Hacker News new | comments | show | ask | jobs | submit login
ReactOS 0.4.7 released (reactos.org)
229 points by jeditobe 10 months ago | hide | past | web | favorite | 59 comments

It's really interesting ReactOS can run both Chrome and Firefox. The latest versions of these browsers do not run on any old Windows: Chrome has required at least Windows 7 since version 50 (April 2016) and Firefox has required this since version 53 (April 2017) - although Firefox 52 ESR is still supported until May 2018.

Although I don't follow ReactOS, I thought they were working on supporting older Windows (Wikipedia says they try to run on Windows Server 2003 drivers). So does this mean they are catching up so fast to newer Windows versions or are we in fact seeing (very) old browsers in these screenshots?

I don't think the ReactOS people are trying to do Windows versions like Russian nesting dolls. They try to aim to fully support the older version, but while they're trying to fill all the nooks and crannies of them, quite often you'll find yourself adding some new API from a more recent version, if that's all you need to support a specific kind of application.

I'd naively assume that finishing the undocumented API calls of the last century is a much more time-consuming task than adding what little came after 2003...

Is it still targeting Windows XP though? Or is it just Windows NT more broadly now, albeit with a much older looking interface?

I'm not seeing the claimed browser support. The ReactOS Application Store has Firefox 45, which successfully updated to Firefox 47... but then Firefox 52 ESR crashes on start. I'll have to uninstall and go back to 47 to get Firefox to work.

The Application Store has Opera 12.18 (from 2012), so I used that to try and download Chrome. Chrome tells me my computer isn't supported, but offers a download anyway. But when I try it, the installer stops after "On your marks!" so I have no idea how they took that Chrome screenshot.

Opera 12.18 is running okay so far, though.

The browsers shown in the screenshot are pretty old. In fact, that screenshot seems to be Chrome ~47 or so, going off the UI icons.

While they target an older driver API, the userspace is mostly Wine's, so it is much easier for them to support recent software.

In fact, for the last 10-15 years, Wine itself to some extent is written against Windows kernel APIs such as ntdll.dll, and only the lowest levels talk to the host OS and wineserver. This means that ReactOS can keep also some of Wine's lower level DLLs, such as kernel32.dll, as long as there is something below them.

It's great to hear that this project is still going on, ReactOS is quite a milestone in terms of dedication and compatibility.

I should try it in a VM again, maybe I can get my old VB6 IDE running again...

On my desktop, it seems to work pretty well in Virtualbox -at least until I tried to install Office 2010 on it.

It doesn't seem to deal well with the audio and network drivers, though.

I had no luck installing visual studio 6 in my vm. My old copy of borland 5 seems to be installing just fine, though.

Edit: I may have spoken too soon; I had to run bcw.exe manually, and got a message about 16 bit WoW not being supported.

The VB6 IDE is distressingly easy to get running on modern versions of Windows - this is fortunate, since we still need to support it, but upsetting, since it diminishes the incentive for killing it off.

The resistance to "killing off" VB6 is in large part based on there being nothing as good to replace it. Unless you're willing to pay thousands for a Delphi licence, your only options are:

1. Modern .NET for desktop apps, whose many semi-obsolete UI frameworks make Java's UI tools look simple

2. Learning HTML/JS/React/Redux/CSS/Bootstrap/Python/Django/SQL/SQLAlchemy/Webpack...

Neither is likely to appeal to someone used to how easily applications could be created in VB6 or Delphi.

(I believe this so strongly I cofounded https://anvil.works, which is basically VB6/Delphi for the web. That's not so much a disclaimer, I guess, as putting my money where my mouth is :-P)

There's also Xojo, which used to be called REALbasic. It's a cross platform (Mac/Windows/Linux) IDE similar to Visual Basic, with the indie developer community of the earlier Delphi versions, and a much cheaper price point (essentially $99 per platform):


If you're trying to replace VB6, the only UI framework to consider on .NET would be WinForms, anyway. Which is obsolete, but still supported.

