>>24800
Nice trips
Half-Life 2 running on ReactOS on real hardware
LOUNGE · Back to threads · Report thread
Anonymous · reply · 2026-07-17 22:40:24
>>24652
Shit bro, I was running HL2 on Linux on day one. Same with Portal
Shit bro, I was running HL2 on Linux on day one. Same with Portal
the explainer · reply · 2026-07-18 02:05:19
>>24651
bro who cares, someone made halflife 2 here work inside a browser, get the fuck out of here
bro who cares, someone made halflife 2 here work inside a browser, get the fuck out of here
Anonymous · reply · 2026-07-18 02:59:09
>>24801
thx
thx
Anonymous · reply · 2026-07-18 03:11:22
>>24651
Wine could run it 19 years ago???
Wine could run it 19 years ago???
Anonymous · reply · 2026-07-18 03:30:10
>>24790
>try running real player
why? i can play my .rmvb's in mpv
>try running real player
why? i can play my .rmvb's in mpv
Anonymous · reply · 2026-07-18 04:35:10
>>24805
20*, the first successful reports (silver/gold) on winehq appdb are from 2006
20*, the first successful reports (silver/gold) on winehq appdb are from 2006
Anonymous · reply · 2026-07-18 10:12:56
>>24803
And?
And?
Anonymous · reply · 2026-07-18 10:20:17
>>24761
How is it Linus' fault? He already does everything right on his part, keeping all kernel interfaces stable forever and getting pissed off the moment someone breaks backwards compatibility.
That's more than NT does. NT breaks compatbility all the time and then expects the .dlls that ship with Windows to work around it to maintain compatibility.
It's the Linux userland that's a clusterfuck and that's because nobody has a common vision for what a Linux based OS is and isn't supposed to provide or what a Linux desktop really is beyond some vague XDG specifications.
How is it Linus' fault? He already does everything right on his part, keeping all kernel interfaces stable forever and getting pissed off the moment someone breaks backwards compatibility.
That's more than NT does. NT breaks compatbility all the time and then expects the .dlls that ship with Windows to work around it to maintain compatibility.
It's the Linux userland that's a clusterfuck and that's because nobody has a common vision for what a Linux based OS is and isn't supposed to provide or what a Linux desktop really is beyond some vague XDG specifications.
Anonymous · reply · 2026-07-18 11:09:53
>>24809
NT does not break backwards compatibility. The design is simply different. The API boundary on NT is the dll. The kernel boundary is an implementation detail that is not meant to be called directly.
NT does not break backwards compatibility. The design is simply different. The API boundary on NT is the dll. The kernel boundary is an implementation detail that is not meant to be called directly.
Anonymous · reply · 2026-07-18 11:13:20
>>24651
Incredible, they've reached the Windows XP era.
Incredible, they've reached the Windows XP era.
Anonymous · reply · 2026-07-18 11:15:19
>>24693
>>24695
Wine is the open source Windows runtime. If you care about running apps you're looking for Wine.
The main point of the reactos project is kernel space/drivers. It's extremely niche and more of a tech demo than anything else.
>>24695
Wine is the open source Windows runtime. If you care about running apps you're looking for Wine.
The main point of the reactos project is kernel space/drivers. It's extremely niche and more of a tech demo than anything else.
Anonymous · reply · 2026-07-18 11:17:09
>>24810
>>24761
>>24809
Windows has a stable *userspace* Linux has a stable kernel. The NT api is *only* available via the DLL and Microsoft changes the internal interface there all the time.
This is why everyone jokes that "wine is the stable Linux ABI" because it's actually true.
>>24761
>>24809
Windows has a stable *userspace* Linux has a stable kernel. The NT api is *only* available via the DLL and Microsoft changes the internal interface there all the time.
This is why everyone jokes that "wine is the stable Linux ABI" because it's actually true.
Anonymous · reply · 2026-07-18 11:18:52
>>24779
>Try getting an old Linux binary running
Almost no one shares Linux applications as binnaries. That's an extremely annoying way to distribute software and none of us want that to become common.
Give us the source or just build an EXE that works in Wine.
