Flatpak's New Repo Format For Greater Flathub Scalability, More Architectures To Come

[oiaohm]

Windows performs perfectly fine for me and billions of other people, I really don't care about missing a few microseconds from my life. I'm not even going to argue if Linux works better, I really don't care. Good on it if it does, Windows works just fine regardless.

For the use cases where linux wins it normally not by small margins.

And Windows is plenty secure. It actually has some extremely sophisticated security measures built directly into the kernel, though they are severely underutilized for the sake of allowing the layman to run whatever program they want, dangerous or not.

That is equally true for the Linux kernel this is your ignoring the Linux Security Modules. Reality is items like selinux are more sophisticated than what is built into the windows kernel. So this is another false arguement. Linux issue of weakness is in userspace with usage of X11 server.

Both OSes have had a history of absolutely horrible buffer overflow bugs in their networking code, difference is that MS couldn't patch them without breaking the OS whereas Linux doesn't have "versions" in the same way Windows did that it concerned itself with compatibility for.

This is a false flag arguement. Redhat and other parties like them does have versions in lots of the same way Windows does based off Linux and have been able to design fixed in their old LTS kernels they use for faults Microsoft have not been able to fix in Windows. There is a man power problem here.

Ironmask you have forgot Linus rule.

https://developers.slashdot.org/story/12/12/29/018234/linus-chews-up-kernel-maintainer-for-introducing-userspace-bug

No Linux kernel modification is allowed to break userspace. This results in some very complex and well developed work around code. This requires lots of developer time that Microsoft does not have. All network stack buffer overflow fixes are inside kernel space and under that rule of you cannot break userspace with Linux. In fact Microsoft is able to break this more often Linux kernel has to support programs using networking from 1993.

So Ironmask another bull crap arguement that is not in fact based in fact. Network stuff that Microsoft cannot fix when Linux kernel can is lack of developers problem with Microsoft like it or not. Linux requirements network area are just as strict as the Microsoft ones.

Linux is actually extremely insecure, but combined with the fact that nobody wants to target it, and software distribution is usually performed with a package manager and curated feeds, there's little issue there.

Those curated feeds in the Linux world have been adding Linux Security Module layers. Nobody wants to target Linux is not true either. There is a general point to be aware of most malware against Windows is using faults that are over 12 months old. Interesting enough same applies to Android. Most major Linux distributions have faults fixed inside 180 days. This window of exploitablity being smaller is a big one. Turns out making effective malware is not a 5 min job.

[Ironmask]

For the use cases where linux wins it normally not by small margins.

Okay? You know Windows outperforms Linux in some areas. That's not some petty argument against any OS. Linux performs some things better than Windows, Windows performs some things better than Linux. Hell, there's been titles of articles on this site saying Windows outperformed Linux. You know there isn't some magical universal algorithm for some things like task scheduling, if there was, then that's what we'd all be using. We use internal combustion engines because they're universally better than steam engines. I really don't know why you're so hooked up on this performance thing. You remind me of 4chan's /g/ back in 2010 when I browsed it, with nVidia and AMD fanboys constantly posting charts saying they outperformed each other. It's pathetic.

That is equally true for the Linux kernel this is your ignoring the Linux Security Modules. Reality is items like selinux are more sophisticated than what is built into the windows kernel. So this is another false arguement. Linux issue of weakness is in userspace with usage of X11 server.

Sorry for the misunderstanding, I was talking about the stock OS. Linux doesn't require SELinux, whereas Windows NT's security is literally architecture'd into it's very core. That's not an argument for or against anything, I'm just clarifying what I meant. SELinux is great security.

This is a false flag arguement. Redhat and other parties like them does have versions in lots of the same way Windows does based off Linux and have been able to design fixed in their old LTS kernels they use for faults Microsoft have not been able to fix in Windows. There is a man power problem here.

I mean, I didn't mean to conflate LTS distros like RHEL with "the spirit of Linux" like non-LTS Ubuntu or rolling distros like Arch. Most hobbyist/desktop users are using those whereas RHEL is more for business. My friend and I used to make fun of LTS distros like RHEL and Debian Stable for being extraordinarily out of date.

