kbolino

26 ms·

541.
by kbolino 2y ago
The hardware will not allow this, at least not without modifications. The encryption keys are not exportable from the Secure Enclave, not even to Apple's own servers.
542.
by kbolino 2y ago
Not exactly. It generates the keys for you and stores them on device in the Secure Enclave. You cannot "bring your own" encryption key, but the primary benefit of doing so--that Apple does not have access to it--is intentionally a
543.
by kbolino 2y ago
The keys are stored only in the Secure Enclave. Encryption and decryption are handled outside the standard CPU and OS. This is hardware-level protection, not just some flag on a cloud account to be flipped. The only way for Apple to break t
544.
by kbolino 2y ago
If you could retry it, then obviously the error was transient. I don't know what the issue is with me describing it that way. But you generally won't have to write retry logic if you have the file open in only one place. The other
545.
by kbolino 2y ago
Subpixel hinting isn't that useful at high DPIs though. Apple has ditched it entirely in macOS, regardless of monitor DPI, and gone back to standard anti-aliasing.
546.
by kbolino 2y ago
I daresay any Turing-complete language will be unable to prevent all memory leaks, since the simplest memory leak is just using more memory intentionally and never letting any of it go. Unless you want to eliminate resizable data structures
547.
by kbolino 2y ago
The smallest available (burstable) instance type on AWS (t2.nano) has 1 vCPU, 3 CPU credits/hour (~= 5% average CPU utilization), and 512 MB of RAM. I don't think you're going to run very much on there and definitely not with
548.
by kbolino 2y ago
They don't just get blocked, they error out. This is a well defined error condition, so you can detect it easily, but it's not going to happen if you stick to the one connection rule. This means you have to intentionally think abo
549.
by kbolino 2y ago
There are no data corruption risks in general for multiple processes on the same local filesystem, but the lock obtained by write operations is pretty aggressive. Multiple readers are okay but a single writer will cause transient errors in
550.
by kbolino 2y ago
The historical answer has generally been: it doesn't (at least, not well). The traditional recommendations were not to open multiple connections to the same file, and especially not to share the same file over a network filesystem. Thi
551.
by kbolino 2y ago
I don't know much about Aurora, there was/is a little too much magic there for my taste, but I feel like once we start with "Postgres- compatible DB", we can't necessarily reason about how things perform under the
552.
by kbolino 2y ago
I upvoted for the details, and I agree that if you try to buy comparable capacity in AWS, it's going to be more expensive. Scaling up in AWS is definitely going to cost more over time too. I don't want to hide these facts. But y
553.
by kbolino 2y ago
This is only half true. Yes, the country switched to first-to-file, but the definition of prior art was expanded . This means it's easier to strike down a patent due to prior art now than before.
554.
by kbolino 2y ago
Pretty much. I don't really see the problem though. You're also getting regular snapshots, which is yet another thing you have to build in that colo setup (and where are your backups going?). This is not for personal projects or s
555.
by kbolino 2y ago
Yeah, I think it should be emphasized that the framebuffer (in this sense) is actually an outdated abstraction and not the "low-level" or "raw" access to the hardware that one might naively expect it to be.
556.
by kbolino 2y ago
Notably, the DOS environment under Windows is running in a hardware mode called "virtual 8086 mode" which allows it to be isolated from the other DOS environments and, even more importantly, from full BIOS access. These limitation
557.
by kbolino 2y ago
For $200/month, I can have 2 ALBs, 2 ECS services, 2 CloudWatch log groups, and 2 RDS instances on AWS (one each for dev and prod) and a GitHub Team account with enough included runner minutes to cover most deployments. A colo is going
558.
by kbolino 2y ago
Windows 10 will go EOL in October, so you should probably try to upgrade sometime before next year. The first rule of PowerShell is to install the "real" PowerShell [1] instead of using the one bundled with Windows (which is out o
559.
by kbolino 2y ago
Whether accelerated UTF-8 decoding breaks RISC-ness is an interesting question, and not one with an obvious answer IMO. Supposing we loaded 4 bytes into a register with a load instruction (quite RISC), we could then have a "decode UTF-
560.
by kbolino 2y ago
The general idea is correct, but a rather significant detail is not. Android did not choose Java bytecode, it chose Dalvik bytecode [1]. This was done for several reasons, but at a high level, Dalvik bytecode more closely matches how real c
561.
by kbolino 2y ago
Good point. I misunderstood the situation. I figured doubling the VRAM size at the same bus width would halve the bandwidth. Instead, it appears entirely possible to double VRAM size (starting from current amounts) while keeping the bus wid
562.
by kbolino 2y ago
Prebuilders get priority access and volume discounts, so while it may not be a good value to buy individually, that doesn't necessarily apply to buying it in bulk.
563.
by kbolino 2y ago
The GPU already has DMA to system RAM. If you're going to make the VRAM as slow as system RAM, then a UMA makes more sense than throwing more memory chips on the GPU.
564.
by kbolino 2y ago
Yes, I'm taking the current situation on mobile and extrapolating it to be the future of desktop computing as well. Tablets and Chromebooks are already there. But Windows and macOS are headed in that direction too. Will there still exi
565.
by kbolino 2y ago
No, I don't think HTTP or JSON are cherry-picked. Depending on what kind of code you write, maybe you haven't needed them as often as I have. Of course, plenty of high-quality third-party libraries exist in the other languages. I&
566.
by kbolino 2y ago
Since any good solution would start with deleting rand and srand from the C standard library entirely, but that's never going to happen, we only have not-good solutions to consider.
567.
by kbolino 2y ago
Ok, yes, there were some algorithms in the past which had problems with carrying input biases to their output and sometimes even getting totally broken with certain seeds/keys. Nowadays, this isn't the case, and the view is that s
568.
by kbolino 2y ago
The problem with using the current time is that "guessing a few bits with confidence" is the same as guessing the whole thing because there's only a few bits that need to be guessed. You might think a time_t value has 55 and
569.
by kbolino 2y ago
> non-determinism in rand() if srand isn't ever called This is the path Go took for its math/rand library eventually: if you don't seed the RNG, it's a CSPRNG, but if you do seed it, it's deterministic. However,
570.
by kbolino 2y ago
The problem is not that OpenBSD doesn't follow the standard, it's that the standard is wrong. The default random number facility should be cryptographically secure. Most developers do not know the difference; giving them a determi

More ›