At last and not least, a distribution that is truly different than most others and most of all clones. Fig OS has a vision that is fresh and possibly the way to do things in the future. One may call it a hybrid, if hybrids didn’t already have a negative connotation. It is best to allow the creator to speak about it on his own than for us trying to describe his work. Promised not to be talking about Devuan any more we are breaking our rule partially, we are allowing someone else do the talking, and Refracta is not officially Devuan anyway. And why did Devuan refused to add this to their list of clones? Did they ban the persons arguing they should? Hmmm….. …! Maybe this makes Fig OS even more interesting.
Using fig, I originally set out to create a program that produced html tables of Puppy Linux distro analysis– in other words, there are countless derivatives and I wanted to examine and output the differences.
Then I created a tool called “distdiff” which mounts two isos (including squashfs) and compares the contents by hashsum and filename.
From there I wanted to know if I could:
* Download two isos
* Use the binaries of a newer Devuan iso to update an older Puppy iso
* Put it back together into a bootable iso
This just-for-fun effort was called “refractahrpup”, and I didnt plan on creating a distro.
Now if youre thinking “Whoa! copying binaries like that is no way to update an iso” then you dont know Puppy. *Of course it is* (No, youre right. But it was a lot of fun.)
The results were inconclusive, but hopeful. I also used it to patch Refractas live iso against Dirty Cow, which I consider a small triumph for a newbie (installing Refracta and then running security updates also works, but I wanted the patch in the live iso.) PLEASE NOTE: this was a while ago and Refracta patched against Dirty Cow in a timely manner. I enjoyed getting there first.
My friend in the Puppy community suggested I drop “pup” from the name, and fsmithred didnt love the name either, so without other ideas handy I decided “Fig OS” worked at least as well as the previous name.
As I got as close to the original goal as I was interested in getting to, I dropped most of the Puppy side of Fig OS. It still allows installation of .pet packages, though its mostly apt-based and exclusively apt-based unless you want to do something weird like install a package from Puppy.
Besides formerly being two modified distros in a single .iso file, the two most unique features of Fig OS are:
Fig OS is created by a fully-automated remaster script. Its intended to be run (you might have to create a folder or something) and then “walk away”; So if you dont like one of the changes it makes, just fine the corresponding lines in the script, delete them or comment them out, and run it again.
This means that it is fairly practical to distribute Fig OS as a 60k plaintext program rather than a cd-sized iso, but allytonx has kindly archived the .iso files.
I think theres potential in this form of customisation and distribution, but Im sure it will be a a while before many people agree.
Fig OS also has a “root desktop.” I dont mean it runs IceWM (the default wm of Fig OS) as root; IceWM runs as user– in fact, everything runs with the same privileges it would under Devuan or Refracta.
But on top of IceWM as user, Fig OS runs a pcmanfm instance as a desktop for icons, so the icons you click (inspired by Puppy) will run as root– unlike Puppy this doesnt permeate the entire system, only the pcmanfm desktop. So you can have your “crazy” run-as-root business from desktop icons, and your hotkeys and IceWM menu will run things as user, as they probably should.
…At least until Debian deprecates its own menus, and lets SystemD take that over too. Which I hear is going to happen *sigh*
Thats ok, we will still have hotkeys, right? (Crap, now Ive said it– and next week hotkeys will be in pid 1 as well.)
On SystemD integration– shouldnt *everything* be in pid 1? I mean, from psctl it *should* list subprocesses as separate pseudo-pids, but shouldnt we reimplement all of /proc/ in SystemD? That way the only process is pid 1, and we can implement pseudo-pid 1 as a subprocess of pid 1– then we can have SystemD in SystemD so you can init while you init!
But I promise here and now, not to put the above-mentioned systemd-pid-tesseract in Fig OS. openrc and the like are possibilities I suppose, but I dont want to open a wormhole.
systemd-free linux community 
🙂 this isnt going to help her think we are two different people, but even if no one cares its nice to have this up here, thanks.
i should also point out that only on the galaxy forum was there any real resistance about treating it as a non-devuan derivative. i have not tried adding it to the without-systemd wiki. i have not bothered with the devuan page, so long as the galaxy forum is going to dictate what is and isnt a derivative with absurd prejudice. i was singled out for sure, not by the community but by someone on a power trip.
and that was sort of the last straw. i think its a good thing that i got treated that way– it helps me understand those with similar experiences. i still dont understand how a person becomes part of devuan. maybe this year we will find out. id rather be part of its derivatives– thats where a lot of the hope is. we do need repos though.
