tosim wrote: Wed May 19, 2021 4:33 pm
wiak-Do you use the full ArchOS, or one of the other versions of it?
WeeDog Arch64, but it is build using archbootstrap so the build is official Arch - except I don't use full install, but via WeeDog special overlayfs initrd/init it becomes a frugal installed Arch system, which is capable of being used in a very flexible modular way since WeeDog provides 100+ overlayfs layers capability via a special _addlayers algorithm that loops round finding the sfs modules rather than fixed adrv, fdrv, zdrv sort of scheme used by Puppy linux:
viewtopic.php?p=21859#p21859
The init further allows the sfs modules to be stored in boot partition, or anywhere else in the filesystem, or even in the initrd itself (or a mix of these) with all the usual changes save partition mechanisms to be expected by a frugal install (some being worked-on/improved at this very minute).
The key to all this is that simple, but really flexible _addlayer approach to loading in filesystem layers though. Per extract from above post:
In simplest to understand form (maybe) WeeDogLinux's _addlayers algorithm is basically (in rough english) as follows (but with lots of extras to deal with many very flexible save persistence options):
Code: Select all
Make a directory in tmpfs for mounting layer filesystems to.
while not all files done:
Look through the directory where the filesystem modules are
looking for .sfs module filesystems and raw module filesystems.
For each one found, mount it ready for use in unionfs layers.
Keep track (variable or array) of each found.
end_while_loop (i.e. keep doing the loop till finished)
Sort all the results found above in the order you want them in the layers.
mount the overlay (be it with aufs or overlayfs) to a tmpfs directory in RAM using the stored list of filesystem modules found above.
The resultant overlay is simplified to use numeric sfs layers and with that additional idea being part of it end looking like:
Code: Select all
overlay ...,lowerdir=save_sfs00:06:05:04:03:02:01,...work
In earlier WeeDogLinux creation (a few years ago), I had the idea to use chroot to the main root filesystem (so early WDL designs use choot in that way) rather than use the usual switch_root (such as traditional Puppy uses), however I now prefer switch_root with more flexibility available through using that alongside the _addlayers algorithm additions. I guess some people, who don't like to acknowledge others, think that if an 'intellectual idea' is simple and elegant enough then they can 'synthesise'/translate-it/modify-it/take-it-freely without credit. Wrong - should withdraw their publication despite the differences it has. I consider such actions a measure of Puppy no longer working well enough and some credit-grabbers trying to seek attention as if all-knowing (disrespectful) Saviours.
I've seen the idea blatantly copied recently, so don't be surprised to find the basics of it turning up in some Pups - it actually comes from WeeDog design, whatever may be claimed (I have heard it said that imititation is the sincerest form of flattery, but giving due credit for inspiration is a minimum form of honesty in acceptable open source development really). Anyway, I've learned from that and don't publish so much or so openly or so quickly/regularly as I did anymore - I wasn't impressed by such translation/modification/publication - I have no respect for any dev work from some quarters therefore, no matter how intensive, good or bad that dev work may be overall. Disgraceful, very low behaviour. I used to work in a postgrad research group (ERG, Univ of Aberdeen), developing TCPIP protocols/RFCs in pre-public-Internet days, so am well versed in spotting and avoiding plagiarism since even a whiff of it invalidates research publications - there is a serious code of conduct involved by competent researchers. Certainly the algorithm is simple, but wasn't used here in Puppyland by others before that I know of, and was certainly very well documented for WeeDogLinux in recent times; yes, simple/elegant algorithm, but good and worth respectfully copying/modifying/simplifying for simpler use and whatever, with due acknowledgement and reference. Anyway, not something I'll ever forget and I'm not an idiot in terms of being able to read code (whether recoded, translated, modified, whatever) and recognise copied ideas - really not good. Users don't care about such things, I understand that, but as a developer who puts hundreds of hours into thought and design ideas, I do. But I'll say no more about it - my memory of the event will live on regardless. Unacceptable, disgraceful, and pathetic - that's all except that I'm still contemplated publishing and contributing nothing further in the future - not on this forum at least - though my behind-the-scenes dev work will continue for my own purposes, albeit much more slowly and with significantly less enthusiasm. WDL init remains pretty much unique, however, and not some modified/code-translated initrd/init from any other source (though full credit is given elsewhere when applicable), however, and I'm pleased with that, despite any uncredited cherry-picking.
But yeah, aside from all these frugal install features WDL_Arch64 is fully Arch compatible so everything in Arch Wiki, Arch repos (including AUR) applies. However, if you want a much much smaller pacman capable Arch system, it is perfectly possible to tweak the firstrib rootfs build script for that purpose (well, actually, for that Arch slimming down purpose, you would tweak the archbootstrap script, which, unlike say debian's debootstrap, is a very simple single script).
wiak
