WTF Bites
-
@Mason_Wheeler said in WTF Bites:
the first sign that something is fundamentally wrong with this picture...
What's wrong with this picture?
-
i'm not the complaints department so.....
You're responsive though.
Kyon!
responsive i may be, but I'm not always responsive in a helpful manner!
-
i'm not the complaints department so.....
You're responsive though.
Kyon!
responsive i may be, but I'm not always responsive in a helpful manner!
Oh, being responsive is almost always helpful...
-
… how do they do the memory management? As far as I can tell the garbage collector (used by JS) was not integrated with WASM yet.
The JS setup code hands in a fixed-length byte array for a stack and a second fixed-length byte array for a heap. The Mono interpreter (as compiled to WASM) includes its own garbage collector which operates on those arrays. The setup code needs to determine the maximum possible heap usage and that amount of memory will always be in use regardless of how much .NET frees, but given today's browsers that's more of a cake problem.
-
@TwelveBaud said in WTF Bites:
but given today's browsers that's more of a cake problem.
They'll go the distance (in using as much memory as they can) but they don't go for speed
-
I don't know about y'all but i really don't consider a language to be a language until it:
I consider a language to be a language when you can write in it and it runs with reasonable efficiency.
Turing complete is Turing complete. Machine code is still transformed and "interpreted" by the CPU but we don't care.
-
@anonymous234 said in WTF Bites:
I consider a language to be a language when you can write in it and it runs with reasonable efficiency.
IOW, you don't consider JavaScript a language
-
In a somewhat related WTF: apparently the
Microsoft.AspNetCore.AppandMicrosoft.AspNetCore.Allpackages have special versioning and deployment rules in .NET Core projects. Because god forbid Microsoft just build a thing on another thing without introducing a bunch of special rules for themselves .Incidentally I just tried "preprocessing" a 12-line ASP.NET Core .csproj file (
dotnet msbuild -pp:output.xml .\RazorPagesProject.csproj). to see what MSBuild actually does and what packages it includes. Well, the output file has 14 424 lines. So yeah.
-
@anonymous234 said in WTF Bites:
I don't know about y'all but i really don't consider a language to be a language until it:
I consider a language to be a language when you can write in it and it runs with reasonable efficiency.
A (programming) language is a language if you can describe a program to a computer using it. Efficiency is not formally required. Types are not required. Compilation is not required. Readability is not required (but God help you if you neglect that!)
Bytecodes are languages, but not ones intended for ever being read or written by humans.
-
@levicki said in WTF Bites:
@Vixen What is a deaf leopard doing in a fox's basement?
don't ask me. I gave them an eviction notice.... and when they were still there I got a court order to remove them... but they countered by saying my basement is the Soverign Nation of the dark side of the moon and they will not be forced off their ancestral homeland....
at least i think that's what they say. the police officer that went down to enforce the order came back up VERY drunk and high on..... something.... so i'm not exactly sure of exactly what words he was trying to say.
-
@levicki said in WTF Bites:
@Vixen Are you saying he purred unidentified drinks down that policeman's throat? And the policeman's record of the events was spotty?
¯\_(ツ)_/¯
-
@levicki said in WTF Bites:
What is a deaf leopard doing in a fox's basement?
Whatever it's doing, it fortunately won't hear Def Leppard.
-
Got an additional drive for my machine, one of those fancy NVMe things. Figured I'd move some stuff from my overfull SSD to the new drive. Figured one of those things might be Visual Studio, because (a) it's huge and (b) it might actually benefit from the additional performance (though, C++ builds are more likely CPU bound).
First: there's no way to move an existing install, recommendation is to uninstall the damn thing and reinstall it. (Or mess with junctions. Maybe that would have been the sane choice).
Second: I tell the installer to put stuff on the new drive. Out of the 30GB of shit, 12GB end up on the new drive, and 18GB end up on the system drive anyway. Thanks for nothing.