I don't see how Delphi + VCL is any easier than C# + WinForms, especially given that the latter is a GC language.

> maybe I can get my old VB6 IDE running again...

Why? Just run delphi 7 so that you at least will have sane exception handling ;)

But on a more serious note I wonder how the lazarus-ide will do on ReactOS.

Mostly because VB6 is the IDE I learned programming with and I think I heavily used it until '06, at which point I switched to C# and .NET and much other fun.

Mostly for Nostalgia...

You bringing up VB6 made me smile. Anything VB makes me nostalgic as well.

During Thanksgiving I talked to my Uncle who gave my family our first computer and also asked him if he remembered giving me his old copy of VB 3.0 to install. We couldn't remember if it was 20-something or 30-something 3.5" floppy disks but I remember feeling really cool being the only person out of my programmer buddies who had a legit VB 3.0 install.

That’s where I learned programming as well. Around that time someone hooked me up with a VB 1.0 IDE. It ran on windows 3,iirc. Or was it straight DOS? Talk about blast from the past.

There was a VB 1.0 for DOS, but also a 1.0 for Windows released earlier I believe. VB for DOS was a bit of a dog, IIRC.

Or just install XP in a VM, then VB6 will work for sure.

The incredible work that goes into ReactOS is truly awe inspiring. It gives me hope for Haiku

Reactos is a great project that can help to preserve old Windows software. Also their work can be used by Wine (not a Windows emulator) developers.

It's more often the other way round; ReactOS uses Wine DLLs (in fact many Wine DLLs can be even used on Windows).

However, the reverse engineering work of ReactOS developers has indeed benefited Wine too.

They've also uncovered and fixed plenty of bugs in wine, and implemented a lot of missing functionality.

I love seeing that GPL license. Nobody can try to pull an Apple with the OS stealing valuable momentum and userbase. Whether you approve or not of such behavior, it's pretty unambiguous that it happens.

Succeeding by making use of open source is not "stealing." Were Sony's developers stealing when they integrated FreeBSD into the PS4? No. They were following the licensing terms set by the FreeBSD project.

Apple's XNU is actually also released as open source (for the most part; last I checked, the parts that pertained to iOS weren't exported). It's based on Mach (which wasn't licensed in an open source way! http://www.cs.cmu.edu/afs/cs/project/mach/public/FAQ/license...) as well as 4.3BSD (which is BSD-licensed).

Stealing momentum and stealing a product are entirely different things, I never said using open source products in your business without contributing is stealing. It can however rob a project of momentum. Yes a business can share their progress, but without the gpl they do not have to. I should have known that was a loaded word, what word would you have used there so that I don't upset you in the future?

No word would avoid "upsetting" me, as it's not the word I take issue with. I think the premise of your argument is wrong.

If a company comes up with a new feature for a piece of open source software, it is in their best interests to push that code upstream. When they do so, the maintainers will continue to develop new features while maintaining this upstreamed feature. As long as someone else finds use in that feature, the company basically recoups their "investment" made by publishing it.

This is true no matter what OSS license your code is published under. And companies that don't upstream their code will fall behind those that do (except when it comes to companies that deal with drivers and hardware; that's a whole different ballpark).

With less-copyleft-than-GPL licenses, companies are allowed to choose what they share. I believe this weans them onto the Open Source train rather than purely allowing them to steal from it.

If you're rallying for software freedom above all else, why not back the AGPL instead? It's so copyleft it's GPL incompatible, because it covers SaaS as well as traditional software.

I do support the AGPL as well, I just don't see it being a large concern that this operating system will become a web service.

