Call for Foundation-supported Project Ideas

Suggestion: When the swap is nearly completely used (less than 2 GiB left), awaken the OOM killer early and make it send SIGTERM 5 seconds before sending SIGKILL (until the amount of free swap becomes greater than 2 GiB). If the swap becomes completely used anyway, go back to the default behaviour and make the OOM killer send SIGKILL without delay. Bonus point: allow the sysadmin to configure those two values (2 GiB of swap, 5 seconds delay) in a config file.
 
I favor

• anything that improves the flexibility and ease of use of security features

• anything that improves the flexibility and ease of use of jails

[The py-focker port looks interesting.]

• a more up-to-date handbook and wiki

On the FreeBSD wiki, there are 29 Jail Management Tools!! "Some of them can be used only on older versions of FreeBSD (4.x / 5.x) ". Which ones is not shown.
 
grahamperrin said:
… add an idea that was not expressed (here) before the summary was published?
Click to expand...

Two or three things repeatedly drifted into, then out of, mind. More than anything: GParted.

forums.freebsd.org

Gparted-like program for disk analysis

Can anyone suggest a FreeBSD program similar to Gparted which can analyse a disk in terms of partitions and filesystems. I have a disk where I am unable to identify the filesystems on the partitions, and don't know how to identify them...
forums.freebsd.org forums.freebsd.org

The idea, here, is not solely for analysis.

It's not something for which I could write a proposal at this time.
 
Looking forward to funding decisions (some time in March, I guess).

jrm@ or anyone at the Foundation: how many proposals were received?

Just curious …
 
This is my proposal :

After 1 month of research we have found the technical reasons why a modern Nvidia GPU if passed through inside a Windows 10 / 11 vm produces the error 43 (actually the error 12). It happens because it misses "line interrupts support for passed through devices" ; actually there is the need of a massive change inside the bhyve source code. As you probably know,this change is not a priority for the developers. To achieve the goal will be a very step forward for bhyve and for all the freebsd community. Let me know your thoughts. thanks.
 
grahamperrin said:
Looking forward to funding decisions (some time in March, I guess).

jrm@ or anyone at the Foundation: how many proposals were received?

Just curious …
Click to expand...

For this call for proposals, we only received a few applications and are considering one for funding. This was fewer responses than for past calls, but anecdotally it's recently been difficult to fill positions for skilled developers across the tech industry. Other than these responses, we do have a few pending applications, one from a company that we have had good results with in the past. Other sponsored FreeBSD work continues: wireless, OCF/Wireguard, and the pull request for RAID-Z expansion looks like it will be merged into the OpenZFS repository soon.

We send these calls out periodically to remind the community that we are keen to fund FreeBSD work (thanks to all your donations) that may be difficult for volunteers to complete. Deadlines are often good motivators, but just because the deadline has ended, doesn't mean that we won't still accept good proposals. When skilled FreeBSD developers or documentation writers are available with good proposals, we are ready to work with them.
 
I am not sure if these suggestions can be considered in the Project Ideas, but here some that will need some funds…

Better HID driver support​

Well, enhance the new HID subsystem, as it can be quite buggy, even with ~16 years old USB keyboard that used to work fine… (Moreover I am not alone in this kind of situation, take a look at the forum).

Make INIT(8) faster​

INIT(8) make any system start/reboot very slow. If there is a way to reduce the boot time, no matter how many time in a year you reboot your machine, it should be done.

Having to reboot for any reason your server (to get ride of zombie processes or whatever) takes too many time, especially for a server that host critical services and applications…

We are no more in 90 or in 2000; now things have to run fast, as people wants things right now.

Yes INIT(8) works fine, but it is very slow. And if “working fine” means no change, well stop working on anything trying to accelerate changes in the operating system, as it just works. Do not try to bring more people to the BSD hold, and just let FreeBSD die…

I insist, I write here MAKE INIT(8) FASTER (through parallelism?), not replace it.

Better workstation experience​

Increase efforts on laptop and desktop usage of FreeBSD. Indeed, we can have over 30,000 open source software packages that are easy to install, but some are outdated, others are buggy or broken, while others are missing…

FreeBSD as laptop is almost impossible, and some users experience some lags with their desktop computer, even with a descent machine. Maybe, somebody should really investigate on the causes of these issues.

