ReactOS 0.4.12 Pulls In Wine-Staging 4.0 DLLs, Many Kernel Improvements

[DMJC]

Every time I've used ReactOS it's been noticeably better than the previous time. I think that ReactOS will hit a tipping point similar to WINE where a ton of the things that don't currently work just work out of the box. It feels like things are rapidly stabilising.

[Michael_S]

This is a project that seems interesting the first time you see it. Then you test it on a VM and... it falls into pieces. Software like Firefox or Inkscape fails to run properly, and you have way better chance of getting a Windows program running under Linux + WINE. I have been following the development for years, and it seems it is never going to take off.

Having a free (as in freedom) drop-in replacement for Windows is a powerful idea. Unfortunately I doubt it is ever going to happen.

My (casual) understanding is that a lot of ReactOS contributors are WINE contributors, so even if ReactOS is never a flawless replacement for Windows XP the project helps WINE. And these days, WINE is amazing.

[edwaleni]

Wine is good for a few things, but as a general replacement for Windows, I would say it is not.

The general focus of its effort is for games pretty much and since i don't use Windows to play games its utility is somewhat limited.

I will keep futzing with it as it progresses. I just finished another round of testing with some Cisco VPN products as a client where Linux is not permitted.

[DMJC]

I'm mostly using ReactOS to emulate things like 3D Studio Max and Caligari TrueSpace. I can copy files into/out of the VM using filezilla (admittedly an older version) But I am considering contributing some code to ReactOS. BTRFS support got very stable very quickly. I think that 64-bit support isn't that important, people need to look at realistic use cases for these operating systems. Sure 64-bit and DirectX 12 support would be cool, but where ReactOS could really gain traction is as a drop in RDP server/Client or a Domain Controller/DNS Server. Stuff where people don't want the complexity of a Linux machine and are happy to just drop in a small VM of ReactOS without all the bloatware. 64-bit is already a work in progress, DirectX 10/11/12 will happen eventually when they implement a Vulkan driver, after that they just leverage WINE's Vulkan->DX12 code. I prefer to think of ReactOS as an opportunity to undo the mistakes of the Windows 98SE/Windows ME to Windows XP transition. Since we have the code to the OS itself, it's possible to write better integrated handling of some of the Windows 95/98 games that won't emulate on modern systems. In theory we could even bring back support for dropped windows items such as gameport drivers.

[willbprog177]

I try ReactOS out when I hear about a new alpha release. Today I tried it out and it's doing much better than the past. You can install VirtualBox Additions on it without BSODing now. You still cannot run newer versions of Firefox on it and while the ReactOS Application manager version of Otter browser works pretty well, I hit a page to download Java, only to have Otter crash. ReactOS never once crashed this time, which is the first time it's never crashed on me. It's getting there!

[ElectricPrism]

Even if ReactOS fails, their efforts in reverse engineering Windows and its kernel will help anyone who wants to preserve Windows for the future, and make it possible to run software in the future.

I assume this is similar to HURD, where the reason for it's failure is that not enough people work on it to make it usable, and it's impossible to keep the needed pace of development.

Failure and Success need not be seen as if they were binary.

Their efforts already have contributed to the major success of WINE, and Proton/SteamPlay among other real world 2019 benefits for Linux users.

Success is incremental, and transactional, meaning small successes like getting a game working on 2001 have had a "Tsunami" like effect gathering power over time so that 2019 has widespread WINE support.

[Nemerian]

Success is incremental, and transactional, meaning small successes like getting a game working on 2001 have had a "Tsunami" like effect gathering power over time so that 2019 has widespread WINE support.

This.
ReactOS is the perfect example of the Pareto principle.
Look at the tests for each new version.
Recently they've added the reasons why they've failed, and the corresponding bugs on Github.
90% of them(barring visual bugs, artefacts and glitches, which are more annoying that breaking, and are tied to rendering, anyhow) are half a dozen major bugs(which are usually tied to flawed or missing architecture elements).
And they chop down a couple every release, which is why people say ReactOS feels significantly more stable and complete with every single new version.

Once they prune away the last of them, ReactOS will speed up tremendously.
I mean, look at Haiku, which isn't even related to other OS families.
Since they've hammered away the last things that were blocking R1, in the past year, since they've released Beta, they've been adding all sorts of crazy modern, new developments.
Even ReactOS has been adding stuff that's missing from Windows itself(Btrfs)

[dragon321]

Just out of curiosity can it load a real graphics driver yet? (i.e. not say QXL or Serious on a vm) Anybody try getting it to run with a Intel/ATI/Nvidia card?

It should but of course only XP/2003 drivers (no support for WDDM yet). On ReactOS forum ("EPIC WIN" thread) somebody posted screen with installed ATI drivers and working OpenGL acceleration (ReactOS doesn't support native DirectX yet and translates it to OpenGL just like Wine). There was some graphical artifacts which is normal in alpha state OS.

Even if ReactOS fail to replace Windows it would be nice to have open source NT compatible OS. I don't think Windows will be open source in near future just like OS/2 isn't open source today. It has probably too much licensed and patented code which can't be simply released to open source even when Windows will be abandoned. ReactOS doesn't have this problem - it's written with open source in mind.

[angrypie]

I don't think Windows will be open source in near future just like OS/2 isn't open source today. It has probably too much licensed and patented code which can't be simply released to open source even when Windows will be abandoned.

They can still open-source the parts they own. Or do it like OpenSolaris/illumos: release closed-source code as binary and replace it later.

Of course I don't think they'll ever do that. At least not until IBM agrees to do the same to OS/2 (not gonna happen either).

[starshipeleven]

They can still open-source the parts they own.

In many long-lived closed source projects it may not be so clear what is truly owned by them and what comes from a third party. Also, opening closed source code will expose them to a ton of litigation over patents, just because there is a crap-ton of them and only closed source obscurity ensures that none can try attacking them.