rfoo

53 ms·

661.
by rfoo 4y ago
The post said that consumer connection does not let you do BGP peering with the ISP. There are real value in doing so in addition to the hack value, by doing so you can steer the Internet traffic to whatever IP blocks you owned to your hous
662.
by rfoo 4y ago
My preferred distro is Debian. Use linux-image-cloud; use a modern hypervisor with no legacy devices (only virtio etc); use tiny-initramfs; mask systemd-networkd-wait-online.service; add cryptomgr.notests noreplace-smp to kernel command lin
663.
by rfoo 4y ago
> and jump to it After CPU setup, memory training, dealing with ACPI shit, loading (and executing! [1]) Option ROM, and others? There are things called "BIOS" doing this before UEFI came into existence. [1] https://e
664.
by rfoo 4y ago
> I am curious about what scenarios would make one want to use systemd-boot (or grub) instead. I have Linux VMs which boots in ~2s (to ssh login; ~700ms to userspace). I use systemd-boot because it has nicer integration with my distro of
665.
by rfoo 4y ago
OK, so that's trying to stop us from being able to use an average USB cable to do anything but charge.
666.
by rfoo 4y ago
I don't know if it is possible where you live, but can you demand a refund without returning the device? That's how people in China countered fake 2TB thumb drives on taobao. They knowingly buy the device and ask for refund (bec
667.
by rfoo 4y ago
Ah, ok. I see where you are coming from :) From what I saw I'd say it's exactly opposite: allowing multiple versions actually means "make teams able to choose for stagnation". And because we are lazy, we certainly do! Th
668.
by rfoo 4y ago
But then another customer still wants gamma@1, they are allowed to do that! But they also want your new features. So now you have to maintain two branches, which I hope we can agree: it is an extra burden. This is unavoidable if we are talk
669.
by rfoo 4y ago
For my CUDA workload WSL2 (or really, Hyper-V GPU-PV) is about 75% native performance. Isn't too bad, I still prefer running non-virtualized Linux for anything longer than overnight. OTOH, run Linux on the metal and Windows in KVM work
670.
by rfoo 4y ago
> To start using beta@1 you need to upgrade alpha@1 to alpha@2 that depends on gamma@2. What's the problem with that? The problem is the team maintaining alpha does not want to upgrade to gamma@2 because it's an extra burden fo
671.
by rfoo 4y ago
How many internal libraries does your "separate services" contain? You service A depends on library alpha@1, your service B depends on library alpha@2. All happy now. Introduce another layer, your service A depends on library alph
672.
by rfoo 4y ago
parquet is just a data file format. Both pandas and DuckDB can query parquet files. It might be confusing that DuckDB does have its own columnar format, but it's more helpful to think it like "just a query engine (or library) for
673.
by rfoo 4y ago
Thanks. I'm not familiar with Java. I thought multiple classloaders are more like dlmopen (which doesn't help much - symbol visibility is hard) cause I saw people struggling on classpath conflict etc.
674.
by rfoo 4y ago
Yes.
675.
by rfoo 4y ago
There are languages / runtimes where there could not be two different versions of the same thing in one binary (and they eagerly fail at build time / immediately crash upon run). That is not the case for JavaScript, Rust, etc. But
676.
by rfoo 4y ago
The downside is if a product no longer have maintainers you are now encouraged to shut it down, even if it still works and it doesn't cost much to run.
677.
by rfoo 4y ago
For other libraries I simply YOLO-ed it and statically linked everything. Can't do the same for glibc. Yes, I ship security vulnerability to my users (and am responsible to fix it after every OpenSSL/curl/libpng/... advi
678.
by rfoo 4y ago
I maintained a sysroot with glibc 2.17 (it was 2.12 before when there are still CentOS 6 presence), built with latest GCC, for $DAYJOB. You can then use whatever modern toolchain you like (be it Clang+LLD or latest GCC+binutils), just point
679.
by rfoo 4y ago
The top 1 reason is that the developer build the binary with newer glibc than users'. And of course your users use older system than yours. Of course, if everyone builds their binary with glibc from 1997 it would just work. /s
680.
by rfoo 4y ago
Without going meta, how about ... the bomb just didn't work?
681.
by rfoo 4y ago
typo: Python was 0.5x instead of 0.7x
682.
by rfoo 4y ago
N3350 is from the Atom product line, which, IMO, is only suitable for embedded / thin clients. It does not tell much if an i7, even very dated, beats it.
683.
by rfoo 4y ago
It would be helpful if you can also bundle a dnsmasq-like (dnsmasq is not available for Windows :( ) thing for split-tunnel DNS. For example I would like *.work.example.com goes through the DNS in work VPN.
684.
by rfoo 4y ago
Yep, Nix not affected at all is pretty impressive. On the other hand this goes against the "verify before parse" principle so I have mixed feelings on Nix's approach.
685.
by rfoo 4y ago
> gpg seems to be what all of them use GPG signs a hash of the message with the private key, and you verify that the signature matches the file hash. Oh wait, what hash? :clown:
686.
by rfoo 4y ago
New git has a flag for keeping the old behavior, so it's not as bad.
687.
by rfoo 4y ago
If we make PathRef an alias of AsRef<Path>, the following code looks much prettier to me: pub fn read(path: PathRef) -> io::Result<Vec<u8>> { fn inner(path: &Path) -> io::Result<Vec<u8>> {
688.
by rfoo 4y ago
> Refuse to take jobs Well, I'm unfortunately not in a place where doing so makes sense. Unless you mean only auditing Rust code. > nag managers I already do so. This doesn't change much. There are still too many must-be-evo
689.
by rfoo 4y ago
Please tell me you have already thrown Firefox, Chrome, old Microsoft Edge and whatever browser out of window and are posting to HN with you rewritten-in-Rust lynx. Not being able to rewrite the world or convincing people to stop using memo
690.
by rfoo 4y ago
> (see also djb’s paper about qmail security) You mean the guy who refused to fix an integer overflow bug, claiming it isn't practical to exploit then 64-bit really happened then years later suddenly the fine guys at Qualys decide

More ›