Here, I have a descent and recent PC (6 cores CPU, 32GB, relatively big GPU), and experience me too some lags even when listening audios, with no reason, while under Windows or any Linux distribution the system works perfectly fine.

Pkg​

Provide a better mirror management​

It can be very difficult to get a decent speed with pkg. It takes sometimes more than 1 hour to get llvm90-9.0.1_3.pkg downloaded (that was the case two days ago on this setup).

Unfortunately, tweaking /usr/local/etc/pkg/repos/FreeBSD.conf does not guaranty to have the fastest mirror (behind a vpn, I can download faster with an US mirror than the EU one)…

Add pkg new options​

search -file​

Provide a way to find files inside packages. When developing, we often have to use some libraries, especially when porting a software from a system to another. Packages that provide a library file have not the same name across systems, so I too often have to go to pkg.org to find out which package I have to install on FreeBSD.

-no-recommended​

We do not need everything that came with a package by default. According to our need, the Internet plan, and disk size, etc, it would be great to control on how to install packages, for example, by installing only required packages. Moreover, it will reduce the attack surface in such system. For now, there is no real policy, so some packages are well sliced, but other not.

Better “decoupling policy” in ports and packages​

Offer ports and packages with less superficial dependencies, particularly useful for limited system (among other, limited SSD size, raspberry):
  • for LibreOffice add in ports support for –enable-split-app-modules, and offer the installation of individual application (writer, cal, and so one) with pkg ;
  • not gluing Gnome with unnecessary applications like Cheese, Glade GTK demo, Print Editor…
  • TexLive in ports is old, if we download it from the official web, we do not need gigs of old TexLive that is glued with Gnome Latex…

Try to get more “languages supported”​

There are missing languages for the documentation (yeah a lot), as well as missing myths dictionaries, for language like Arabic, Hebrew, Turkish (yeah I know, these are not inside the OS "core"), etc… Maybe we should find a way to increase widely language support.

Jail​

Provide us unprivileged jail​

For now, we have to be root or in wheel group (via sudo) to create and to start a jail. Please provide us a way to do the same thing as unprivileged user. jailme allows to start a jail as unprivileged user, but not to create jail, like lxc-create.

Orchestration tools​

Some orchestration tools like lxd could be very helpful to manage lot of jails.

Better documentation and Wiki​

It will be very helpful to get a better documentation, as well as a better Wiki especially for new users.

Here some subjects that are not well documented:
  • how to get Internet access for jails, virtual machines, etc.
  • how to setup different kind of virtual network, for which use case;
  • how to use vale(4).
  • how to get Wayland working with any supported DE.

Try to have a better upstream cooperation​

Maybe by dedicating, upstream, some developer on some project, could be beneficial, as well as more “meeting” between “FreeBSD” and leaders of major projects too.

The goal? To get more abstraction layer upstream that will facilitate the work downstream for FreeBSD ports and package, and maybe to stop this Systemd assimilation, acting like Borgs, in more and more projects.

I see too often (to my taste) FreeBSD maintainers acting alone, even when being stuck on issue for months, without trying to reach the maintainers of the initial project (if ever)…

LibreOffice​

Why LibreOffice cannot be build directly from sources while there is some work already done? Why we need to add more patch downstream?

Eclipse and other IDE​

To be competitive as a workstation, FreeBSD must have more Opensource IDEs. For now, the choices are pretty limited, and even Eclipse is outdated and a real pain to get work outside the port. Atom is missing too, and very painful to build from upstream sources. It probably contributes to the fact that BSD are used by only 0.18% developer, when Linux distributions are used by 25.32% of them, competing with MacOS.

Host some ports inside FreeBSD​

Make some critical ports and packages (like Firefox, and so one) run not by benevolent, but by some FreeBSD developers. Maybe through a vote, or after some consultation, decide which package will be chosen. Yes, it means to maybe increase the FreeBSD "core" responsability.

"The Power to Server", a failure?​

The Power to Server what? How? Whom? When?

Try to understand why FreeBSD keep failing to attract people that prefer to go to Gentoo or ArchLinux, and the like, while these distributions often do a bad coping/pasting of FreeBSD (if only by choosing Systemd). Yes they have a better hardware support, but what lead Intel and others to start ignoring FreeBSD and to keep ignoring it?

Now FreeBSD is lagging even in server field. Again, why almost all web hosters propose only some Linux based distributions, and not FreeBSD (it hasn't always been the case)? What these Linux based distributions do better?

