Thread 106294122

Will it ever be useful in production?

>>106294122 (OP)
Maybe in like 27 years

>>106294122 (OP)
this is another jeebuntu-based distro?

>>106294170
No its a windows clone

>>106294183
ouch. I think it will be a better windows than windows ever was. So linux will be better technically, but this will have a huge install base of legacy software.

>>106294170
It's an attempt to clone Windows as a free and open-source operating system through reverse-engineering.
Problem: It's based on Windows Server 2003 (with very limited userspace support for some Windows Vista programs).

>>106294254
ReactOS has many issues. But being based on windows 2003 is not one of them. Vista is where windows started going very wrong. What would even be the point of ReactOS if they moved to NT6+, at that point you might as well just use modern windows.

>>106294298
Drivers and outdated security model by design.
Now, to be fair, ReactOS was intended as a drop-in replacement for Windows XP/2K3, but will production machines running for 40 years under XP really switch?

Probably not. For games in particular DXVK / Wine is good enough now. For legacy software you should just write new software.

>>106294170
Fucking zoomer newfag

>>106294254
>Problem: It's based on Windows Server 2003 (with very limited userspace support for some Windows Vista programs).
They're slowly refactoring the OS to a moving NT6+ target (team has decided on eventually dropping srv03 during CLT2025), with NT6.0 being the fundamental base for the while being.
All of the works are in that repo until they bringup to the mainline when it's ready.
https://github.com/ReactOS-Longhorn-Initiative/reactos
Main releases should be affected (like x86/AMD64), pretty sure the original xbox and pc98 ports will continue to be built with NT5-compliant code in mind.

>>106294361
Security is a meme.

As for Drivers, adding support for newer drivers does not require NT6, its possible to add the missing features without the massive redesign of NT6.

The only real pain point is graphics drivers. WDDM will never work for NT5 due to the complete removal of GDI acceleration.

The ideal solution would be not to bother with WDDM, instead write custom drivers for ReactOS/NT5. I don't think it would be that impossibly difficult, the NT5 driver model is pretty simple. And the amdgpu driver is open source, so it should be possible to figgure out how to initialize the gpu and set it up for receiving basic command buffers.

>>106294122 (OP)
If anyone could fork it and make a Media Center clone it could be a pretty decent HTPC OS imo

production is a meme

>>106294122 (OP)
It already is.

>>106294361
>XP
>40 years
>check calendar
Don't scare me like that anon. WTF

>>106294525
righteous fury
>>106295635
This.

>>106294674
Oh, interesting. So they're not going to maintain both side by side or something, they're really going to target NT6?
I wouldn't be surprised if they tried to keep the two models.
>>106294680
It's work that hardware manufacturers won't bother with. Now, to be fair again, they won't bother making WDDM Vista drivers again either, but it would make porting easier.
If we're making drivers from the ground up then the problem is indeed the same.
>>106295679
I meant that as in "ReactOS will be ready in 20 years so the machines it targets will have been running XP for 40 years".
>>106295635
>>106295706
Uh... You might be able to run Minesweeper in VMs just fine, but on bare metal it's not. It also still crashes on various random things.

>>106295795
>Uh...
>still le random crashes lolidkwt \o_o?/
Thank you for your incredibly thoughtful and information filled post.

>>106294122 (OP)
All of the windows reverse-engineering expertise that went into this toy instead of improving wine on an OS worth running is a tragedy.

>>106294122 (OP)
Jack of all trades, master of none. They decided to waste time and resources trying to make it compatible with Vista+ slop os' so they never ever came close to finishing the 2000/XP stuff, which would have been it's strength in a better reality.

>>106294298
>Vista is where windows started going very wrong
The only thing wrong with Vista is that it significantly bumped hardware requirements. It had bugs but they were fixed with Vista SP3 (Windows 7)

>>106298806
Vista completely fucked up the graphics stack. Compositing added 1-2 frames of lag. And you could not really opt out either since WDDM removed all the hardware acceleration from GDI so it runs like garbage without the compositor. It was a crime against humanity, classic windows had a zero lag UI, everything was rendered instantly.

>>106294122 (OP)
I've always liked ReactOS. It's a shining example of what happens when you let freetards try to write something like Windows.
Even having the real thing right in front of them for 30 years, this software equivalent of some west Pacific cargo cult's coconut radios and P-51 Mustangs made out of dried grass, is the best they could come up with.

>>106294122 (OP)
>ReactionaryOS

>>106294122 (OP)
no, because Microsoft will go out of their way to fuck it up. Besides, React is based on stolen code.

>>106298827
I still laugh everytime I see this revisionist history post. It's the "we wuz" of the Vista hater.

>>106298889
Its an objective fact that classic windows had a more responsive UI than Vista or any version of windows since. Classic windows when coupled with a decent gpu could update/repaint a window in under 1ms.