So Ironmask another bull crap arguement that is not in fact based in fact. Network stuff that Microsoft cannot fix when Linux kernel can is lack of developers problem with Microsoft like it or not. Linux requirements network area are just as strict as the Microsoft ones.

No, literally, the infamous networking bug in Windows 2000 that caused worms like Sasser were not able to be fixed because the fix would have broken how the networking layer worked. They fixed it in Windows XP, but they couldn't fix it in 2000 because it would have broken everything. You don't seem to know much about MS' history with fixing bugs in Windows. Is that just because you disregard the entire system as being not noteworthy to you? Because that's not very productive.

[oiaohm]

Okay? You know Windows outperforms Linux in some areas. That's not some petty argument against any OS. Linux performs some things better than Windows, Windows performs some things better than Linux. Hell, there's been titles of articles on this site saying Windows outperformed Linux. You know there isn't some magical universal algorithm for some things like task scheduling, if there was, then that's what we'd all be using. We use internal combustion engines because they're universally better than steam engines. I really don't know why you're so hooked up on this performance thing.

http://blog.zorinaq.com/i-contribute...an-other-oper/
There is more to it than that. There is a issue with Windows development.

Sorry for the misunderstanding, I was talking about the stock OS. Linux doesn't require SELinux, whereas Windows NT's security is literally architecture'd into it's very core. That's not an argument for or against anything, I'm just clarifying what I meant. SELinux is great security.

Linux Security Modules framework is in fact backed into the Linux kernel core. You don't have SELinux the LSM framework is still present. So security has come architecture into the Linux core its more of a question is it enabled or disabled. Yes just Windows NT design you can decide to build with security disabled as reactos developers proved for Windows.

By the way most distributions in fact build their linux kernels with multi LSM parts. Even that they are called a Linux Security Module its not in fact loadable so have to be built into the kernel image when its built. So most Linux distributions build many LSM parts like selinux,apparmor,TOMOYO then only configure one.

cat /boot/config-5.9.0-3-amd64 |grep CONFIG_SECURITY this is on current debian testing you will notice 4 LSM included with the framework.

I mean, I didn't mean to conflate LTS distros like RHEL with "the spirit of Linux" like non-LTS Ubuntu or rolling distros like Arch. Most hobbyist/desktop users are using those whereas RHEL is more for business. My friend and I used to make fun of LTS distros like RHEL and Debian Stable for being extraordinarily out of date.

https://store.steampowered.com/hwsurvey/Steam-Hardware-Software-Survey-Welcome-to-Steam?platform=linux

By the way your idea of Linux users is off by surveys. Over 33 percent of steam survey are LTS Ubuntu something. Turns out the stable release users make up over 70 percent of the Linux user base.

The reality is in networking they are not in fact that far different. The LTS kernel that Debian and RHEL have backported patches from the mainline kernel. Remember Microsoft has a Windows Insider version that is your rolling distribution. You have your normal retail versions that are like your debian testing. Then you have your windows enterprise version that are your Redhats. Microsoft has about 5 to 6 different Windows 10 kernels with complete userspaces on the go at any one time per platform.

No, literally, the infamous networking bug in Windows 2000 that caused worms like Sasser were not able to be fixed because the fix would have broken how the networking layer worked. They fixed it in Windows XP, but they couldn't fix it in 2000 because it would have broken everything. You don't seem to know much about MS' history with fixing bugs in Windows. Is that just because you disregard the entire system as being not noteworthy to you? Because that's not very productive.

No bad example and you are telling complete lies. MS04-011 (LSASS) vulnerability, which is caused by a buffer overrun in the Local Security Authority Subsystem Service( This is the fault Sasser Worm used). Has to be fault you are talking about right Microsoft did finally publish a patch for that for windows 2000 after the viruses were exploiting it. The fault was know over 12 months before the viruses and the fault was in fact fixable without breaking anything just required effort.

