GrantMoyer

9 ms·

61.
by GrantMoyer 2y ago
I also think a lot of speedrunnimg tecniques demontrate the "anything can happen" nature of undefined behavior in an viscerally, not-purely-theoretical way. What happens when you don't take undefined behavior seriously? Well,
62.
by GrantMoyer 2y ago
The thing is there's not much nuance to be had. Netflix's approach is just worse than existing best approaches on all axes.
63.
by GrantMoyer 2y ago
Each update from Mozilla about this issue has Mozilla claiming users are confused (which may be true; I don't follow the larger social media ecosystem), then doubling down on the part I'm personally concerned about. I'm worri
64.
by GrantMoyer 2y ago
More like Bic needs unrestricted access to your letters so their pen can write them for you.
65.
by GrantMoyer 2y ago
Sure, but neither Distribution Policy for Mozilla Software[1] nor Mozilla Trademark Guidelines[2] make any mention of the Terms of Use. Services Firefox connects to have their own Terms of Service, like you say, but those are unrealated to
66.
by GrantMoyer 2y ago
The term itself is bad here. The problem isn't just because it's hard to understand. > Second clause establishes that when you input that data you give firefox a license to that data which you otherwise own Firefox doesn't
67.
by GrantMoyer 2y ago
Firefox's blog post in the change[1] has an update: > UPDATE: We’ve seen a little confusion about the language regarding licenses, so we want to clear that up. We need a license to allow us to make some of the basic functionality of
68.
by GrantMoyer 2y ago
Firefox's own about:license page (reachable through the about firefox dialog) says the sources are available under a wide variety of open source licenses. Does that mean only Mozilla distributed binaries are governed by the "Firef
69.
by GrantMoyer 2y ago
> If it was restricted to bubbling the maximum or minimum to the end of the list, the full sort would use one bubble, it would complete in linear time, and the list wouldn't be sorted afterwards. Yes, one bubble op is linear time. A
70.
by GrantMoyer 2y ago
Last time I checked, while tutanota's emails are ostensibly E2E encrypted, all public keys are provided by their server and there's no way to pin keys or verify them over a side channel, so a compromised server could trivially sen
71.
by GrantMoyer 2y ago
Your step 3 is an insertion op, not a bubble op. An insertion op inserts one element into a sorted list. A bubble op bubbles the maximum (or minimum) element to the end of an unsorted list.
72.
by GrantMoyer 2y ago
Interaction nets: https://dl.acm.org/doi/pdf/10.1145/96709.96718
73.
by GrantMoyer 2y ago
> Surely there must be a middle ground between allowing free abortion and having women die of sepsis. You're absolutely right, and we know so, because that middle ground actually existested before Roe v. Wade was overturned. Abort
74.
by GrantMoyer 2y ago
Actually, the opposite is true. The rationals have measure zero, so a number selected uniformly at random from 0 to 1 (formalized way of placing a dot on a number line), has a 0% chance of being rational.
75.
by GrantMoyer 2y ago
One important aspect (to me, at least) that the article doesn't mention is that physical media comes with actual ownership rights (first-sale doctrine), and that by whatever lucky coincidence publishers weren't able to introduce w
76.
by GrantMoyer 2y ago
The linked article doesn't say the submitted driver code was awful. In fact, it says Paragon submitted the driver after Linus suggested they submit it. What the article quotes Linus complaining about is a process issue. Paragon apparen
77.
by GrantMoyer 2y ago
In addition to being very technically interesting, I think this is among the best executions of a "humorously flippant" tone I've read. I particularly enjoyed: > They’ll simply describe an interactive protocol with the rig
78.
by GrantMoyer 2y ago
C++ has had a few options since C++11[1]. - Linear congruential generator[2]: xₜ = a xₜ₋₁ + c mod m - Mersenne Twister[3]: too complicated for a few lines - Subtract with carry[3]: k < p carryₜ = [xₜ₋ₖ - xₜ₋ₚ - carryₜ
79.
by GrantMoyer 2y ago
Note that you need to use 3 shifts. All the triplets of shifts listed in the paper should generate sequences with period 2^32-1 or 2^64-1 depending on the state size, because the tables were generated with that constraint. For example the f
80.
by GrantMoyer 2y ago
I've hit issues with bad quality PRNGs when using them for Monte Carlo path tracing. With a non-uniform PRNG, the simulation can take a lot longer to converge to a clear image, because some ray paths can be severely under or over repre
81.
by GrantMoyer 2y ago
Xorshift is really cool because of how simple it is to comprehend and implement despite how well it works as a PRNG. The initial paper describing it, "Xorshift RNGs"[1], is surprisingly short and approachable, and I encourage anyo
82.
by GrantMoyer 2y ago
Seems like a great time to stop eating eggs.
83.
by GrantMoyer 2y ago
DMCA and Trademarks
84.
by GrantMoyer 2y ago
Follow up to my earlier comment; The docs say creating a reference to uninitialized memory is undefined behavior, but I couldn't think of any cases where it would go wrong if the data is never actually accessed. It turns out that'
85.
by GrantMoyer 2y ago
1. Yep, this is an abstraction HW works hard to preserve. 2.The specific scenario I had in mind was: write value X to address A from core 1, write value Y to A from core 2, read value from address A from core 1 (still X in core 1's cac
86.
by GrantMoyer 2y ago
In Rust, creating a reference to uninitialized data is undefined behavior, even if you don't access the data through it[1]. The optimizer assumes all references are intialized. [1]: https://doc.rust-lang.org/std/pt
87.
by GrantMoyer 2y ago
The issue isn't writes to uninitialized memory, it's reads from uninitialzed memory. The compiler doesn't know how much of the buffer `read` writes. The docs say it returns a unsigned integer with how many bytes it wrote, so
88.
by GrantMoyer 2y ago
> That's correct at a hardware level. Not even there, due to memory caches.
89.
by GrantMoyer 2y ago
Okay, I think I see. In the linked article, due to the behavior of `read` uninitialized memory is never read. The same would be true in equivalent C code. However, in C, the programmer doesn't need to prove to the compiler that uniniti
90.
by GrantMoyer 2y ago
You have some misconceptions about C and undefined behavior. > C code that reads uninitialized data is presumed to be buggy, because it wouldn't be doing that unless it thought the memory was initialized. But the read itself is not

More ›