>>106298827
Vista was deliberately programmed to be incompatible with everything XP/2000, which they'd later "fix" with the sloppiest emulation possible. Truly the beginning of the end for MS os'.

>>106298918
Because it was designed to run on a Pentium with 32MB of RAM. Vista wasn't.
And that's why you seethe. You're not just admitting your current socioeconomic status - you're admitting it's affected you your entire life.

>>106298927
If you install vista on a trillion gigahertz computer its UI latency will still be worse.

>>106298943
Ahh, the classic "if I can't do something, nobody can".
Classic fallback of the lazy, stupid, incompetent and/or poor - but believes the participation trophies he got as a kid actually mean something.

>>106294122 (OP)
This OS would have developed rapidly if they were not overshadowed by Microsoft who could sue them at any time for copyright infringement, currently they are reverse engineering it using a clean room method, which is slow compared to copying the Windows source code that was leaked several years ago.

>>106298960
What the fuck are you on about you are just speaking gibberish now.

Also, Vista UI lag is not caused by lack of performance. Its by fundamental. When the compositor is enabled there is 1-2 frames of lag caused by the rendering pipeline. And with the compositor disabled gdi has to render to system memory, and then at some point copy that system memory to the gpu. But its not done very well, sometimes the OS won't bother copying for like 30 ms, so the lag is terrible.

>>106294680
as long as it runs my software and lets me rice it to look like windows 95 I dont care what they do, but I strongly doubt this crap can run ancient rpg maker, 2hu games, or other doujin products. this is the area where linux vulcan fails and shits bricks.

>>106298827
Worked on my machine

>>106298998
>people laughing at me for seething, foaming at the mouth and pissing myself because im poor and have been for at least 20 years
>is gibberish
Glad we had this chat.
Guys, jacket and mouthguard for this one.

>>106295795
>Oh, interesting. So they're not going to maintain both side by side or something, they're really going to target NT6?
Yeah, the NT5/NT6 gap increasing that turned out to be too huge for the then-goal of full srv2003 compatibility.
I think what was their initial goal with NT6 back before the LH Initiative happened was creating static libraries for NT6-only kernel-mode stuff, but that's more of like putting lipstick on an NT5 pig. And like you've mentioned - trying to implement NT6+ API calls in the user-mode.
There were some "strikebreakers" in the team that wanted ReactOS to still be NT5-only, including the then-release engineer (guy still uses XP as host, complained about RosBE dropping VS2010 support, so he can't build anymore; guy's also got that autistic ritual of updating release ISOs with fixes backported from latest master, including those from several years back). But none of that matters as the team decided to reset the goal to NT6+.
>they settled on Vista (NT6.0) as baseline, as it wasn't a complete restructure as 7 (NT6.1) was internally, yet brang many improvements and additions for what makes modern Windows
>although certain parts of the NT6 architecture that will be used in ROS can be compliant with later NT6 releases and made compatible: e.g. ACPI/PCI being Win10 1607-compliant, PO being 9600 (NT6.3)-compliant
>there will be shims to run applications that won't run on baseline Vista, for the while being
The guy who's primarily working on the NT6 rewrite is the same guy who was working on SMP and UEFI support. (especially the latter - guy's working for one of the UEFI vendors, Phoenix IIRC)

>I wouldn't be surprised if they tried to keep the two models.
Well, it would be hard to imagine NT6 working well on the original XBOX, let alone the PC-98, especially considered the latter's architectural differences over regular x86.

>>106294240
the thing about reactos, is that it shared a ton of code with wine. which makes sense since a lot of wine is just a clone of windows userspace. reactos goes a step further and clones windows kernelspace as well.
what this means in practice is that wine will basically always be better than reactos at running most windows software (outside of drivers and the like that operate in windows kernelspace) since the userspace is literally the same code just running on an already-stable kernelspace (linux)

>>106294361
reactos didn't start off targeting xp/2003, it started back when windows 95 was the current version of windows. the initial effort was with freewin95. i'm sure sometime in the future reactos will start targeting later versions of windows, it's likely just that xp/2003 is still useful enough at this stage in development that there isn't a great pressure to try to move to vista/7/2008 just yet

i'd see flying cars before reactos reaches the full capability of windows 7

>>106301259
It was worth continuing to target 2000/XP compatibility because of legacy software. It's taken so long that that's pretty much over now and the few examples left could probably get away with running in a container.

>>106301416
after writing that i realised there's another issue here. while it's obvious they couldn't target the current version of windows as they couldn't possibly keep up, even moving to a newer-but-still-old target has similar issues. like if development moves to nt6.x, then nt5.x won't be completed. so there's the question, do you try to finish one regardless of how old it becomes, or do you just keep trying to chase after newer ones?
it's not an easy thing to answer.

objectively speaking, a production ready clone of xp/2003 is more useful than a perpetual alpha of anything, but idk, it depends on the specific goals of the project

>>106301540
Best car scenario it becomes the XP version of dosbox. More realistically it'll be alpha forever.