Its funny how often Microsoft says X fault is not fixable because it will break everything then virus appears exploiting it and magically their developers work out it is in fact fixable while breaking nothing. Putting in code to block a buffer overflow other than performance is 99.999% of the time harmless. The 0.001 percent times its not harmless it will be causing the OS to have random stability issues as well fixing it is still worth it. Its a insanely rare kernel buffer overflow flaw/exploit that is not in fact fixable invisible to applications and drivers the only known example happened in early Linux as in 1992.

There is a horrible fix in the Linux kernel today for a hard case for the userspace Synaptic touch driver that does expect to be able to overflow a particular buffer so the buffer is allocated larger for it so no more overflow. Yes this is a invisible fix results in more ram usage and the buffer overflow no longer going over anything exploitable.

https://lkml.org/lkml/2014/4/29/626

I can point to identical issue on Windows and Linux. espfix issue. Microsoft dropping win16 support from 64 bit windows was because they would not invest the time to fix it this was in fact admitted on the Linux kernel mailing list. Yes working out how to do espfix so 16 bit protected mode was safe on 64 bit systems was not simple.
.

[uid313]

It had monopoly for years and it still has such position. When comes to security it's insecure by design. It has backdoors introduced by MS itself, so stop kidding.

Windows does not have a monopoly, there are plenty of alternatives like Ubuntu, Fedora, openSUSE, Arch Linux, Manjaro, etc, people just don't want to use them. Despite Linux is free people still don't want it even when its free. They wouldn't even want it if you paid them to use it.

It is not insecure by design, it was designed by legendary Dave Cutler. In fact the design of Windows is more secure than Linux. Windows uses a microkernel while Linux uses a monolithic kernel. NTFS has a rich file permissions with ACL and inherited groups while Linux is limited to POSIX which is very basic and only has user, group and world.

Backdoors are conspiracy theories.

I wonder what drivers you're using? Maybe nvidia proprietary crap? There's a huge difference between Linux and Windows I/O performance. I hope you didn't miss Phoronix benchmarks..

I use the Intel open source graphics drivers to power my integrated GPU on the CPU.
I do not have any dedicated graphics card.

I don't notice any difference in I/O performance, perhaps it is visible during synthetic benchmarks, but when I just use the computer normally, both Windows and Linux are fast and responsive for me.

Windows if probably the most bloated OS ever. It's nowhere the speed of Fedora (Gnome) and Ubuntu is MS b*tch now, so I don't care. Windows 'usability' is a nightmare. It's inconsistent mess. For example it's a mix of old control panel and the newer one.

I don't know how you can say it is bloated when it is so fast. Windows boots, logs in, and shutdowns faster than Windows. The desktop responsivity is similar. My experience with Windows is that it feels fast and is not slow.

Windows usability is great and millions of end-users people use it everyday with great success and they use it to run their businesses. Windows has more usability research and a bigger usability team than any other system.
Yes, it does still have an old control panel that has not fully been migrated over to the new control panel, but it's not really a problem since normally people are not messing around with the control panel anyways. If this is Windows biggest usability problem then the usability of Windows is great!

Fanatics like Volta who say Windows is bad because it violates their arbitrary morality system and that's it's insecure "because I said so" put me off from bothering with Linux as a desktop OS. The Linux desktop community is so volatile and they never actually offer a real reason to use it. I mean, Linux is a wonderful OS, that's why I like it both as a server OS and in WSL/2, but, people like Volta never give a reason to actually install the thing on a desktop. They just continue to rant incessantly whining "why aren't you agreeing with me!?"

Yes, unfortunately the Linux community suffers from these radical fanatics. They insist that Linux is great and has no flaws and that Windows is terrible.

I use Linux as my desktop operating system at home, and Windows at work. I can recommend that you try out Linux on the desktop, I use Ubuntu. I do like both Ubuntu and Windows. I don't strongly favor either. Neither is really any better than the other for me. Both are just operating systems, which I find mainly similar, but a bit a different. It's like having a green t-shirt and a blue t-shirt.

There are some things I like with Linux, some things I like with Windows, overall it really does not matter. I mostly use Linux at home out of habit, because its what I've been doing, and it works good for me, but I wouldn't mind using Windows it's nice too. I use Windows at work, because its what everyone else does, but I wouldn't mind if used Linux, it's nice too.