I don't see issue with other open source licenses, they all push toward the correct direction. I just also prefer licenses that prevent corporations from hiding the code in a binary after selling it. Sometimes the corporation benefits from taking an adversarial relationship, especially those who are hoping to hide their flaws and backdoors from users. You can argue once the users find out it really hurts the corporation, and they'll always eventually find out, but you can bet that the shareholders will have cashed out long before that happened. While the best growth is always mutual gain, some are willing to take a less effort win/lose, and I would like to be able to trust that some of our software is safe from that. It's my personal opinion that FreeBSD throws pearls to swine, and that we should be using such a license for libraries that we don't mind being incorporated into commercial software, but not for our entire operating system. We both benefit from the competition, but I don't see much benefit in giving the most powerful party even more power than they already have.

Even with the GPL, as long the software is only run in house, you don't have to contribute back.

There is no software license where that isn't true. I don't even know how such a license could be enforced anyway so it's a bit silly. Neat fact though.

> Apple with the OS stealing valuable momentum and userbase.

What did they "steal" that is of value that nobody can access after the "theft"?

Also their not alone because there is a `etc\hosts` file in my windows driver directory.

Stealing momentum and userbase are not theft, but rather meant to convey a loss for the open source project despite being the ones who got the ball rolling. It was expressly permitted by the BSD license, so I'm not saying the Apple did anything wrong, but rather pointing out a opinion about a strategic misstep by the BSD license.

for context https://forums.freebsd.org/threads/41109/

Please, Linux is the direct competitor of FreeBSD, not anything Apple-related. This is obvious GPL-vs-BSD trolling attempt and you know that.

It's not about the users it's about the developers, and no I'm not trying to troll. I'm sorry that what I said upsets you, but I don't think I was saying anything offensive. I admit "Stealing" is a loaded word and I was using it loosely, I didn't think it would upset people because I wasn't meaning it literally.

Thought I'd try the new release, and while there's plenty of glitches (eg it blue-screened after install), there's some useful things that are working. Sublime Text installed and seems to work, though the antialiasing is not very good & makes the text difficult to read.

I keep testing my own Win32 Photoshop plug-ins on ReactOS, and they seem to work fine in my makeshift graphics editor. I'm having a hard time finding any glitches at all with those plug-ins now.

As I mentioned elsewhere, Opera 12.18 works okay, and I'm posting this comment from ReactOS & Opera (installed via Parallels Desktop).

Wait... you have written a "makeshift" graphics editor that can already use Photoshop plugins? I would love to hear more about that!

It was made with Delphi 2005, ImageEn [1] (a Delphi component for loading & saving image file formats) and the Centaurix Delphi Photoshop component. Sadly Centaurix went out of business. Their SDK was designed for writing Photoshop plugins, but it also had some functions for loading & testing Photoshop plugins, and I used those. There were a lot of bugs that I had to patch to get it working.

I made it because my products are Photoshop plug-ins [2], but a lot of people were downloading them without having Photoshop. People would complain after installing that there wasn't a program to run. I thought including a bare-bones standalone graphics editor (no layers, no selections) with the Photoshop plugin download would solve that problem.

Actually, it didn't move the needle at all. Seems the people most likely to buy already had Photoshop / PaintShop Pro etc anyway. But it's been useful for running tests on ReactOS, or back when I was using Wine instead of Parallels on my Mac.

[1] https://www.imageen.com/

[2] https://www.namesuppressed.com/

Could you make some demo videos about your program and ReactOS?

That's a great idea, I hadn't thought of making a screencast of it running on ReactOS. Might take me some time to put together, I'm not sure Camtasia / BB Flashback will work on ReactOS... but it's worth a try!

Wonderful story. Thank you

Not sure where the 'with Four different browsers supported' part of the headline came from -- searching for 'browser' gives no results on the page. Is this from the screenshot of Chrome / Firefox / Opera / K-Meleon?

I would hope the OS can support a truckload more than four browsers! I would be interested to learn how well ancient Internet Explorer versions work.

I commented elsewhere about it, but I can't even see how they took that screenshot. Firefox 52 ESR crashes at startup for me, and Chrome refused to even install on ReactOS. Opera 12 is the only one I've got to work reliably-ish.

I assume the screenshot must be old versions of the browsers.

