ninjaoxygen

People who grew up with assembly language or locking primitives are fairly likely to use compare-and-swap because it's literally the commonly understood term for it.

https://en.wikipedia.org/wiki/Compare-and-swap

At one point it was definitely not so deep... carriers were literally looking at the IP TTL and seeing whether it was a recognised value from the phone or a few hops less than one of the common defaults, in which case it was considered tethering traffic.

You could spoof it by finding out your mobile's TTL, overriding the TTL in the connecting device to be one higher than the mobile.

Great to see you posting on HN! I still remember making the decision to replace Crunchy Data PGO. CNPG was newer and smaller at the time, so maybe riskier, but seeing the way your team had responded to issues, had a real engineering mindset and clear knowledge of how PostgreSQL should be operated made it an easy choice!

Thank you for your efforts, I hope the balance of commercial and community works out well for you, the product is great.

Totally agree with you there, as much as I love to hate non-transferability, revokable licenses, permanent VAC bans on accounts that got hacked, I still find Steam the most convenient path to "owning" games in one place.

The Linux work done for Steam Deck is fantastic and I do credit their efforts with inspiring others to work on similar projects that extend and complement what Valve achieved. Much of the hard effort did go into Windows games on Linux before Valve looked at it; everything the WINE project, Codeweavers did, gaming via Lutris since 2009, however Valve have definitely been a force multiplier.

Trust is earned and I think Valve are doing pretty well on that front, especially when you look at the differences to other PC stores, Ubisoft, EA, and to some extent Epic. GOG and Itch are very different beasts.

To some extent I miss the time where Steam was totally curated, you had to make an impact to get your game on the platform, back before it was a free-for-all of shovelware and low-effort slop. Occasional controversies aside, at least on Steam the tools / marketing funnel are there to keep the popular games at the forefront of the store whilst also being fairly open to allow devs to publish without being the chosen one.

Is there a danger of doing to games what Spotify has done to music? Maybe, but I reckon the super deep-discount sales have calmed somewhat and are happening later in game's long-tail part of the lifecycle or used as promo for sequels.

There are plenty of publishers that choose to mainly avoid going that route, often the traditional established publishers with console outlets they don't want to cannibalise, for example Sony and Konami.

We use Barman inside Kubernetes via CloudNativePG's plugin, as it is the default backup plugin.

Barman has always been solid for backup and restore, however configuring backup in CNPG is a little more interesting - WAL limits need to be set carefully or you just end up filling WAL volumes and the database becoming unavailable.

I remember switching from Win95 to NT4.0 just to be able to use SoftICE properly under Windows without all the stability problems, it was an incredible time! SoftICE felt like absolute wizardry at the time.

Did you use SELECT FOR UPDATE at all, or just never had to update dependent data? If the complex operations are implemented using stored functions / procedures then the a transaction is implicit.

If the data is fairly straightforward like just one-to-many CRUD with no circular references, you would be able to do it without transactions, just table relationships would be enough to ensure consistency.

Respect to Bose for taking a better stance on this. I still remember Sonos implementing "Recycle Mode" that they thankfully backtracked on.

Yes, you're right, the last sentence is definitely a mistake on my part, I should have written less! Thanks for the links, paulirwin's sibling response is helpful too.

We had code using WCF and AppDomains that were always out of scope for .NET Core. WCF has a Core replacement now that is not quite one-for-one but AppDomains will never be supported in .NET Core / .NET 5.0+ and would indeed have to stay on 4.8 / 4.8.1 if they were still running.

Yes, you are right, if you are on 5.0+, however the 4.x stuff is definitely out of support.

Sorry, I did not know they had actually brought non-Core ASP.NET forward into 5.0+, but it makes sense given how much of .NET Framework they continued support for and how much ASP.NET and Forms stuff is still around in enterprise with no budget for bringing it forward.

Totally agree with breaking the chain though, we moved to Core around 2.0 and never looked back, as an ecosystem it is so much better.

There is no .NET Core or .NET Framework since .NET 5.0 in 2020. Maybe you mean ASP.NET Core, but then there is no ASP.NET Framework so the comment still does not make sense to me.