[oiaohm]

It is not insecure by design, it was designed by legendary Dave Cutler. In fact the design of Windows is more secure than Linux. Windows uses a microkernel while Linux uses a monolithic kernel.

Lets start by ripping another stack of lies to bits. Windows started off designed to use a microkernel but then changed to what is called a hybrid putting everything in one single address space to gain performance. Then proceed to weld everything into ntoskrnl.exe. The reality is Dave Culter gave up the all the security advantages of a microkernel for performance like it or not. The Linux kernel is not a pure monolithic either https://www.kernel.org/doc/htmldocs/...per-setup.html a pure monolithic does not have kernel drivers spawning usermode helper programs. Windows starts with a microkernel plan and rapidly drops most of it getting all the faults of a monolithic. Linux kernel starts out monolithic and over time has been growing more microkernel features because it makes sense from security.

NTFS has a rich file permissions with ACL and inherited groups while Linux is limited to POSIX which is very basic and only has user, group and world.

Next Windows NT was not designed on NTFS at first this is important. Windows NT 3.1 would in fact install on fat and if you wanted NTFSv1 you had to convert from fat to NTFS.v1 its quite a few versions of Windows before windows could straight up install on NTFS. When you get right into the core of NT design the rich file permissions with ACL and inherited groups are in fact added on not in fact designed into Windows NT core as a lot of fools will attempt to have believe. Your base permissions of Windows is fat single user single set of permissions that is it. It was later versions of windows that went to having ACL on all the time.

Does Linux have permissions past the POSIX ones the answer is hell yes. There is a thing you have that is called xattr that allows quite a bit more.
First you have the Posix ACL this allows exceeding the user, group and world limit this is stored in the system namespace of xattr of the file system,

https://help.ubuntu.com/community/FilePermissionsACLs

Yes ubuntu does not turn Posix ACL on by default but you can turn them on.
Next you have your mandatory access control permissions from the likes of SeLinux these are in fact stored in the security namespace of the xattr.

uid313 what looked at Ubuntu with what comes out the box and thought that was all there was.

https://help.ubuntu.com/community/FilePermissionsACLs

Basically the out the box is not all their is with Linux permisions.

The reality with Linux is the base posix permissions are not absolute horrible. Windows NT stuck to fat level of permissions is absolutely horrible so windows has to auto step it self up one level.

Here the horrible reality windows is coming basically out the box set on its maximum. Most Linux distributions come out box with permissions options set to min and for a lot of usage cases this fine. There is reason for this processing complex ACL hinders IO performance. If you are not in fact need inherited permissions having the processing for that on is hurting your performance. Remember windows your choice is Fat level permissions or NTFS level permissions that is not a great deal of choice.
This is the Linux permissions choices.
1) Posix file permisions (yes the user,group and world)
2) Posix file permisisons + posix acl( this added inherited and lot more slots than user,group and world)
3) Posix file permisionns + posix acl + LSM MAC permissions(these are your mandatory access control setting windows does not have these)
4) Posix file permissions + LSM MAC permissions

The 4 one is interesting you look in a normal file manager as a normal user and the LSM MAC permissions are totally hidden heck even as root they be totally hidden that right LSM like selinux have the rights to hide information from root. The posix acl on gives fairly close to what NTFS does. But the LSM MAC permissions go well past.

There are quite a few Linux distributions that have Posix file permissions+ LSM MAC permissions from selinux on out the box. Fun right Linux can have a way more comprehensive permission system than Windows enabled in this combination and 95 percent of it is hidden from the user. I am not kidding on the 95% part the selinux mandatory access control permissions can do a hell load of things.

Of course a lot of people would think this was all. Its not so far that been data access control.

https://www.howtoforge.com/how-to-manage-linux-file-capabilities/

Then we have file capabilities that are permissions that are only for executable under Linux as well. This again is stored in xattr the system namespace. Most Linux distributions these days are using file capabilities all the time.

I don't notice any difference in I/O performance, perhaps it is visible during synthetic benchmarks, but when I just use the computer normally, both Windows and Linux are fast and responsive for me.