Yes, I know, the foundation have increased marketing support, and try "to increase the awareness and use of FreeBSD", but there is a big OLD issue there. Even Deb Goodkin “now” sees Linux as a threat for FreeBSD.

Finally! 15 years ago, FreeBSD fanboys mocked me when I was saying and writing that Linux was a threat. The most stupid of them keep answering that Linux is only a kernel, FreeBSD is not a desktop OS, and do not need to evolve…

Anyway, if nothing is done, FreeBSD will probably die, soon or later, as it will deem useless.

Whonix, QubesOS, Tor browser​

Indeed, why Whonix choose to use Debian as a base, 10 years ago? Why Qubes Os did not choose FreeBSD but Fedora, 10 years ago? Yet, at that time, Debian and FreeBSD were pretty equivalent.

FreeBSD is supposed to be more secure, yet it was not chosen. There is obviously something wrong…

Why Tor browser is not available for FreeBSD?

Why “big projects” for those privacy and security count do not choose FreeBSD as their base or add support for FreeBSD?

Kill the Mascott​

Finally, maybe more subjective, I really thing that FreeBSD should discard its mascot that looks like Elmer Fudd looking at himself experiencing satanic acid trip.

Yeah, I hate it. I disagree with Robert Watson (with all due respect it should let marketing people working on that issue). FreeBSD need a new mascot or simply get ride of it, as it is tacky, and stink old fashion.

Did you ever asked people how they find him? I have been told by young IT people that they find him crappy, and pretty representative of the redneck FreeBSD users…

Pay survey if needed, as a bad image will not attract people and moreover young people, toward FreeBSD.
 
twoface Your list contains items that are strange, false, have nothing to do with what FreeBSD has control over and personal preference. pkg is NOT slow. I've used FreeBSD as my workstation for nearly 20 years. And any attempt to make FreeBSD a personal computer for the average user is misguided.

btw, Beastie, the current version, was created by John Lasseter of Pixar. Maybe you've heard of him. He offered me a job once.

But posts on lists like this happen every month and we have to put up with it as we push forward with ease and get real work done.
 
hbsd said:
There is nothing special about Tor Browser
Tor Browser is just: Firefox + Tor

[HOWTO] use Tor network and web proxy
Click to expand...
sorry for being offtopic again, but I have a strong need to educate users.
aaaand again: NO, IT IS NOT! Tor Browser is a modified version of Firefox! If you use your normal firefox behind tor you can pretty easily be traced.

forums.freebsd.org

Tor Browser

Hello everyone, I have been looking for Tor Browser (torproject.org) for x86 and ARM. I've noticed a few old threads but not actually any solutions. Any experience or suggestions how to get Tor browser on FreeBSD? Thank you.
forums.freebsd.org forums.freebsd.org
 
twoface Your list contains items that are strange, false, have nothing to do with what FreeBSD has control over and personal preference.
Click to expand...

You are too vague…

pkg is NOT slow. I've used FreeBSD as my workstation for nearly 20 years.
Click to expand...

Well, I found on the forum at least one topic about pkg being too slow… Maybe you are in some area where pkg works fine, but seems not being the case for everyone.

And any attempt to make FreeBSD a personal computer for the average user is misguided.
Click to expand...

Never wrote that, but why this would be misguided?

btw, Beastie, the current version, was created by John Lasseter of Pixar. Maybe you've heard of him. He offered me a job once.
Click to expand...

Ok, so no need to redraw Superman, nor Batman. Yeah, let's keep their crappy old design from 30's in 2022.

But posts on lists like this happen every month and we have to put up with it as we push forward with ease and get real work done.
Click to expand...

Wow, what arguments…

@hbsd what you wrote is just dangerous.
 
rootbert said:
sorry for being offtopic again, but I have a strong need to educate users.
aaaand again: NO, IT IS NOT! Tor Browser is a modified version of Firefox! If you use your normal firefox behind tor you can pretty easily be traced.

forums.freebsd.org

Tor Browser

Hello everyone, I have been looking for Tor Browser (torproject.org) for x86 and ARM. I've noticed a few old threads but not actually any solutions. Any experience or suggestions how to get Tor browser on FreeBSD? Thank you.
forums.freebsd.org forums.freebsd.org
Click to expand...

rootbert said:
If you use your normal firefox behind tor you can pretty easily be traced.
Click to expand...
Yes you can always be traced even with tor browser. tor is not secure as you think it is:

Is Tor Really Anonymous and Secure?