The vulnerable component is ASP.NET Core, which did not change name when .NET dropped the Core name to distinguish it from legacy ASP.NET.

--- edit: cut here - the sentence below is incorrect! ---

If somehow you were still using legacy ASP.NET / Framework 4.8 etc, you have much bigger problems - legacy ASP.NET has been unsupported since 2022 so will definitely not be receiving security updates.

Ignoring CR is often how two systems end up parsing the same file differently, one as two lines the other as a single line.

If the format is not sensitive to additional empty lines then converting them all CR to LF in-place is likely a safer approach, or a tokenizer that coalesces all sequential CR/LF characters into a single EOL token.

I write a lot of software that parses control protocols, the differences between the firmware from a single manufacturer on different devices is astonishing! I find it shocking the number that actually have no delimiters or packet length.

At the very least Windows Update can offer you two versions, a stable version of the driver and a newer but potentially less stable version through the "Optional Driver Updates" page in Windows Update.

True, it has never been in an HTML standard, however it was definitely a documented part of early HTML.

The blink element was in Netscape Navigator's HTML dialect in 1993/94, when early HTML was still just hitting IETF RFCs / DRAFTs, you can find blink in the Netscape HTML developer documentation from just after that era, DevEdge. It was never in NCSA Mosaic, the other big GUI browser of the era.

Later on in the process of being standardized, when it was more W3C than IETF albeit still mainly the same people, Netscape agreed to drop blink from the proposals if Microsoft dropped marquee, so in that sense yes, it was never in a standardized version of HTML, but many tags in active use at the time were never in a standards doc.

See here https://www.w3.org/People/Raggett/book4/ch02.html for some history from w3c, who went on to become the formal custodians of HTML after the IETF days.

Edit: here's the earliest Netscape Developer Docs I can see on archive.org https://web.archive.org/web/19961115043739/http://developer....

The act of turning AIS off can attract unwanted attention (higher resolution local satellite monitoring), less likely if you are entering waters where piracy is common and many vessels disable AIS.

If a vessel turns AIS off then cuts the cable, but their position is known by other means, they will be giving up plausible deniability.

It may be an unpopular opinion, but I suspect treating juveniles with kid gloves all the time and the popularization of the view that they are entitled not be exposed to anything that may upset or offend them has lead to less personal resilience.

The social media environment does also mean that when subjected to bullying it is amplified by a massive factor that previous generations were not exposed to, so the change in expectations is certainly not the only factor at play here.

As a user of Keycloak on a production project, I'm a little sad there is currently no support for opaque tokens. Sure, you can treat the access token as an opaque token... but at the end of the day it could be a lot smaller.

Discussed here https://github.com/keycloak/keycloak/discussions/9713 and https://stackoverflow.com/questions/75082532/keycloak-suppor...

We also experience a few front-end issues, like when a token expires, the browser tab goes back to the login page. If you leave the tab a while then press login, the token it is using will have expired. Rather than automatically retrieving a new token and posting the login again, the user gets an error message and has to authenticate again.

If you have two tabs in that state, you log one back in, switch to the other tab, if you refresh that tab, all is well, login proceeds automatically. If you press "login" instead, you get an error page telling you "already logged in" rather than just redirecting you back to the app... it also loses the redirect url so you have to press "back" instead.

Will see if we can fix these when we have time, it would be nice to contribute back.

gpg supports using public / private keypairs to encrypt any amount of data you like. I use it for uni-directional backups from machines where trust is an issue.

Or is the reality of this that it's just encrypting a symmetric key with the asymmetric cipher, and then encrypting data using that key?

Funnily enough based off Lenovo's reputation for their higher end laptops, in April I bought a Lenovo X1 Carbon Gen10, pretty much as they came out. The machine is a nightmare, and has been through about 10 firmware updates over the last 6 months, but has barely improved.