>Try getting an old Linux binary running
Almost no one shares Linux applications as binnaries. That's an extremely annoying way to distribute software and none of us want that to become common.
Give us the source or just build an EXE that works in Wine.
Anonymous · reply · 2026-07-18 11:18:55
>>24810
Fair. And yet, people depend on it and other undocumented parts of Windows all the time and then complain when their software breaks when Windows updates, while well behaved software even from the 90s still runs, unchanged, on modern windows...
Fair. And yet, people depend on it and other undocumented parts of Windows all the time and then complain when their software breaks when Windows updates, while well behaved software even from the 90s still runs, unchanged, on modern windows...
Anonymous · reply · 2026-07-18 11:30:32
>>24815
The kernel boundary interface is pretty much never called directly by anyone. I would not even consider it an undocumented feature. Not unless you also consider random unexported worker functions undocumented features. The kernel boundary should be considered purely an implementation detail. And the fact that the system service table numbers are shuffled every single release of windows, and also cannot be used from WOW64 means that this one thing is something that is truly never touched by anyone.
The kernel boundary interface is pretty much never called directly by anyone. I would not even consider it an undocumented feature. Not unless you also consider random unexported worker functions undocumented features. The kernel boundary should be considered purely an implementation detail. And the fact that the system service table numbers are shuffled every single release of windows, and also cannot be used from WOW64 means that this one thing is something that is truly never touched by anyone.
Anonymous · reply · 2026-07-18 11:38:30
>>24816
You'd be surprised at the absolutely retarded things anti cheat and "security software" developers do.
They do exactly that. YOLO Syscalls directly, by their ever shuffling numbers.
You'd be surprised at the absolutely retarded things anti cheat and "security software" developers do.
They do exactly that. YOLO Syscalls directly, by their ever shuffling numbers.
Anonymous · reply · 2026-07-18 14:58:03
>>24814
>Almost no one shares Linux applications as binnaries.
Remember, they tried to get that shit working with the Linux Standard Base, but the work was too boring and too high skill level for them. Need at least a master's degree and experience writing your own compiler.
>Almost no one shares Linux applications as binnaries.
Remember, they tried to get that shit working with the Linux Standard Base, but the work was too boring and too high skill level for them. Need at least a master's degree and experience writing your own compiler.
Anonymous · reply · 2026-07-18 15:10:57
>>24651
If AI is so good, why can't someone just vibe code a Windows clone?
If AI is so good, why can't someone just vibe code a Windows clone?
Anonymous · reply · 2026-07-18 15:18:11
>>24819
presence of leaked windows code in llm training data means if you vibecode anything microsoft doesn't like it can shut you down.
presence of leaked windows code in llm training data means if you vibecode anything microsoft doesn't like it can shut you down.
Anonymous · reply · 2026-07-18 15:26:02
>>24820
I want to see courts participate in that mental gymnastics olympics, cuz Faceberg believes its fair use after they got sued for using Anna's archive for LLM training data
I want to see courts participate in that mental gymnastics olympics, cuz Faceberg believes its fair use after they got sued for using Anna's archive for LLM training data
Anonymous · reply · 2026-07-18 15:30:04
>>24821
It's a very easy thing: If I let faceberg win I get more money if I don't I won't
It's a very easy thing: If I let faceberg win I get more money if I don't I won't
Anonymous · reply · 2026-07-18 15:30:17
>>24795
nowadays they can vibe code it under a year
nowadays they can vibe code it under a year
Anonymous · reply · 2026-07-18 15:30:35
>>24821
The courts already ruled that AI code has no copyright, so a Windows clone made by AI would be fair play.
The courts already ruled that AI code has no copyright, so a Windows clone made by AI would be fair play.
Anonymous · reply · 2026-07-18 15:31:12
>>24821
books are meant to be read by someone though.
leaked code is always illegal.
books are meant to be read by someone though.
leaked code is always illegal.
Anonymous · reply · 2026-07-18 15:40:44
>>24813
>>24816
This is true, and Microsoft has the right approach in making the userspace the focus of the OS instead of the kernel. The kernel barely matters in Windows nowadays, hell, the Win32 API was not even created for the NT kernel but for DOS, to begin with. If Microsoft wanted it, they could create another kernel and built Windows and all its dependencies on top of it and call it a day.