Unfortunately, there is no such thing as 100% security. You are at risk even if you use JavaScript in your browser!
Can you name some of the features that made Tor Browser better than Firefox or even Chromium?
 
twoface said:
You are too vague…
Click to expand...
You list too many items this early in the morning. We tire of these repetitive lists posted by the occasional new user ad nauseam.
twoface said:
I found on the forum at least one topic about pkg being too slow
Click to expand...
You can find one of anything if you search long enough. In the meantime, pkg works great for the rest of us and I use it every day.
twoface said:
let's keep their crappy old design from 30's in 2022.
Click to expand...
Beastie was first created in the 1970s and updated in the 1980s. That it is old and not to your liking is personal preference and has no meaning to the majority who find him entertaining. Your statement is no reason to change it.
 
twoface said:

Better HID driver support​

Well, enhance the new HID subsystem, as it can be quite buggy, even with ~16 years old USB keyboard that used to work fine… (Moreover I am not alone in this kind of situation, take a look at the forum).
Click to expand...

It describes many of the improvements in the HID system in FreeBSD 13.0, and it describes how HID functions on NetBSD. I think it will be a lot closer by the time of the next FreeBSD version. I bet it can be improved on from there. It's a matter of time before the foundation is layed for getting multimedia keyboard keys and other devices to work. Once that happens, people will start getting more functions and devices to work, and start documenting those steps.
twoface said:
I insist, I write here MAKE INIT(8) FASTER (through parallelism?), not replace it.
Click to expand...
When I used linux, there were different run levels, and that was confusing. The only way I can see parallelism, is if instead of being numbered runlevels, each run level is separated by the type, for instance: a runlevel for Internet start up, while the other runlevel is labeled desktop, one for filesystems, and another runlevel for everything else starts up starting at the same time. Firewall startup tasks may be better inside an Internet runlevel, that starts up in the section before the net comes up.
twoface said:
Add pkg new options
Click to expand...
There are many package and ports management utilities in ports, it has its own category of ports-mgmt. I use portmaster.
twoface said:

Better documentation and Wiki​

Click to expand...
Use the How-to Guides section in the forums and other documentation. The official wiki in my opinion took a step back from what it was before. From what I understand, editing it became easier. I've tried before, and one day will try it with the new system. What's really discouraging is what it takes for commits, and how in comparison Bugzilla for unmaintained ports doesn't get addressed.
twoface said:

LibreOffice​

Why LibreOffice cannot be build directly from sources while there is some work already done? Why we need to add more patch downstream?
Click to expand...
For office software: the best way requires a company to get behind it. Apache OpenOffice makes more sense for companies to get behind, but this one has dependencies on Java. Also, this is something that would be better of all of the BSD's got together and collaborated on, and I don't see this until there's a native or improved graphical toolkit. Maybe QT5 or Lumina will do in the meantime. LibreOffice is good for now.
twoface said:
Finally, maybe more subjective, I really thing that FreeBSD should discard its mascot that looks like Elmer Fudd looking at himself experiencing satanic acid trip.
Click to expand...
It's odd to me how the Beastie mascot wears Converse and ends up advertising for that brand. Any mascot should avoid any clothing brand. I know feet are difficult to draw, and it often comes out wrong. In video games and cartoons, they tend to draw the feet and hands of characters correctly. If it's going to wear shoes, they can be changed to something more generic. Keep the mascot, except loose the current shoes.


Each item needs to be refined so what it takes is shortened and so nonessential wants can be removed. Learning about what already exists, and as much about how it works. Coming up with something that makes it easier and what works with the bulk of what's already there, functionwise. A lot of learning in the How-tos and other places can reduce a few needs on a list, and reduce the requirements to get them. It's about what's essential, and what makes it easier yet better for implementation of better features. The list being unrefined seems to offer more problems than solutions.
 
twoface said:
Ok, so no need to redraw Superman, nor Batman. Yeah, let's keep their crappy old design from 30's in 2022.
Click to expand...
Absolutely the next generation should design and implement something new and "modern".

If you think you can produce a design that is popular and captures the legacy of FreeBSD, who knows, it might be accepted by the community. For example, I am not perfectly content with the strange red sphere with horns that has been adopted as the official logo.
 
You must log in or register to reply here.
Back
Top