Windows is fucking retarded when it comes to this shit. (Moving home folder? Nah - perhaps some select subfolders. Moving applications? Don't think so. At least Steam apparently has an option for its stuff.)
-
Moving applications? Don't think so.
You can do this in the Apps control panel depending on what type of app it is.
At least Steam apparently has an option for its stuff
Yes Steam makes it easy.
-
@loopback0 said in WTF Bites:
You can do this in the Apps control panel depending on what type of app it is.
Some google search results claimed the same thing. I didn't check all the applications, but for the ones that I did, either the "Move" option is grayed out, or it's replaced by "Modify" (again, moving isn't an option after selecting that).
Perhaps it works for "store apps", but the only ones that I have are the ones that ship with Windows already (and frankly, "Photos" or whatever can stay on the old SSD -- moving that isn't going to make any difference).
-
First: there's no way to move an existing install, recommendation is to uninstall the damn thing and reinstall it.
Well yeah. Do you know how hard it would be for Windows to do that? Applications don't really have a "place". I mean, they have a place (Program Files), but each one handles the installation, uninstallation and other logic by itself and they're bound to do it wrong, and it's going to break all paths.
If they had added just a tiny little bit of abstraction/encapsulation so the app didn't know where its files are located Windows would be 10,000x better (give or take a couple zeros).
-
@anonymous234 said in WTF Bites:
Well yeah. Do you know how hard it would be for Windows to do that? Applications don't really have a "place". I mean, they have a place (Program Files), but each one handles the installation, uninstallation and other logic by itself and they're bound to do it wrong, and it's going to break all paths.
I would have been OK with copy-pasting it over to a different location and then using a junction/symlink to make it believe it's still in the old place. But there are quite a few recommendations against doing so, claiming random breakage.
Somewhere there was a suggestion for doing so with Visual Studio's "Shared" folder, which is apparently where the majority of the can't-go-in-a-custom-path stuff ends up *edit:* I figured I might as well check, and it's not. However, some follow up comments claimed that this might even end up interfering with Windows Update (for whatever reason?).
-
However, some follow up comments claimed that this might even end up interfering with Windows Update (for whatever reason?).
Because Windows Update breaks when you look at it funny. Duh.
On my last Windows laptop, having a dual boot set-up prevented it from installing some updates.
(I think the underlying reasons was that the language packs, which change the text of about one trillion components, couldn't change the about two sentences in the boot loader)
-
Second: I tell the installer to put stuff on the new drive. Out of the 30GB of shit, 12GB end up on the new drive, and 18GB end up on the system drive anyway. Thanks for nothing.
Visual Studio has a whole bunch of different installation paths (app, SDK, Python, reference assemblies, Dotfuscator, etc). Up until 2015, you had to change each of these individually; changing the path for the Visual Studio IDE component doesn't change the path for the Windows SDK component. With the new 2017/2019 installer, which lets you change but a single path, I don't know how to make that happen...
-
@anonymous234 said in WTF Bites:
If they had added just a tiny little bit of abstraction/encapsulation so the app didn't know where its files are located Windows would be 10,000x better (give or take a couple zeros).
In other words, Store apps...
-
-
@anonymous234 said in WTF Bites:
If they had added just a tiny little bit of abstraction/encapsulation so the app didn't know where its files are located Windows would be 10,000x better (give or take a couple zeros).
The Alternate Filesystems threads are
and 
-
Anyone taking bets on what sort of problems we'll see in 10 years time?
-
Not sure if it's been posted already, but... What?
Boeing Employees Mocked FAA In Internal Messages Before 737 Max Disasters
The documents include emails and internal communications. In one message, employees mock the Federal Aviation Administration and brag about getting regulators to approve the jets without requiring much additional pilot training.
In another document, an employee ridicules colleagues involved in the development of the troubled plane, saying, "This airplane is designed by clowns who in turn are supervised by monkeys."
-
responsive i may be, but I'm not always responsive in a helpful manner!
— Modern web in nutshell.