>>24816
This is true, and Microsoft has the right approach in making the userspace the focus of the OS instead of the kernel. The kernel barely matters in Windows nowadays, hell, the Win32 API was not even created for the NT kernel but for DOS, to begin with. If Microsoft wanted it, they could create another kernel and built Windows and all its dependencies on top of it and call it a day.
Anonymous · reply · 2026-07-18 15:45:40
>>24825
Yes, but you have to consider that Facebook pirated books. That is illegal by all definitions and interpretations of copyright law even if the book is meant to be "read", reading outside the contexts the publisher allows it to be read(i.e. after it was bought) it is still illegal.
>leaked code is always illegal.
For all we know microsoft could have fed Windows's source code to the LLMs as training data, at that point it becomes murky whether it is illegal or not
Yes, but you have to consider that Facebook pirated books. That is illegal by all definitions and interpretations of copyright law even if the book is meant to be "read", reading outside the contexts the publisher allows it to be read(i.e. after it was bought) it is still illegal.
>leaked code is always illegal.
For all we know microsoft could have fed Windows's source code to the LLMs as training data, at that point it becomes murky whether it is illegal or not
Anonymous · reply · 2026-07-18 15:47:14
>>24826
> If Microsoft wanted it, they could create another kernel and built Windows and all its dependencies on top of it and call it a day.
This statement has problems. The user space and kernel space are deeply intertwined. Like, the entire windowing system on windows is implemented inside kernel space. And the semantics of the filesystem is highly dependent on kernel. You could make a new kernel in the sense of implementing something similar to NT, but you can't just swap the kernel out for something fundamentally different.
> If Microsoft wanted it, they could create another kernel and built Windows and all its dependencies on top of it and call it a day.
This statement has problems. The user space and kernel space are deeply intertwined. Like, the entire windowing system on windows is implemented inside kernel space. And the semantics of the filesystem is highly dependent on kernel. You could make a new kernel in the sense of implementing something similar to NT, but you can't just swap the kernel out for something fundamentally different.
Anonymous · reply · 2026-07-18 16:04:13
>>24827
good point. I'd love to see llm successfully launder leaked code personally. resurrect opera presto.
good point. I'd love to see llm successfully launder leaked code personally. resurrect opera presto.
Anonymous · reply · 2026-07-18 16:43:30
>>24828
Yet you could run an .exe in both 98 and 2000 with no problems even when both of them had totally different kernels.
Yet you could run an .exe in both 98 and 2000 with no problems even when both of them had totally different kernels.
Anonymous · reply · 2026-07-18 16:46:08
>>24740
At this pace, it's never going to be
At this pace, it's never going to be
Anonymous · reply · 2026-07-18 16:47:41
>>24830
only an exe without any W functions (unicode) would work on both.
only an exe without any W functions (unicode) would work on both.
Anonymous · reply · 2026-07-18 18:35:17
>>24651
What's stopping them from recycling leaked Windows code through LLMs to make this now? Everyone is doing it with MIT/GPL code
What's stopping them from recycling leaked Windows code through LLMs to make this now? Everyone is doing it with MIT/GPL code
Anonymous · reply · 2026-07-18 18:36:11
>>24833
nobody enforces mit/gpl
nobody enforces mit/gpl
Anonymous · reply · 2026-07-18 18:36:44
>>24833
again, better wait for a good precedent in court than rush to become the bad one.
again, better wait for a good precedent in court than rush to become the bad one.
Anonymous · reply · 2026-07-18 18:42:17
>>24833
Because the most important property of an OS is correctness. Debugging an operating system is close to impossible. It has to be correct from the get go, vibe coding will not work.
Because the most important property of an OS is correctness. Debugging an operating system is close to impossible. It has to be correct from the get go, vibe coding will not work.