For me it comes out in big renders and code builds. It also comes out in krita when doing some 4K and above image work or using particular transformations options guided by pen(yes you can be putting in a hell load of I/O operation requests quickly). I would agree a user only ever doing light work on their computers will most likely never notice. But use doing heavy work we down right do notice.

[RahulSundaram]

Only when libraries have the same versions. Which is unlikely over time.

Nope. Libraries have nothing to do with it and it doesn't have to the same version at all, it is file level deduplication.

[Volta]

Windows does not have a monopoly, there are plenty of alternatives like Ubuntu, Fedora, openSUSE, Arch Linux, Manjaro, etc, people just don't want to use them. Despite Linux is free people still don't want it even when its free. They wouldn't even want it if you paid them to use it.

It does. Just like Android on mobile market (ignoring iOS for now) and Windows have nothing to say out there. Most of the people grew up on Windows, because it has monopoly and most of the software is designed for Windows. It's the main reason people use it over Linux. Linux kills it in terms of performance, security and usability.

It is not insecure by design, it was designed by legendary Dave Cutler. In fact the design of Windows is more secure than Linux. Windows uses a microkernel while Linux uses a monolithic kernel. NTFS has a rich file permissions with ACL and inherited groups while Linux is limited to POSIX which is very basic and only has user, group and world.

I don't care about Dave Cutler or any other idiot who ever worked on this shit. Good to know you have no idea about kernel differences. Windows is insecure by design and it's a fact. You're seeing much less viruses now thanks to Linux that keeps internet clean. I hope you remember old times when connecting to internet could be bye bye to Windows? It tells everything about its 'security' and kernel.

Backdoors are conspiracy theories.

Windows security is a conspiracy theory. Furthermore, M$ begs to differ:

https://privacy.microsoft.com/en-us/privacystatement

In order to provide this computing experience, we collect data about you, your device, and the way you use Windows. And because Windows is personal to you, we give you choices about the personal data we collect and how we use it. Note that if your Windows device is managed by your organization (such as your employer or school), your organization may use centralized management tools provided by Microsoft or others to access and process your data and to control device settings (including privacy settings), device policies, software updates, data collection by us or the organization, or other aspects of your device. Additionally, your organization may use management tools provided by Microsoft or others to access and process your data from that device, including your interaction data, diagnostic data, and the contents of your communications and files.

And a lot more about windows spyware:

https://www.geckoandfly.com/25083/free-tools-disable-stop-windows-spying-tracking-you/

I don't notice any difference in I/O performance, perhaps it is visible during synthetic benchmarks, but when I just use the computer normally, both Windows and Linux are fast and responsive for me.

Then you have chess reflexes. Linux I/O and scalability performance is superior to Windows.

I don't know how you can say it is bloated when it is so fast. Windows boots, logs in, and shutdowns faster than Windows. The desktop responsivity is similar. My experience with Windows is that it feels fast and is not slow.

It's slow shit, takes ages to load (I load it normally).

Windows usability is great and millions of end-users people use it everyday with great success and they use it to run their businesses. Windows has more usability research and a bigger usability team than any other system.

Rather availability of software, because its usability is an unholly nightmare.

Yes, it does still have an old control panel that has not fully been migrated over to the new control panel, but it's not really a problem since normally people are not messing around with the control panel anyways. If this is Windows biggest usability problem then the usability of Windows is great!

No, it's one of the less serious problems.

Yes, unfortunately the Linux community suffers from these radical fanatics. They insist that Linux is great and has no flaws and that Windows is terrible.

Windows 'community' suffers from unaware and naive people.

[Ironmask]

Volta, as stated before, your entire argument boils down to "I disagree with you because I do not like the thing you like." You've deluded yourself into thinking a piece of software is somehow evil, you think it's insecure without stating why (well, it seems more like you just really want it to be insecure). You're desperately trying to drive politics into something inherently not political and attempting to make reality bend the way you want it to. And the saddest part is that nobody's insulting the thing you like, everyone here is saying both Windows and Linux are great, and they are, and they have their own pros and cons. And then you come back insisting we're wrong because you want us to be.