-
Moving home folder? Nah - perhaps some select subfolders.
The fun thing is that you can define an alternate location for the
Usersfolder, but only during Windows (re)installation. And even then you want to be careful: https://www.tenforums.com/tutorials/1964-move-users-folder-location-windows-10-a.html
-
@dkf
You assume they extended past 2021? Brave man...
-
-
@TwelveBaud said in WTF Bites:
but given today's browsers that's more of a cake problem.
They'll go the distance (in using as much memory as they can) but they don't go for speed
-
@TwelveBaud said in WTF Bites:
a second fixed-length byte array for a heap
That's retarded. Not even actual native applications work like that - heap is always handed in by the OS in chunks on demand.
-
@Tsaukpaetra said in The Official Status Thread:
quite nicely...
Oh, reading the box for shits. I don't think we have a proper engrish thread so here.
They want you to scan the code to discern truth, but too much of the code is hidden to actually do so...
What trollery!
Edit: oh, it's a scratch-off thing. The code itself redirects to a non-working link. Looks like a QQ authorization redirect?
I'm not sure I want to discern their truth anyways....
-
@levicki said in WTF Bites:
@Tsaukpaetra said in WTF Bites:
I'm not sure I want to discern their truth anyways....
You know you do... you always get to the bottom of every dead horse.
... Shut up! :tsundere.png:
-
@TwelveBaud said in WTF Bites:
a second fixed-length byte array for a heap
That's retarded. Not even actual native applications work like that - heap is always handed in by the OS in chunks on demand.
...Can someone explain why they gave a downvote here? Who knows, maybe I'll learn something new.
-
@Tsaukpaetra said in WTF Bites:
@levicki said in WTF Bites:
@Tsaukpaetra said in WTF Bites:
I'm not sure I want to discern their truth anyways....
You know you do... you always get to the bottom of every dead horse.
... Shut up! :tsundere.png:
It's actually a redirect to a wechat... something.
That tries to redirect you to weixin12315 (a "big data" company?), which sends an SNS message and redirects (or something?) again to f315 . cc. Never actually invoked it so I don't know what else is intended to happen.
-
Don’t you have a Do Not Call list?
Doesn't do any good when the bad guys simply ignore the list. And supply fake caller id so complaining does no good.
Exactly.
So all telemarketing is basically fraud because legitimate telemarketing has almost nobody left to call? Or do people just not bother to get put on the list?
Legitimate telemarketing is an oxymoron.
If you have a legitimate product to sell you don't need hundreds of people in call centers, hiding their identity behind fake phone numbers. You spend that money advertising your product.
-
@anonymous234 said in WTF Bites:
Incidentally I just tried "preprocessing" a 12-line ASP.NET Core .csproj file (
dotnet msbuild -pp:output.xml .\RazorPagesProject.csproj). to see what MSBuild actually does and what packages it includes. Well, the output file has 14 424 lines. So yeah.18,000 API functions is considered "very minimal" by the .NET Core guys. Any more questions?
-
-
-
Strange. All the LEDs on RAM and other components actually start making sense at that point.
-
-
I'm happy today that I don't have an account with WTFbank, one of our clients.
As a banking service provider, we regularly have to remind clients to send us fresh certificates for their sites. Maybe their Outlooks are too forgetful, their staff turnover is too high or two years validity is an unfathomably long period for them, anyway they rely on our monitoring to tell them when they should renew. Their certificates and obviously the corresponding keys live on our servers so theoretically we could just do it for them but for legal reasons we can't. We just generate a Certificate Signing Request for them so they can pass it on to the CA.
Two years ago, WTFbank ignored our CSR, generated one for the wrong domain name, and had a certificate issued for that. This year, we told them they could just reuse our old CSR and have a new certificate made from that. Instead they decided to stick their old certificate and the key in an unencrypted email and send it merrily across the intertubes. To us, who already have them. To do nobody-knows-what with them
@LaoC said in Linux locks and a kinder, gentler Linus:
ATMs are an instructive example of the few IoTish devices that run Windows. They're also a pretty extreme example in that they're not cheap mass-market devices, buyers are usually not completely clueless, [...]