Anonymous · reply · 2026-07-18 18:43:18
>>24833
No precedent has been set in court yet and ReactOS in the past already got accused of using leaked source code and had to do a full audit to prove that wasn't the case
Basically, they're not going to take that risk
No precedent has been set in court yet and ReactOS in the past already got accused of using leaked source code and had to do a full audit to prove that wasn't the case
Basically, they're not going to take that risk
Anonymous · reply · 2026-07-18 19:42:00
You can run hl2 in a browser: https://hl2.slqnt.dev/
Anonymous · reply · 2026-07-18 19:52:34
>>24838
This is probably xash3d and not the original HL2, no? That's a significantly easier target.
This is probably xash3d and not the original HL2, no? That's a significantly easier target.
Anonymous · reply · 2026-07-18 20:12:02
>>24832
>unicows.dll laughs in your face
>unicows.dll laughs in your face
Nintendo delenda est · reply · 2026-07-18 21:53:28
>>24781
>16-bit software
lol, lmao even
>If I ran 10 LTSC
funny you didn't mention it's the 32-bit version. Really kills your argument
>16-bit software
lol, lmao even
>If I ran 10 LTSC
funny you didn't mention it's the 32-bit version. Really kills your argument
Anonymous · reply · 2026-07-18 22:14:19
>>24841
Windows 11 doesn't have WOW32, you moron.
Windows 11 doesn't have WOW32, you moron.
Anonymous · reply · 2026-07-18 22:52:28
>>24841
>failed to look at screenshot
>failed read
Look harder
Then KYS in shame
>failed to look at screenshot
>failed read
Look harder
Then KYS in shame
Anonymous · reply · 2026-07-19 00:28:02
>>24841
Everyone that's not retarded knows AMD 64 can't run 16 but code from 64 bit mode.
Everyone that's not retarded knows AMD 64 can't run 16 but code from 64 bit mode.
Anonymous · reply · 2026-07-19 03:42:03
>>24844
wine must be doing something illegal!
wine must be doing something illegal!
Anonymous · reply · 2026-07-19 03:45:09
>>24840
you need to compile a program specifically targeting these unicows of yours, it won't help you running an arbitrary wide-unicode exe.
you need to compile a program specifically targeting these unicows of yours, it won't help you running an arbitrary wide-unicode exe.
Anonymous · reply · 2026-07-19 04:06:54
>>24844
That is not the reason. The NTVDM codebase has an x86 emulator which could have been used on amd64. The actual reason is because the number of significant bits in handles was increased to 32bits. NTVDM passes handles through 16bit code, so would have required a significant redesign to work on 64bit windows.
That is not the reason. The NTVDM codebase has an x86 emulator which could have been used on amd64. The actual reason is because the number of significant bits in handles was increased to 32bits. NTVDM passes handles through 16bit code, so would have required a significant redesign to work on 64bit windows.
Anonymous · reply · 2026-07-19 04:13:01
>>24826
>This is true, and Microsoft has the right approach in making the userspace the focus of the OS instead of the kernel.
No.
It enables the ape-like ecosystem they've created so if people want that they should use the Win32 API/ABI.
I would argue that's the wrong objective and therefore certainly the wrong approach.
>This is true, and Microsoft has the right approach in making the userspace the focus of the OS instead of the kernel.
No.
It enables the ape-like ecosystem they've created so if people want that they should use the Win32 API/ABI.
I would argue that's the wrong objective and therefore certainly the wrong approach.
Anonymous · reply · 2026-07-19 04:46:05
>>24846
Yes. But nobody's complaining that an "arbitrary wide-unicode exe" doesn't run on Windows 95 - and if they are, they need to be laughed at.
This is only about backward compatibility: the post that started this subthread was about 9x EXEs on NT kernals.
Yes. But nobody's complaining that an "arbitrary wide-unicode exe" doesn't run on Windows 95 - and if they are, they need to be laughed at.
This is only about backward compatibility: the post that started this subthread was about 9x EXEs on NT kernals.
Anonymous · reply · 2026-07-19 05:13:00
>>24803
wtf im playing it now this is pod racing
wtf im playing it now this is pod racing
‹ Newer posts Page 4 of 5 Older posts ›
Reply to this thread
Posting requires a durable account or a TripPass. Reading remains available.
Sign in or create or attach a TripPass to post. Browsing is open.