LikeLike
I didn’t / wouldn’t “promise to not talk about devuan”. The royal We peppered across blog articles here is offputting, is confusing, and quite frankly it smacks of mental illness.
LikeLike
fig, I hailed as “izzi” at unofficialdistros. I trust that you will not “out” my other personas (and, yes, I still have one embedded at dev1galaxy). As you will recall, that “friend in the Puppy community who suggested dropping pup from the name” begged out, decided not to continue participating at unofficialdistros due to sensing “too much negativity” in our posts there.
“And why did Devuan refused to add this to their list of clones? Did they ban the persons arguing they should? Hmmm….. …! Maybe this makes Fig OS even more interesting.”
typo: “why did… refused”
and
yet again, yeah, readers will wonder WTF? “posted by fungalnet” but the article is quoting (written in first person) something written by YOU? Wait, codeinfig _is_ fungalnet??? I’ve read your comment stating “no, not the same person”, just letting you know that until I read that comment… I too was sitting here wondering “same person?”
LikeLiked by 1 person
hey! incidentally the too much negativity guy wasnt the same guy as the one that made the fig os boot screen for me (who recommended dropping pup) but i think they both joined at one point.
nope, im not fungal but yes it now makes plenty of sense that you know each other. i havent made a new version of fig os since we most recently spoke, but there isnt a new release of refracta for quite a while either– now is the time where all the devuan girls and boys are playing with udev replacements and trying to figure out how to remove libsystemd again.
theyre not technically going slower than debian yet but theyre not catching up either– i sincerely hope that the selection of tools theyre going for will help, for every users sake. also for fig os.
i think it would be useful to have a page that talked about different things systemd/fdo has caused the deprecation of (udev, hwclock, sysvinit, debian menus– pending, *kit?) and what devuan has replaced them with.
given that antix has their own simpler workarounds, im equally curious what they have done to patch these things. the best approach for devuan imo would be to throw in antix-like workarounds before shifting to whatever longer-term solutions they think are better… when they werent 3 years in it made enough sense to say “well lets just fix these things permanently” but these “permanent” fixes take years. throw in some more patches, already!
LikeLike
Ok, now I am getting confused. With eudev being available as an easy drop-in replacement of udev, why would they still be working on udev? Why is that important to stick to udev instead of eudev? I think it was about 4 months ago that it became available together with OpenRC, I used it, and I never had a single problem with it.
There is a reproduction of an article by Steve Litt on here, it is making a distro into a bicycle and systemd into a bicycle’s braking system. Steve has a very valid point. If you keep running behind a fast vehicle you will always be out of breath and you will never catch up. Debian is going out of its way making it hard for those not using systemd to catch up with it. Siduction is proof that if you use systemd even Debian-unstable can be quite stable.
Besides all the bitterness about Devuan, and the recent discovery of AntiX, at this point I see the goal as fading away. Debian will never allow any clone to catch up to it. Especially all those that claim they are Debian without systemd.
LikeLike
last i checked, the eudev maintainer joined devuan. however, i think the problem with eudev (as usual) is that its a fix, not a replacement.
those keeping track might try to say “hey, you just said devuan takes too long with full replacements, not patches.” right– they should use more patches like eudev, until they have these full replacements.
eudev is not a full replacement, its a bandaid over the old thing. and if im wrong– and eudev will work long-term, thats good news. but thats not the impression that i got from before. the udev replacement for the long run is vdev. you know the guy that wrote that article you like, jude c. nelson? he is the vdev author. but i havent heard about vdev in a year, i assume theyre still trying it out as a replacement for eudev– which makes plenty of sense to me.
LikeLike
someone i dont really know tried fig os yesterday (he tracks distros a la distrowatch)
he says of fig os: “It’s lightweight, runs smooth, and comes setup with essential software like a web browser, photo editor, etc all ready to go.”
this is pretty much the first real review of fig os, so thats fun. my guess is that he ran it in qemu or virtualbox.
LikeLike