From day one and still now... on Win11 that it shipped with it won't hibernate, just a black screen on resume, then pot luck over the next boots whether it starts or not. Sometimes from shutdown it does the same thing. Under little load, it gets hot as hell, you literally can't type on it, I use a bluetooth keyboard. If I'm doing light dev work, 3h battery life, tops. Audio drivers are broken, the internal 3.5mm just emits noise above around 10% volume and is distorted below that. Shipped with an underpowered 45W adapter... same SKU now ships with a 65W (an i7-1260p w/32GB).

I don't think it's faulty, there are plenty of people with the same story about Gen10s. Mine is in for repair now, so we'll see, if it comes back with the same issues, it's getting warranty returned.

I know the X1 Carbon Gen9 is a really good machine, maybe Lenovo got screwed by Intel's 12th Gen platform and Microsoft's support for the processor, but the machine should not be going out the door like that.

Come to think of it... my XPS15 9560 from 2017 is STILL a better machine, which is what I switched back to while it's in for service (and cost half as much!)

If you want to use it, the firmware will need to be updated, and therefore quickly no longer vulnerable once Sony patches it.

If you wait until later, the units in the shops will also have patched firmware.

I remember having NT3.51 and NT4.0 checked build CDs from MSDN.

I think they also came with full symbol definitions, it was an absolute treasure-trove of information about the OS.

What scares the hell out of me, it would be so easy to backdoor an implementation of the open core in a well-hidden way.

Whereas the "big" CPU providers are staking their reputation and therefore future business on providing a non-backdoored CPU, it would be fairly trivial for an individual device manufacturer to provide a backdoored CPU design for their chip design.

It could become the whole cheap-device OEM firmware situation all over again (as we saw with many backdoored routers), but this time the blob is located on-die, so is significantly harder to reverse engineer or audit.

My only annoyance with GitLab on that front is the lack of tags for "14.8-ce" to enable use of watchtower to keep the minor/security patch up to date, like we have on other images.

I definitely do not want to use "latest", in the past there have been updates or intervention required between major versions.

I will be interested to see how HTTP/3 fares on Virgin in the UK. I believe the Superhub 3 and Superhub 4 have UDP offload issues, so it's all dealt with by the CPU, meaning throughput is severely limited compared to the linespeed.

Waiting for Superhub 5 to be officially rolled out before upgrading here!

Out of interest, do you have Kubernetes enabled in Docker for Desktop?

I had regular problems very like yours, then one day it just stopped happening, suddenly WSL2 Docker Desktop worked fine.

I'm suspecting it might be around the time I turned Kubernetes off again?

We also remained on Yarn 1 because developing packages using the workspaces feature is really nice.

We did make an attempt at moving to Yarn 2, it did not work out well. At that time (probably 12 months ago or more) lots of upstream packages had broken package.json files, we never managed to fix them all. Lots of things with React were a bit screwy too!

I work in high-end AV. Perceptually, in a good space, those fantastic speakers often do not make the sound people associate with "loud", no distortion, no top end becoming hissy, just the concert-level bass to clue you in to how loud it is. People who have paid a lot for their system want that "wow" factor that immediately makes anyone think it's loud. The other issue is that different listening material definitely needs different amplifier and processor settings - there is no setting that "just works". We find many customers do not wish to get engage with those settings these days.

Wow, you triggered a real flood of memories for me. I remember being bored in in an unrelated training course and using debug.com to write 320x240 scrollers with rep movsb. Being a kid I had no internet access at the time so a copy of Ralf Brown's Interrupt List and Borland's Turbo C++ documentation was pretty much all the documentation I had. Then later discovering djgpp and protected mode, what a time to be alive!

I really want to let my kids (7 / 11) have a simple programming environment where they can get stuff done quickly. They have struggled a little with Python so far (it always comes down to needing complex libraries to do what they want e.g. write simple games.) Anyone have suggestions for an alternative language to let them be fairly self-starting?

I believe the point he is making is that the games can run in the native Windows above Hyper-V and then WSL2 / other Linux installs can be run under Hyper-V from the native Windows OS.

In the stock trading world, HFT on FPGA is closer to the edge than GPU / TPU solutions and they are using ML models, if you count that as AI. When the logic and the NIC are on the same hardware, it's really fast. An ASIC would be even faster, but you can't really iterate on that.