Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
orf
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
58 ms
·
901.
▲
by
orf
4y ago
It's not possible to do with no translation penalty. Rust (and other languages like it) need the sizes of objects known at compile time in order to allocate space on the stack (and within struct fields). One way you could do it would b
902.
▲
by
orf
4y ago
It’s more correct to say that it always prefers zero-cost abstractions. Defaulting to dynamic dispatch would go very much against this, even if in some cases it’s not a bad idea.
903.
▲
by
orf
4y ago
Totally agree, and I do hope that at some point this will be standardised. Ideally you’d have an endpoint that returns a set of patterns to check for and a way to submit a set of tokens for checking, along with some kind of context. It woul
904.
▲
by
orf
4y ago
Indeed, so you’re building all this tooling and complexity and introducing more issues for the very small intersection of people that: 1. Are not malicious 2. Have access to a key 3. Are unable or unwilling to commit it to GitHub It would b
905.
▲
by
orf
4y ago
Not really: it depends on the permissions assigned to the keys. I wouldn’t like to wake up to an email that says “your key has been disabled because someone anonymously reported is as leaked, sorry if this has broken your entire system”. Wh
906.
▲
by
orf
4y ago
How would this work in practice? In order to effectively filter for tokens you need a set of pretty specific regular expressions. These specific expressions may contain non-public information about the structure of the tokens. But imagine w
907.
▲
Custom Allocators in Rust
(nical.github.io)
1 points
by
orf
4y ago
|
0 comments
908.
▲
by
orf
4y ago
I set up a project[1] to automatically leak AWS secrets published to the Python package index, which then triggers the secret scanning process and quarantines the keys[2] 1. https://github.com/pypi-data/pypi-aws-secrets
909.
▲
by
orf
4y ago
Cool motive, still pollution.
910.
▲
by
orf
4y ago
> There is no requirement to have a written contract with a tenant Sounds horrific > The IRS in fact never asks and does not know the identity of your tenants. Why is that a requirement to registering a contract? You’re just registeri
911.
▲
by
orf
4y ago
Because you have a contract, and that contract would be registered? And if you don’t have a contract, that’s a recipe for abuse of rights and tax evasion.
912.
▲
by
orf
4y ago
The tax authority doesn’t know if you’re a landlord? That hard to believe.
913.
▲
by
orf
4y ago
The original code doesn’t use a single transaction.
914.
▲
by
orf
4y ago
Not sure how juggling two phones, then having to manually copy addresses with potentially foreign characters between them is better. No, just let me view the damn map on my phone and the car display?
915.
▲
by
orf
4y ago
CarPlay is awesome, but I found the lack of being able to use maps very restricting. If you’re driving with someone who is co-piloting then them being able to use your phone to scout ahead for stops, restaurants etc is really useful. We end
916.
▲
by
orf
4y ago
Because they are not being fired?
917.
▲
by
orf
4y ago
This is a pretty shallow post complete with code samples as screenshots rather than text. The “expert level” technique uses an unbounded number of threads rather than utilising a thread pool. Using concurrent objects here also seems a waste
918.
▲
by
orf
4y ago
> I'd propose writing a query more appropriate for whatever part of the code contains the iteration. This is an identical solution to adding “select_related()” in the right places, and thus we loop back to the start of our thread.
919.
▲
by
orf
4y ago
I’m not sure that position lives up to reality. It’s easy to avoid N+1 queries in Django, there are lots of ways to reduce that to 1 joined query or 2 in the case of a m2m. The issue here isn’t that “orms are bad”, the issue here is that “k
920.
▲
by
orf
4y ago
Whilst shallowly funny the sarcasm actually shows a lack of understanding of the problem. The problem can be described as this: given an arbitrary point in a program, how can you infer what data is required at a future point without executi
921.
▲
by
orf
4y ago
There are some valid points buried in this comment, but they are deeper than you perhaps intended or even know. Metaclasses are widely used, but unless your specific problem calls for their use then you won’t use them. Usually these problem
922.
▲
by
orf
4y ago
With the exception of exception groups, these are all very simple and small additions. Does it mean Python is small? No. But that doesn’t make it big.
923.
▲
by
orf
4y ago
I started Python with 2.4 or 2.5. It’s not big, it’s been many years since I’ve been actually surprised about something. Evolution, or incrementing version numbers, don’t necessarily equal increased size. For example Python now has async fu
924.
▲
by
orf
4y ago
Fantastic story! Did the results get used in a claim?
925.
▲
by
orf
4y ago
We are not talking about purely US market share, because the argument being made is that the headcount is required to compete globally. Job board popularity varies greatly by country, but from the look of it LinkedIn has more of the global
926.
▲
by
orf
4y ago
Yes, about 25%-30%. And they do way more, in a lot of geos. Or are you saying LinkedIn has 14,000 employees just for their job board functionality, and only 6k for the other larger, more complex parts of their system? No? So what explains t
927.
▲
by
orf
4y ago
Yes I understand the rationale behind “lots of bodies”. As OP was rightly pointing out, there is a discrepancy here between output and size. LinkedIn for example has many of the same issues and a lot more complex functionality. They manage
928.
▲
by
orf
4y ago
Even with the 60 or so countries listed here ( https://www.indeed.com/m/countries ), that’s ~230 employees per country. Still seems way to high, especially given the uneven size of some of the job markets in those listed
929.
▲
by
orf
4y ago
A lot of people don’t look at your portfolio, or even with considerable verifiable open source work still ask you to do a test. I interviewed at one place that used my work and ran my code in production. Still had to do a take home test. Wh
930.
▲
by
orf
4y ago
It’s very hard to tell the difference between both your comments and conspiratorial ramblings (psychopaths have an unnaturally high alcohol tolerance?), hence the flagging and downvoting. It’s also hard to tell what your point actually is.
More ›