Use Firefox 48 from RAPPS.

If you want Chrome you need to find Chrome 49 (Chrome 50 ends support for Windows XP) and run it with --no-sandbox key

It supports software that only runs properly in older versions of Windows so its almost a XP type of OS but can run the latest of browsers at the same time. Its kind of a big deal when you think about it from that point considering modern browsers do not support XP and co anymore

Indeed the text in the screenshot implies that it can run all browsers, not four.

Maybe it's time to undust those Win32 ASM skills...


  c:\dir\ssvn update
  This might take a while, so please be patient.

  Local Revision: 76032
  Online HEAD Revision: 76032

  Your tree is up to date.

So, can anyone confirm for me that 76032 is the reversion that they're calling 0.4.7?

As said in the article, they switched to git.

Nothing makes me more sad than this project.

We've asked you time and time again to start commenting substantively. If you won't, we'll ban the account.


Not sure why you should feel that way. There is nothing wrong in trying to create a free clone of a popular software product. (There's also FreeDOS - a high-quality clone of MS-DOS.)

Maybe he feels like it would be more beneficial if the work were put into WINE instead.

There are lots of people who could switch to Linux if certain apps worked well under WINE. With the anti-user shenanigans Microsoft is doing in Windows now (e.g. using upstream bandwidth to seed updates, sucking up people's Internet connections, and potentially busting their data caps), this is needed now more than ever.

But no one's going to replace Windows with ReactOS. For example, is ReactOS going to keep up with drivers for current hardware? Of course not. Do they regularly release security updates? No. But Linux and Linux distros already do. So improving WINE helps people actually switch off Windows to a safer, less anti-user system.

> is ReactOS going to keep up with drivers for current hardware? Of course not.

Unlike Linux, ReactOS doesn't need to: one of the design goals was to provide binary compatibility with regular Windows drivers.

That's really neat if it works out. But doesn't this mean that users have to own a copy of Windows to use those drivers legally? Maybe they can download some drivers from manufacturer sites, and maybe some of their EULAs don't require the user to own a copy of Windows, but I'd be surprised if many users could run a machine with only Windows drivers without any of the drivers included in Windows releases.

But even if that all works as well as it does in Windows, it's still inferior to Linux. Users are stuck with whatever binary drivers they can get, and whatever reverse-engineered, bug-for-bug compatibility the ReactOS developers can manage. It would be much better for users if they could run WINE on Linux.

This isn't an either/or situation. The ReactOS developers collaborate heavily with the WINE developers, so many updates to ReactOS also benefit WINE and vice-versa.

That sounds nice, but wouldn't it be better if the effort were put toward WINE alone?

I just don't understand. WINE is like a necessary evil, a noble sacrifice of its developers' time to enable users to switch from Windows to a free, open platform for the applications they can't run on one natively. It's a sensible abstraction that enables something that wouldn't otherwise be possible.

But why reimplement Windows, the whole OS? If the devs enjoy it, it's their time to use as they choose. But the end result would be much more beneficial for users if it were put toward WINE so users could leave the Windows platform behind.

>is ReactOS going to keep up with drivers for current hardware?

They don't need to. They do implement whatever needed to have ABI compatibility with actual Windows drivers, so the vendor driver should work.

That sounds really cool. But:

1. I presume this requires "bug-for-bug compatibility." How feasible is this without access to either the Windows source code or a lab full of machines and QA staff?

When a user gets a ReactOS BSOD because of a bug in the ReactOS driver ABI, what do they do? Even if they can report a bug and it eventually gets fixed, what do they do in the meantime? The answer is, probably, boot back into Windows and get their work done. Then they have to justify to themselves taking the time to try ReactOS again sometime in the future.

2. How many machines in the wild can run without any Microsoft-made drivers included in Windows releases? If the answer is less than 100%, those users will still have to own a copy of Windows.

I don't know why either. As it stands right now, the parent poster is vaguebooking.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact