Poe’s Law is vastly over-invoked. In any social setting where some amount of context can be assumed, it doesn’t apply.
And even if someone wanted to challenge me on that, I’d argue that the tiny number of people who would think “HN is a tech support site” is being said seriously aren’t worth the two seconds it takes to type “/s”.
The time it takes to type `/s` is a decent bit lower than two seconds, and yet the amount of time you've spend trying to argue that it's not worth it is vastly more.
I invoked Poe’s law sarcastically. I’m surprised you didn’t realize that, given how over-invoked it is here. I guess there’s always someone who doesn’t get the joke. /s. <—- shouldn’t be necessary, but is, obviously. /s <—- not sure if necessary here.
But this is exactly what I’m saying: if you’re sarcastic and the audience gets it, it’s good sarcasm. If you’re sarcastic and the audience doesn’t get it, it’s bad sarcasm. If you don’t have a clear enough context, just don’t be sarcastic.
There’s always someone who doesn’t get the joke. Even today, there are people who, at least at first, are fooled into thinking that Jonathan Swift’s “A Modest Proposal” was a serious proposal.
In spoken word, sarcasm in even the cleanest context always has the benefit of delivery paving the way for virtually guaranteed reception. A written indication does look silly when reception succeeds without its help, but guaranteed parity with speech seems like the winning maneuver.
The whole point of sarcasm is to say something you don't mean, and people who know you or other facts well enough, can get a dopamine hit for being able to infer that. It's just an ingroup/bonding thing and carries no information or value outside of that. At any rate, adding a /s is like saying "just being sarcastic", might as well not be sarcastic in the first place.
I think what you're describing is a subset of sarcasm, or potentially a separate thing that often utilizes sarcasm (significant overlap). A friend of mine will often suffix such statements in speech with "if you know you know" in the event that the delivery alone isn't sufficient to trigger everyone's memory, and "IYKYK" in text.
Pure sarcasm has no such prerequisite, which means that delivery alone (tone, cadence, and smirk) is almost always sufficient. These give away the fact that sarcasm is happening to roughly the same extent that "/s" does, just that we have more experience with the former so the latter tends to feel too spoon-fed.
Oracle VirtualDoxx likes to phone home. I'd rather use qemu.
* edit: Oracle VirtualBox even calls home to check if you've got a valid Windows Server License, and it will Narc and Cop you out. Use qemu or something else; people! Oracle is the cousin of cancer. If an Oracle product is free, please re-think! Stay safe, people! Oracle VirtualBox virtually, not a pun, logs tons of things that you do within it.
Virtualbox has been controlled by Oracle since that company's acquisition of the former Sun Microsystems in 2010.
Oracle has a remarkably chequered past regarding its hosting of Free Software projects: OpenOffice, OpenSolaris, Mysql, and AFAIR several Apache projects, off the top of my head.
Virtualbox itself has been somewhat closed down / restricted over the course of Oracle's tenure.
This could well be an early indication of further such moves on Oracle's part. That's very decidedly not clear at the moment, but would well bear further observation.
> I seem to recall trademark troubles with Oracle, which led to everyone forking and starting Jenkins.
Well, you have a keyword in there, "everyone". All but 1 major contributor (not even the main contributor and definitely not the project architect and lead) disagreed with their decision.
Oracle persisted with the help of that major contributor - an Oracle employee, BTW - and while Jenkins was being created, they did not want to contribute and instead tried to create a competing effort over at the Eclipse Foundation.
Jenkins was receiving 95% of the contributions because, as you said, "everyone" was helping out, while Hudson was slowly dying on the vine at the Eclipse Foundation, and Oracle was putting out cheerful release announcements for it <<for years>> even though anyone who checked the commit logs could clearly see that Hudson was dying.
Oracle kept pushing with Hudson internally for quite a few years but even they gave up and buried Hudson a while ago.
Similar story with OpenOffice. They <<really, really>> wanted control so they forced everyone else to force through their actions and OpenOffice basically had the same face, Apache Foundation, death. Though OpenOffice is more of a zombie these days, I don't think it's been shut down completely like Hudson.
So a super small toy CI server from 18+ years ago still has its source available yet Hudson, while the OG big CI server is completely gone from the internet, including Oracle's modifications.
You've got it completely backwards with regard to OpenOffice. LibreOffice was prematurely forked from it for the sole reason that "Oracle owns it now"; what played out was that Oracle managed to relicense the entire suite to be under a single license (the Apache 2 license) while LibreOffice, to this day, continues with the triple-licensed mess that Sun had implemented. LibreOffice got popular, OpenOffice died off; Oracle handed it over to the Apache Software Foundation for it to die.
1. Why does it matter LibreOffice is triple-licensed, though?
2. Why do you think LibreOffice became popular? Hint: most of its contributors moved over from OpenOffice because few people trusted Oracle. Trust matters.
I was addressing the question of why the issue itself might gain attention on HN, and be worthy of deeper investigation, particularly in light of a rather snarky "slow news day" sibling response to the OP.
Oracle deserves a very close watch, and I've strong suspicions that overreliance on Virtualbox will end in tears.
It's not a belief that the cause is malice, it's a suspicion that it could be, or might portend same.
> The extensions pack was always under a proprietary license.
Yes, while I'm not sure if the actual terms of the license changed, enforcement changed very much.
With SUN ist was "Of course, you can personally install it on your work computer and use it for work. Only if you deploy it to your fleet of computers please pay us."
What I meant is, a lackadaisical approach to billing and asking for money was common at Sun for everything software related, same thing for Java which was definitely a line item.
Cost of revenue for penny-ante fluff would have been orders of magnitude greater than actual revenue. Goodwill and other benefits substantively greater.
Blaming Sun's demise on Virtualbox largess is ludicrous on its face.
Sun has its lunch eaten by Linux and commodity hardware. Selling proprietary UNIX and super expensive hardware was a great business until Linux and (comparatively) cheap Intel boxes were good enough for more and more customers.
For example they removed PCI pass-through, they never updated their GPU guest drivers to support modern OpenGL, they never improved their (experimental) support for Mac OS X guests, in general the development of VirtualBox has been pretty much been completely halted over the last decade, in comparison to VMWare they practically went backwards.
> in comparison to VMWare they practically went backwards.
Well that's interesting. I recently went from free virtualbox to a paid vmware workstation license because of some frustrations with vbox, and found that vmware is not better or worse but just frustrating in different ways (e.g. clipboard sync doesn't have a directional option and shared folders are harder to set up, but usb passthrough is much more stable). Performance is identical for various workloads as near as I can tell, except that vmware's VM displays freeze for 0.5 seconds every time you focus/unfocus one (virtualbox is seamless). If vmware went forwards compared to that, one would have to have been crazy to pay if you can just apt install virtualbox
>Something is amiss. It's been reported, should be fixed soon. Meanwhile user "marmel" has noticed that the downloads are still available on https://download.virtualbox.org/virtualbox/
Fine with me I recently moved to KVM from Vbox. It took a few hours of paper cuts to get the first VM up and running but once done it's very easy to use. Good timing for me.
I've always wondered why people would use 3rd party virtualization solutions instead of the native ones. Is it for better performance? Is it so that they could use the same tools as their colleagues using a different OS? Are there things which flat out don't work or are much more difficult to configure?
In my case, Hyper V and KVM always allowed me to do what I needed.
Shared folders don't work as well (still haven't figured out how to mount a read-only folder in the guest or a read-write folder where permissions translate correctly), linked clones aren't straight forward (and you can't modify the parent while a child clone exists), and the UI is much worse compared to VirtualBox, which is mostly self-explanatory. With KVM/QEMU I found it hard to even find the right man page with the information I'm looking for.
I still prefer KVM over VirtualBox, but I'd be lying if I said that it wasn't more work.
TIL Hyper-V is more than system calls, I didn't know it has a GUI (edit: apparently it's command-line only, and the GUI I saw on an image search is on Windows Server only, so 99% of people can't use this already). Thought that it was a set of APIs for vmware and the likes to use under Windows, and that Virtual PC was what Microsoft offered as front-end.
But which other "native" virtualisation platforms do you mean? I don't know that the Linux Foundation has an official virtualisation product, and couldn't name one for macOS or iOS either.
Edit: and Hyper-V is apparently missing basics like any audio support at all, so that makes it useless for most users. Not sure this would be anyone's default choice when you can use cross-platform, open source, and more featureful software
> edit: apparently it's command-line only, and the GUI I saw on an image search is on Windows Server only
It's not server-only, but AFAIK does require the pro edition, though. But I suppose that or enterprise is what people usually use in a professional setting. The feaure is called "Hyper-V management tools" or similar.
> But which other "native" virtualisation platforms do you mean? I don't know that the Linux Foundation has an official virtualisation product, and couldn't name one for macOS or iOS either.
I meant KVM for Linux. I don't know what the Linux Foundation recommends (I never interact with it), but KVM is integrated in the Linux kernel, which seems as "native" as can be.
There's a macOS virtualization framework, but I've stopped using that platform right around when it came out, so I can't recall its name.
> Edit: and Hyper-V is apparently missing basics like any audio support at all, so that makes it useless for most users.
I didn't know it missed audio support. I also wonder whether "most users" actually require audio from their VM. I'd expect the main use of a VM to be running random programs that don't work on the native OS. On Windows, that usually means running some Linux program. Those don't often require sound. KVM supports sound, without issues.
> Not sure this would be anyone's default choice when you can use cross-platform, open source, and more featureful software
Because it's already there, mostly, and doesn't require you to carefully read a known-hostile company's EULA.
---
edit: Don't these nice functions (sound, mouse integration and folder sharing) actually require the VBox "Extensions"? IIRC, those aren't open-source.
VBox extensions are needed for remote desktop integration and USB passthrough newer than USB 1.1 iirc. Hmm, and maybe 3d acceleration? Not sure exactly, but anyway, iirc: sound, folder sharing, and definitely mouse integration are part of the open source / freely licensed part.
I didn't know that KVM was a Linux kernel API actually, seems odd that they're not all using that as that sounds like it should indeed give the best performance and stability! Apparently proxmox does, but not vmware, virtualbox, qemu, or others whose name I might have heard.
> I'd expect the main use of a VM to be running random programs that don't work on the native OS
At work we use it for security, so things like meetings (video calls) are in a separate VM, so sound is a nice to have there, one could say. Privately, I indeed use it for running software that WINE has trouble with, like games, where sound is also not exactly optional if you want "your base is under attack" alerts. Lastly, I use it for some Windows server software that I fell in love with a decade ago and didn't manage to find a good Linux replacement for: there, indeed, I didn't need sound.
> Because it's already there
My start menu doesn't have virtualisation software out of the box, and apt install virtualbox is as simple as figuring out what front-end might be able to talk to KVM. Or, for Windows, finding where to get that command-line utility that talks to Hyper-V in the kernel for non-pro Windows versions
> Apparently proxmox does, but not vmware, virtualbox, qemu, or others whose name I might have heard.
qemu is just a frontend, it can run on top of KVM.
> At work we use it for security, so things like meetings (video calls) are in a separate VM, so sound is a nice to have there
Indeed. But are you running meetings in a Linux VM on top of a Windows install? I was specifically thinking about non-windows apps on top of Windows. KVM supports sound. Windows guests on HyperV also support sound because you actually connect to those via RDP, which can "share" the local sound card. Recent versions of Ubuntu with Gnome also come with some kind of RDP server out of the box that supports sound, so such a guest on Hyper V would also have sound.
I get that Virtualbox does have some mind share, so if you just want to set up a VM without going down some rabbit hole, you'll just install that.
I use 7.0.x for an old Debian install (3.1), FreeBSD and Windows (all guests, the host is Windows 11 or Fedora 37). All seem to work fine, at least none of the major issues that thread brings up (3daccel on windows, broken efi, etc).
Compatibility with Windows seems a lot better in 7 compared to 6, both as host and as guest. They also seem to have added a number of features such as secure boot. It seems that there was quite a time gap between 6 and 7 so a whole lot of changes built up.
Whether those things matter to you, or not, is in the eye of the beholder.
VirtualBox 7 came out almost exactly one year after the initial release of Windows 11. The bypass of the Secure Boot requirement was needed for quite a while.
Sure, I know it was required for a while, but ISTR there being a very visible burst of work toward it and TPM support right after that requirement change became public.
I tried it last week, and it was a terrible experience. I used one of the official windows eval VMs, it was crashing all the time, and when it ran, it was laggy, with a lot of visual bugs.
My guest is a Fedora laptop with AMD CPU and GPU, and running the same VM with VMWare worked like a charm.
Virtual box seems to have a lot of issues with amd GPU's.
Whatever host or guest OS I use, I always find graphical bugs inside the guest VM. Some programs like Chrome seem to fall back to some software rendering compatibility mode, despite 3d acceleration being enabled, and things that don't fall back have issues with some things being overly transparent, like the windows start menu.
I lost about a day trying to debug packer builds of vagrant boxes for vbox last week. The vbox builds were glacially slow and would reliably fail at rebuilding the tools dkms components.
That was on 7.0.8 I think. "Lucky" me 7.0.10 was released that same day, but didn't show up in an update check. I just happened to see the reference on their site to 7.0.10 and it resolved both issues from the previous day.
I don't have much day to day use to report with it though, I don't use the vbox images much myself.
I tried it when I re-installed my workstation with Kubuntu 23.04. It was a shit-show of invalid links and errors when I tried to get the extensions and runtime modules for the VM's. After about 45 minutes I purged it and installed libvirt and Virtual Machine Manager.
I'm using it now and appears to work well on Manjaro, but I had to roll back from a rt kernel I was trying because it couldn't build the necessary modules.
I would migrate to virt-manager in a second if directory and clipboard sharing wasn't that convoluted requiring multiple reads/watches of manuals and videos to do what on Virtualbox has been a lot more intuitive for ages. (solution: never ever let developers design user interfaces :^)
I've been using 7.0.x for over half a year, currently on 7.0.8, on a Windows host. Most of my guests were installed during 6.x and have continued to run fine on 7.0.x, and new guests made in 7.0.x run and behave normally. No problems on my end.
I would be cautious or even distrustful of using anything from Oracle. VirtualBox components come under three different licenses - GPLv2, personal use & evaluation license, and an enterprise license. Their VirtualBox license FAQ [1] gives them enough leeway to change future licenses at will. If an exploit is discovered in your old VirtualBox and they've changed the license, you're out of luck.
Be specially careful when installing their extension pack, as it is an evaluation license.
We've moved our development to KVM and Virtual Machine Manager on Linux [3] and UTM on Mac [4]. There are other options to run your VM, such as Multipass [5] or VirtualBuddy [6].
On a digressive topic - it was fun migrating our legacy application server stack from Oracle Java (old & poorly considered decision) to OpenJDK, thanks to their license [2].
In past oracle lawyers have send huge bills to companies after someone in their IP range had used virtualbox guest editions (which is not free for commercial use).
Sure, but it's reasonable for most people to expect an extension they've downloaded from within the app itself might have similar licensing as the app itself.
Even if it isn't reasonable to do so, I suspect most people do indeed do so, and that's exactly what Oracle is counting on.
It was required (at least used to) to use USB2.0/3.0, RDP compatibility, PCIe passthrough. I remember having to download that, so my guess is either I'm not typical user, or there are two different worlds with respect to how VBox is used.
I think it's extremely uncommon for most people to need either rdp or PCIe passthrough, and I'd guess it's still quite uncommon (though more likely than the previous two) to need usb passthrough.
I have been using Virtual Box for years without needing any of those.
When it finally came down to PCIe pass through to get better 3D support I moved to qemu/kvm, no one forced me to pirate Oracles software for that. Of course qemu/kvm can be an absolute shit show to configure, so I still default to Virtual Box.
You are confusing the Guest Additions (free, FOSS, included in many distros) and the Extension Pack (commercial/proprietary, freeware for personal/non-commercial use.)
I have had to explain this within the last fortnight; read that.
Sorry, I did in fact mean extension pack. The FOSS one have not always been available/complete hence in past a lot of companies had no choice other than using the Oracle guest addition extension pack. Which resulted in situations like this:
The Guest Additions (GAs) are not part of the Extension Pack (EP). As far as I can recall, after I think 15 years of using VirtualBox now, the GAs never were part of the EP.
The GAs are FOSS.
The EP is not FOSS. Since VirtualBox 4.0, released December 22, 2010, the core product has been FOSS and the proprietary bits were in the EP and only the EP.
VirtualBox 2 and 3 had separate freeware PUEL and Open Source editions (OSE), but that is back in the noughties. You are reaching back nearly 15 years here to get to when the product was not all-FOSS.
I think you are confusing the names and the descriptions of the two totally separate add-ons, one of which is a proprietary addon to the hypervisor itself, and the other of which is a FOSS addon to the OS running inside the VM.
Can you expose some more on the licensing issues with migration from Oracle Java to OpenJDK? I will admit that nothing jumps at me from the FAQ, but then it's rather... legalistic one.
> Their VirtualBox license FAQ [1] gives them enough leeway to change future licenses at will. If an exploit is discovered in your old VirtualBox and they've changed the license, you're out of luck.
That only applies to the optional Extension Pack. The base software for host and guest is GPL v2, as described in the FAQ you linked.
----
Will the licensing terms of the VirtualBox base package be changed in the future?
We are not currently planning to do so, but it doesn't really matter. The license that applies to the VirtualBox base package is perpetual and cannot be taken back unless you violate the license. Even if Oracle were to change the license again, this would only affect future versions, and anybody still in compliance would still be permitted to redistribute the existing VirtualBox code under the terms of the GPLv2.
I think the point is that the patch for a critical security vulnerability _could_ be gated behind a license change.
Seems unlikely though even for Oracle. If it did happen I expect someone else would come up with a freely licensed fix and that would be shipped by all the distros.
Out of curiosity since it seems you have a lot of experience in this field, how does UTM compare to VMware Fusion on macOS? I used to run VirtualBox for years but switched to VMware due to recurring performance issues. Trying out VMware was like a breath of fresh air, things that used to take a lot of effort to configure suddenly worked right out of the box. How does UTM fare?
Oh I remember about 10 years ago doing the migration from Oracle JVM to OpenJDK. I think it would have been around the Java 6 period, but there were all sorts of small things that I didn't think would be an issue that ended up cutting us.
The two main ones were a font, which I don't remember exactly which or why but I think it was to generate PDFs and it came bundled with the Oracle JVM so the PDF library defaulted to it. For Open JDK there was an extra setup step to install a RedHat font, but it was confusing when a pdf comes out garbled because you switched your JVM. But I could be mixing up my issues.
I think the other was built in certificates, but it was too long ago now.
Having a fun time at the moment with Java 11 to 17. The application has extensive Lombok, so nothing quite works the same. I've had less pain migrating from Python 2 to 3.
Yeah, Lombok doesn't use standard, supported APIs. They even say in their docs, what they're doing could become incompatible for future java versions at any time. Not java's fault really.
It is for this reason I don't use it (also I don't like too much magic, but that's me).
It just magically understands your query based upon the name of the function.
There is far too much "magic" going on these days which has a massive learning curve for the newbies, and a headache for everyone else when something stops working because they've changed how the incantation works.
You don't need to have the "int" there, though, right? You can just do "val a = 1" [0]
Don't get me wrong, I still pick Java over Kotlin, but I thought ktl had type inference. Even Java has "var a = 1" now. I just wish we had optional named parameters in Java, that's like, my biggest issue with the language.
I took that example from the Kotlin homepage. I've not used Kotlin, but since Scala recommends that you include the type declaration if you know what it is, I assumed that Kotlin did as well. If their example is bad, then apologies!
Optional named parameters and also default values would be great.
The stuff that gets used by the most in this codebase is the `allParameterConstructor` for enums and the `slf4j` annotation. Without those two I could probably remove the Lombok dependency.
> Be specially careful when installing their extension pack, as it is an evaluation license.
Our company was targeted by Oracle Sales because they detected mere _downloads_ from our IP range. They couldn't prove any installations, but of course wanted the company to pay for commercial licenses.
Im not sure about the details, but in the end I think we bought a couple of licences & then blocked all URLs for the extension pack company-wide.
Probably not. But how would you feel about going up against Oracle's legal team to make the point? They, and other large companies, rely on this a lot when deciding upon whether or not to use dubious practises. I think Oracle's main business has become litigious & licensing arse-hole-ary, with everything else in their portfolio existing to support those functions, so they are rather good at it.
> After such an intrusion I wouldn't ever want to have anything to do with Oracle.
This sort of thing is why many people are wary of Oracle. The company has considerable form for this sort of thing and worse.
I had sales drones get me just for downloading VirtualBox as a whole from our IPs. They sent me an email, tagged High Importance even.
"My name is <name> and I am part of Oracle Open Source Linux & Virtualization team. I have got an internal notification from my team that you have shown interest in Oracle Virtual Box... I would like to schedule a time with you to discuss the use cases in your organization, our recent version 7 announcements, and how Oracle can assist you in getting the maximum value from this offering..."
What took me a bit to figure out is that this does not mount a single folder, but a virtual parent folder to which you can add and remove directories without changing the mount. This even works while the VM is running.
For example, mounting
share /mnt/host virtiofs rw,nofail 0 0
in the guest and then sharing
~/Downloads
in UTM will share the Downloads folder as
/mnt/host/Downloads
When using the QEMU backend, VirtFS is used instead, which does not use such a parent folder and requires restarting the VM to change shares iirc.
license issues apart, virtualbox guest tools for windows are awesome: clipboard sharing, drag and dropping files, network shares, auto window resize, etc just work out of the box.
With KVM/splicetools it is always a pain to setup that.
I use VirtualBox for trying things out or sandboxing, mainly because of the friendly GUI, but wouldn't mind switching to something as friendly that's not Oracle's.
From oracle, I only use an older version of Virtualbox and in cloud I only use their free VM tier (4 ARM cores & 12GB RAM) because it's the best current "permanent" offer out there. That said, I take care to not invest in to any of their specific cloud guff and just run my own or generally available software & OS on Linux VM's there, which is not too much work to migrate anywhere.
I've seen enough in the past to have zero trust in the company to not start doing more anti-dev stuff.
Last week I downloaded the latest VMware player and installed Windows 11. (I had to run the efi boot file manually) It works well, and it has a sane non-commercial license that doesn't allow the company to send tax collectors to your home.
Bridging wifi shouldn't work. Something something only one mac per wifi client. If it works somewhere else they're doing some ebtables/nftables l2 magic.
It can be this. The device is both a wifi client, does L2 magic for the ethernet port (you can't put a switch behind it) and it can rebroadcast wifi signals, because to do so you don't have to be a client to rebroadcast, but you have to half your airtime, so you're halfing your speed.
At least the ones I've used also had their own IP address (for the config interface). Hell, one even had a switch built in! (It was a home router with a pretend-to-be-an-extender mode.)
Oracle does a very good job in ensurnig their software is used by as least users as possible. My company has recently issued a global e-mail requiring all employees to uninstall everything related to VirtualBox and switch to different solutions (vmware, parallels). It's a shame, because VirtualBox was one of my favorite virtualizers, but at this point I'm afraid that Oracle lawyers will get me for staring on an icon for too long without paying an extended icon license.
Can I ask why? The "big selling point" for a long time has been the $0 cost - it's essentially common knowledge now that Parallels or VMWare will provide anywhere from slightly to much better performance on the same hardware.
Well, yes, that was the main thing, it was free to use, and it was easier to "convince" it to virtualize macOS on non-Apple hardware. Also I liked the fact that it was easy to run headless VMs (VBoxHeadless), and the "VBoxManage" command-line tool was pretty capable. Plus, the dual monitor functionality worked on my setup pretty well (Linux host and Windows guest).
But anyway, today I have everything I need through KVM and virt-manager when I want to run Linux or Windows guests, on macOS I use UTM for home/manual use. On macOS VirtualBox was a no-go because VirtualBox is an x86 virtualizer, so they don't have plans to support arm64 anyway.
> On macOS VirtualBox was a no-go because VirtualBox is an x86 virtualizer, so they don't have plans to support arm64 anyway
So apparently their long term goal[1] is to support Arm-on-Arm VMs, but they didn't even intend[2] to ship the x86-on-Arm version yet really, so don't hold your breath for Arm-on-Arm.
My understanding is that they never "intended" to release it at the time it was (based on a forum post from their devs).
Have you actually been using it?
It wasn't clear to me from the high level posts/commentary about it, until I had an Arm Mac to try it with, that it does not in fact run Arm VMs - it runs 32bit x86 VMs, extremely slowly.
on windows use hyperv as best solution, on linux we have best. same for mac. so just lacking, i did no searched so, one mgmt script set to abstract best per platform. vbox sucks if it is not open, there are closed best alternatives per platform. and there is vmware.
Just spotted - while looking around, seems like the links have returned: (for v7.0.10). Honestly, it was as I was looking at the history to see what happened on the page, it just re-appeared when I went back. (Clicked on the latest version number in the history, now it appeares every time)
151 comments
[ 3.0 ms ] story [ 206 ms ] threadWonder if 7.0.10 wasn't an "intended" update?
And even if someone wanted to challenge me on that, I’d argue that the tiny number of people who would think “HN is a tech support site” is being said seriously aren’t worth the two seconds it takes to type “/s”.
That amount is larger than is available in most (probably all) HN comment sections.
Pure sarcasm has no such prerequisite, which means that delivery alone (tone, cadence, and smirk) is almost always sufficient. These give away the fact that sarcasm is happening to roughly the same extent that "/s" does, just that we have more experience with the former so the latter tends to feel too spoon-fed.
Not clear sarcasm.
* edit: Oracle VirtualBox even calls home to check if you've got a valid Windows Server License, and it will Narc and Cop you out. Use qemu or something else; people! Oracle is the cousin of cancer. If an Oracle product is free, please re-think! Stay safe, people! Oracle VirtualBox virtually, not a pun, logs tons of things that you do within it.
Oracle has a remarkably chequered past regarding its hosting of Free Software projects: OpenOffice, OpenSolaris, Mysql, and AFAIR several Apache projects, off the top of my head.
Virtualbox itself has been somewhat closed down / restricted over the course of Oracle's tenure.
This could well be an early indication of further such moves on Oracle's part. That's very decidedly not clear at the moment, but would well bear further observation.
Oracle buried OpenOffice and Hudson just out of spite. Especially for Hudson I was watching it happen almost day by day when it occurred.
I seem to recall trademark troubles with Oracle, which led to everyone forking and starting Jenkins.
But I don't think I've heard of Oracle intentionally burying a product.
Well, you have a keyword in there, "everyone". All but 1 major contributor (not even the main contributor and definitely not the project architect and lead) disagreed with their decision.
Oracle persisted with the help of that major contributor - an Oracle employee, BTW - and while Jenkins was being created, they did not want to contribute and instead tried to create a competing effort over at the Eclipse Foundation.
Jenkins was receiving 95% of the contributions because, as you said, "everyone" was helping out, while Hudson was slowly dying on the vine at the Eclipse Foundation, and Oracle was putting out cheerful release announcements for it <<for years>> even though anyone who checked the commit logs could clearly see that Hudson was dying.
Oracle kept pushing with Hudson internally for quite a few years but even they gave up and buried Hudson a while ago.
Similar story with OpenOffice. They <<really, really>> wanted control so they forced everyone else to force through their actions and OpenOffice basically had the same face, Apache Foundation, death. Though OpenOffice is more of a zombie these days, I don't think it's been shut down completely like Hudson.
I checked now and Hudson seems to be completely gone: https://git.eclipse.org/c/hudson/
The Eclipse Foundation has taken down even the Git repo, let alone the website.
DamageControl, the original inspiration for Hudson/Jenkins still has a repo on GitHub: https://github.com/codehaus/damagecontrol
So a super small toy CI server from 18+ years ago still has its source available yet Hudson, while the OG big CI server is completely gone from the internet, including Oracle's modifications.
Kind of shameful from the Eclipse Foundation, especially after the initial boasting: http://www.h-online.com/open/news/item/Eclipse-Foundation-an...
2. Why do you think LibreOffice became popular? Hint: most of its contributors moved over from OpenOffice because few people trusted Oracle. Trust matters.
Oracle deserves a very close watch, and I've strong suspicions that overreliance on Virtualbox will end in tears.
It's not a belief that the cause is malice, it's a suspicion that it could be, or might portend same.
Chequered would be for cheques, as in pay cheque.
<https://sapling.ai/usage/checkered-vs-chequered>
Failure accepted.
How so? VirtualBox was GPL-licensed before the acquisition and it remains so today. The extensions pack was always under a proprietary license.
Yes, while I'm not sure if the actual terms of the license changed, enforcement changed very much.
With SUN ist was "Of course, you can personally install it on your work computer and use it for work. Only if you deploy it to your fleet of computers please pay us."
https://web.archive.org/web/20100204053823/https://www.virtu...
Cost of revenue for penny-ante fluff would have been orders of magnitude greater than actual revenue. Goodwill and other benefits substantively greater.
Blaming Sun's demise on Virtualbox largess is ludicrous on its face.
Well that's interesting. I recently went from free virtualbox to a paid vmware workstation license because of some frustrations with vbox, and found that vmware is not better or worse but just frustrating in different ways (e.g. clipboard sync doesn't have a directional option and shared folders are harder to set up, but usb passthrough is much more stable). Performance is identical for various workloads as near as I can tell, except that vmware's VM displays freeze for 0.5 seconds every time you focus/unfocus one (virtualbox is seamless). If vmware went forwards compared to that, one would have to have been crazy to pay if you can just apt install virtualbox
>Something is amiss. It's been reported, should be fixed soon. Meanwhile user "marmel" has noticed that the downloads are still available on https://download.virtualbox.org/virtualbox/
https://forums.virtualbox.org/viewtopic.php?p=538844#p538844
I've always wondered why people would use 3rd party virtualization solutions instead of the native ones. Is it for better performance? Is it so that they could use the same tools as their colleagues using a different OS? Are there things which flat out don't work or are much more difficult to configure?
In my case, Hyper V and KVM always allowed me to do what I needed.
I still prefer KVM over VirtualBox, but I'd be lying if I said that it wasn't more work.
But which other "native" virtualisation platforms do you mean? I don't know that the Linux Foundation has an official virtualisation product, and couldn't name one for macOS or iOS either.
Edit: and Hyper-V is apparently missing basics like any audio support at all, so that makes it useless for most users. Not sure this would be anyone's default choice when you can use cross-platform, open source, and more featureful software
It's not server-only, but AFAIK does require the pro edition, though. But I suppose that or enterprise is what people usually use in a professional setting. The feaure is called "Hyper-V management tools" or similar.
> But which other "native" virtualisation platforms do you mean? I don't know that the Linux Foundation has an official virtualisation product, and couldn't name one for macOS or iOS either.
I meant KVM for Linux. I don't know what the Linux Foundation recommends (I never interact with it), but KVM is integrated in the Linux kernel, which seems as "native" as can be.
There's a macOS virtualization framework, but I've stopped using that platform right around when it came out, so I can't recall its name.
> Edit: and Hyper-V is apparently missing basics like any audio support at all, so that makes it useless for most users.
I didn't know it missed audio support. I also wonder whether "most users" actually require audio from their VM. I'd expect the main use of a VM to be running random programs that don't work on the native OS. On Windows, that usually means running some Linux program. Those don't often require sound. KVM supports sound, without issues.
> Not sure this would be anyone's default choice when you can use cross-platform, open source, and more featureful software
Because it's already there, mostly, and doesn't require you to carefully read a known-hostile company's EULA.
---
edit: Don't these nice functions (sound, mouse integration and folder sharing) actually require the VBox "Extensions"? IIRC, those aren't open-source.
I didn't know that KVM was a Linux kernel API actually, seems odd that they're not all using that as that sounds like it should indeed give the best performance and stability! Apparently proxmox does, but not vmware, virtualbox, qemu, or others whose name I might have heard.
> I'd expect the main use of a VM to be running random programs that don't work on the native OS
At work we use it for security, so things like meetings (video calls) are in a separate VM, so sound is a nice to have there, one could say. Privately, I indeed use it for running software that WINE has trouble with, like games, where sound is also not exactly optional if you want "your base is under attack" alerts. Lastly, I use it for some Windows server software that I fell in love with a decade ago and didn't manage to find a good Linux replacement for: there, indeed, I didn't need sound.
> Because it's already there
My start menu doesn't have virtualisation software out of the box, and apt install virtualbox is as simple as figuring out what front-end might be able to talk to KVM. Or, for Windows, finding where to get that command-line utility that talks to Hyper-V in the kernel for non-pro Windows versions
I agree about the Oracle hostility thing, though
qemu is just a frontend, it can run on top of KVM.
> At work we use it for security, so things like meetings (video calls) are in a separate VM, so sound is a nice to have there
Indeed. But are you running meetings in a Linux VM on top of a Windows install? I was specifically thinking about non-windows apps on top of Windows. KVM supports sound. Windows guests on HyperV also support sound because you actually connect to those via RDP, which can "share" the local sound card. Recent versions of Ubuntu with Gnome also come with some kind of RDP server out of the box that supports sound, so such a guest on Hyper V would also have sound.
I get that Virtualbox does have some mind share, so if you just want to set up a VM without going down some rabbit hole, you'll just install that.
https://www.reddit.com/r/virtualbox/comments/y56gpv/anything...
Whether those things matter to you, or not, is in the eye of the beholder.
October 2022 https://www.virtualbox.org/wiki/Changelog-7.0
October 2021 https://en.m.wikipedia.org/wiki/Windows_11
My guest is a Fedora laptop with AMD CPU and GPU, and running the same VM with VMWare worked like a charm.
Whatever host or guest OS I use, I always find graphical bugs inside the guest VM. Some programs like Chrome seem to fall back to some software rendering compatibility mode, despite 3d acceleration being enabled, and things that don't fall back have issues with some things being overly transparent, like the windows start menu.
That was on 7.0.8 I think. "Lucky" me 7.0.10 was released that same day, but didn't show up in an update check. I just happened to see the reference on their site to 7.0.10 and it resolved both issues from the previous day.
I don't have much day to day use to report with it though, I don't use the vbox images much myself.
Everything is working as it should in my case.
P.S.: I'm using Debian's packages via unstable.
Fresh new install works good.
Only Windows-Guest under Linux-Host had some graphical issues.
Be specially careful when installing their extension pack, as it is an evaluation license.
We've moved our development to KVM and Virtual Machine Manager on Linux [3] and UTM on Mac [4]. There are other options to run your VM, such as Multipass [5] or VirtualBuddy [6].
On a digressive topic - it was fun migrating our legacy application server stack from Oracle Java (old & poorly considered decision) to OpenJDK, thanks to their license [2].
[1] https://www.virtualbox.org/wiki/Licensing_FAQ
[2] https://www.oracle.com/java/technologies/javase/jdk-faqs.htm...
[3] https://ubuntu.com/blog/kvm-hyphervisor
[4] https://mac.getutm.app/
[5] https://multipass.run/
[6] https://github.com/insidegui/VirtualBuddy
In past oracle lawyers have send huge bills to companies after someone in their IP range had used virtualbox guest editions (which is not free for commercial use).
This is false. https://www.virtualbox.org/wiki/Licensing_FAQ says:
> the VirtualBox Guest Additions are licensed under the GPLv2.
You might have been thinking of the VirtualBox Extension Pack, which is an entirely different thing.
Oracle could not milk VirtualBox for money so they used this as a way to extort unknowing commercial users.
Even if it isn't reasonable to do so, I suspect most people do indeed do so, and that's exactly what Oracle is counting on.
Where in the App does it download the extensions? I know it will download the Guest Additions for you, but those are something completely different.
1: https://docs.oracle.com/en/virtualization/virtualbox/6.0/use...
When it finally came down to PCIe pass through to get better 3D support I moved to qemu/kvm, no one forced me to pirate Oracles software for that. Of course qemu/kvm can be an absolute shit show to configure, so I still default to Virtual Box.
You are confusing the Guest Additions (free, FOSS, included in many distros) and the Extension Pack (commercial/proprietary, freeware for personal/non-commercial use.)
I have had to explain this within the last fortnight; read that.
https://news.ycombinator.com/item?id=36679707
https://www.reddit.com/r/sysadmin/comments/d1ttzp/oracle_is_...
The Guest Additions are capital F Free.
The Extension Pack is not.
There is no such thing as "the Oracle guest addition extension pack." Never has been.
You are conflating 2 totally different separate things, as is that Redditor.
Exactly what it was called is not important. The point is that you needed it and Oracle later made it a torpedo cost.
I think you are wrong.
The Guest Additions (GAs) are not part of the Extension Pack (EP). As far as I can recall, after I think 15 years of using VirtualBox now, the GAs never were part of the EP.
The GAs are FOSS.
The EP is not FOSS. Since VirtualBox 4.0, released December 22, 2010, the core product has been FOSS and the proprietary bits were in the EP and only the EP.
https://www.virtualbox.org/wiki/News_pre_4_3
VirtualBox 2 and 3 had separate freeware PUEL and Open Source editions (OSE), but that is back in the noughties. You are reaching back nearly 15 years here to get to when the product was not all-FOSS.
I think you are confusing the names and the descriptions of the two totally separate add-ons, one of which is a proprietary addon to the hypervisor itself, and the other of which is a FOSS addon to the OS running inside the VM.
That only applies to the optional Extension Pack. The base software for host and guest is GPL v2, as described in the FAQ you linked.
----
Will the licensing terms of the VirtualBox base package be changed in the future?
We are not currently planning to do so, but it doesn't really matter. The license that applies to the VirtualBox base package is perpetual and cannot be taken back unless you violate the license. Even if Oracle were to change the license again, this would only affect future versions, and anybody still in compliance would still be permitted to redistribute the existing VirtualBox code under the terms of the GPLv2.
----
I think this is what the parent commenter was saying:
> If an exploit is discovered in your old VirtualBox and they've changed the license
Seems unlikely though even for Oracle. If it did happen I expect someone else would come up with a freely licensed fix and that would be shipped by all the distros.
I prefer libvirt/kvm anyway. :)
The two main ones were a font, which I don't remember exactly which or why but I think it was to generate PDFs and it came bundled with the Oracle JVM so the PDF library defaulted to it. For Open JDK there was an extra setup step to install a RedHat font, but it was confusing when a pdf comes out garbled because you switched your JVM. But I could be mixing up my issues.
I think the other was built in certificates, but it was too long ago now.
Having a fun time at the moment with Java 11 to 17. The application has extensive Lombok, so nothing quite works the same. I've had less pain migrating from Python 2 to 3.
It is for this reason I don't use it (also I don't like too much magic, but that's me).
Not just you. 'Magic' seems great until newer members of the team have no idea how it works and just use it - then things get painful.
It just magically understands your query based upon the name of the function.
There is far too much "magic" going on these days which has a massive learning curve for the newbies, and a headache for everyone else when something stops working because they've changed how the incantation works.
I haven't picked it in the past except for a toy project, to experiment with Lombok.
It has some interesting ideas which would be nice if they folded it into Java, but while it's fringe it ends up being a liability.
> val a: Int = 1
I wouldn't want to inflict that on anyone.
Don't get me wrong, I still pick Java over Kotlin, but I thought ktl had type inference. Even Java has "var a = 1" now. I just wish we had optional named parameters in Java, that's like, my biggest issue with the language.
[0] https://www.oreilly.com/library/view/kotlin-programming-the/...
Optional named parameters and also default values would be great.
The stuff that gets used by the most in this codebase is the `allParameterConstructor` for enums and the `slf4j` annotation. Without those two I could probably remove the Lombok dependency.
Our company was targeted by Oracle Sales because they detected mere _downloads_ from our IP range. They couldn't prove any installations, but of course wanted the company to pay for commercial licenses.
Im not sure about the details, but in the end I think we bought a couple of licences & then blocked all URLs for the extension pack company-wide.
For instance Lead Forensics, but there are others in this space as well.
Probably not. But how would you feel about going up against Oracle's legal team to make the point? They, and other large companies, rely on this a lot when deciding upon whether or not to use dubious practises. I think Oracle's main business has become litigious & licensing arse-hole-ary, with everything else in their portfolio existing to support those functions, so they are rather good at it.
> After such an intrusion I wouldn't ever want to have anything to do with Oracle.
This sort of thing is why many people are wary of Oracle. The company has considerable form for this sort of thing and worse.
"My name is <name> and I am part of Oracle Open Source Linux & Virtualization team. I have got an internal notification from my team that you have shown interest in Oracle Virtual Box... I would like to schedule a time with you to discuss the use cases in your organization, our recent version 7 announcements, and how Oracle can assist you in getting the maximum value from this offering..."
How easy is it to setup directory sharing? Reading https://docs.getutm.app/guest-support/sharing/directory... seems a bit more difficult than VirtualBox
My use case is relatively simple: build some tiny OSes (into ISO image) on Linux VM, then run the ISO on macOS's qemu
What took me a bit to figure out is that this does not mount a single folder, but a virtual parent folder to which you can add and remove directories without changing the mount. This even works while the VM is running.
For example, mounting
in the guest and then sharing in UTM will share the Downloads folder as When using the QEMU backend, VirtFS is used instead, which does not use such a parent folder and requires restarting the VM to change shares iirc.With KVM/splicetools it is always a pain to setup that.
Do you mean Red Hat's Virtual Machine Manager, AKA virt-manager (https://en.wikipedia.org/wiki/Virt-manager), or Microsoft's Virtual Machine Manager (https://learn.microsoft.com/en-us/system-center/vmm/?view=sc...), or Synology's Virtual Machine Manager (https://www.synology.com/en-global/dsm/feature/virtual_machi...)? I'm guessing the first one?
I use VirtualBox for trying things out or sandboxing, mainly because of the friendly GUI, but wouldn't mind switching to something as friendly that's not Oracle's.
I've seen enough in the past to have zero trust in the company to not start doing more anti-dev stuff.
Stay tuned for more changes as soon as Broadcom takes control
But nobody tell them that hardware stores figured this out decades ago…
Can I ask why? The "big selling point" for a long time has been the $0 cost - it's essentially common knowledge now that Parallels or VMWare will provide anywhere from slightly to much better performance on the same hardware.
But anyway, today I have everything I need through KVM and virt-manager when I want to run Linux or Windows guests, on macOS I use UTM for home/manual use. On macOS VirtualBox was a no-go because VirtualBox is an x86 virtualizer, so they don't have plans to support arm64 anyway.
So apparently their long term goal[1] is to support Arm-on-Arm VMs, but they didn't even intend[2] to ship the x86-on-Arm version yet really, so don't hold your breath for Arm-on-Arm.
1: https://forums.virtualbox.org/viewtopic.php?p=523070#p523070
2: https://forums.virtualbox.org/viewtopic.php?p=537634#p537634
https://download.virtualbox.org/virtualbox/7.0.10/
Have you actually been using it?
It wasn't clear to me from the high level posts/commentary about it, until I had an Arm Mac to try it with, that it does not in fact run Arm VMs - it runs 32bit x86 VMs, extremely slowly.
I'm really curious who, if anyone, is using it.
Had a look on their page edits and couldn't see anything there: https://www.virtualbox.org/wiki/Downloads?action=history
Perhaps some form of glitch?
https://www.virtualbox.org/wiki/Downloads