweet, cmp::min_by_key and cmp::max_by_key have finally been stabilized. Was sorely missing this little convenience in Rust, as it helps express a particular idea more clearly and succinctly. Also the improvement to compile times on macOS d
88.
▲
by
Svetlitski
5y ago
I spun up another instance, and now know that the filesystem used on the Linux benchmarks is xfs.
89.
▲
by
Svetlitski
5y ago
No, it is not. I would've submitted a Homebrew formula for it, but that would go directly against Homebrew's guidelines, which frown upon authors submitting formulas for their own tools [1]. [1] https://docs.brew.sh
90.
▲
by
Svetlitski
5y ago
Definitely valid points, as I mentioned in a comment below I'll be updating the README to mention this is geared towards systems with SSDs. As for lacking most of cp(1)'s options, that's a deliberate choice. My goal with fcp
More ›