I'd like to qualify that:
… in relation to your average IoS user
-
@dkf
You assume they extended past 2021? Brave man...I assume that they only changed one digit of the regular expression they use to validate the date.
-
-
Two years ago, WTFbank ignored our CSR, generated one for the wrong domain name, and had a certificate issued for that. This year, we told them they could just reuse our old CSR and have a new certificate made from that. Instead they decided to stick their old certificate and the key in an unencrypted email and send it merrily across the intertubes. To us, who already have them. To do nobody-knows-what with them
By the way, they've been fucking around long enough with this for the certificate to expire today. So they managed to have both a compromised key and an expired certificate. WTFbank FTW!
It's not quite as bad as it sounds though because for raisins they've been including some graphics via HTTP for a long time so the site always shows up as having broken security in browsers anyway—apparently their customers don't mind (enough).
-
@LaoC There's something I don't understand though : if you are working with Certificate Signing Requests, how come they know the private key? The whole CSR procedure is meant to avoid sending private keys, so wasn't the private key generated on your server then?
EDIT: Just realized that there might be some shitty bussiness-continuity rule in play here where your customer wants to be "in charge" of all key material so they could "transparently" jump ship if they want another service provider.
Seems they haven't considered that maybe their secrets aren't really safe if they're going to be handing copies of it anyway (whether or not it happens via a secure channel).
-
@TwelveBaud said in WTF Bites:
a second fixed-length byte array for a heap
That's retarded. Not even actual native applications work like that - heap is always handed in by the OS in chunks on demand.
But because the same cannot be done for stacks, the operating systems have long learned to handle the already “allocated” buffers in chunks anyway. So it's not actually a big problem.
For example about a month ago I was checking what eats up most memory in our project and noticed that because there is zillions of threads all over the place, and each gets 8MiB stack allocated, though most fortunately only ever use a couple of kilobytes, the allocated virtual memory is around 10 times what the device even has available.
-
@TwelveBaud said in WTF Bites:
a second fixed-length byte array for a heap
That's retarded. Not even actual native applications work like that - heap is always handed in by the OS in chunks on demand.
But because the same cannot be done for stacks, the operating systems have long learned to handle the already “allocated” buffers in chunks anyway. So it's not actually a big problem.
It's not a problem because the system knows exactly which chunks are used and which aren't. Which leaves us with 3 options:
- The browser will have to replicate the entire functionality of the system memory manager and provide address space virtualization so the huge ass heap array they pass to WA only takes as much space in physical memory as needed. We need to entrust browser developers with making this system 100% bulletproof, because otherwise every WA program will have to deal with OOMs and random memory corruption all over the place.
- Same as above, except the browser doesn't implement memory management itself, but forwards all allocations and deallocations to the OS. Still have to get the virtualization right, and now the stakes have risen from crashing the single WA program to crashing the entire browser process (which, admittedly, might not be a real difference since all browsers have already switched to 1-process-per-tab model).
- Browsers give up and provide a dense array, and now web developers must manually assign the maximum heap size, carefully balancing between OOMing their program and OOMing the browser.
-
The browser will have to replicate the entire functionality of the system memory manager and provide address space virtualization so the huge ass heap array they pass to WA only takes as much space in physical memory as needed.
No, the only thing it needs to do is not touch content of the huge ass array except in locations explicitly requested. Then the system still knows what memory is used and all is fine.
-
@Bulb this will work for the most part, but it won't allow the OS to reclaim freed memory as long as the program is running. Also, it relies on very efficient in-program memory allocator that aggressively reuses freed memory for new allocations.
-
@Gąska It does not allow OS to reclaim freed memory, but that's the case with most other VMs as well.
… sooner or later they'll probably add a
mmap/VirtualAllocequivalent, but for the time being we have to contend withsbrk().
