00:09:25 <zedmatrix> ok building curl and wget and openssh, next kernel and then need to copy the grub uefi info over
00:42:22 <nsITobin> zedmatrix: https://gitlab.com/binaryoutcast/snowball-os/baseos/-/commit/117725c520d6fe187c39bbb4cb98573067eca27f
00:45:38 <zedmatrix> and you can call those from an rpm spec file ?
00:46:10 <nsITobin> ... i suppose I could but it would be nuts
00:50:35 <zedmatrix> oh so you are just bash building ch5+6 ?
00:50:54 <nsITobin> i am moving back and forth
00:50:58 <nsITobin> refining 
00:52:02 <zedmatrix> but once you build the chapter5-7, you can save that and use that as a base for building from its only about 130Mb or so
00:52:51 <zedmatrix> i believe that is how most distros do that is start with there basic setup
00:53:23 <nsITobin> i have the rpmbuild part stage3 where i want it to make spec files
00:53:25 <zedmatrix> atleast gentoo does, you start with a bootable iso, download a stage3 build tar and build
00:53:39 <nsITobin> i am now making stage 1 and 2 less annoying
00:54:12 <zedmatrix> thats y i broke my auto build into 6 parts
01:00:55 <zedmatrix> this auto build is almost done, just building the kernel 
01:02:58 <zedmatrix> if malina was around it would nice to ask about the /sbin/init as a script
01:07:20 <malina> well shoot?
01:07:51 <malina> one can specify init=/bin/sh say on the boot line
01:08:31 <malina> the scripts in lfs are precisely that, they don't use any auxiliary binaries [almost], except for devices. and so on. but are called from the init script
01:08:42 <malina> what is the question zedmatrix ?
01:08:46 <zedmatrix> well i know you should mknod console and null?
01:09:36 <malina> ye, hmm, I do forget a bit how to do that. I recall having a bit issues myself when i was 'mixing' some inits, so t so speak. but using eudev or udevadm thingie should bring them up
01:09:39 <zedmatrix> the kernel looks for /sbin/init, if that was a bash script to create all the dev nods, instead of sysv, rc systemd
01:09:46 <malina> but yes, in the init scripts I think you write in the mknod/null somewhere 
01:09:50 <malina> to preppoulate some stuff
01:10:18 <malina> well init is usually just calling in more scripts (or binaries) . but ye, you coulod write one LONG init in one file or split it up, (sysv style)
01:10:37 <malina> ye, you can reeplace syv, with your own init
01:10:46 <malina> which would have to take care of the various instnatiations.
01:10:49 <zedmatrix> just wondering if you have ?
01:11:06 <malina> well, I .. I did something like a hybrid from sysv + superman + eudev
01:11:20 <malina> I do recall I had to hammer a lot befor getting it going , esp on musl I think
01:11:38 <malina> hmm, I had on musl at least issues with devpts or /dev/ptmx not coming up properly iirc
01:11:45 <malina> and had to piss about a while before i got it all going
01:11:46 <zedmatrix> yea i was looking at my musl , used sysv+mdev+skarnet
01:11:54 <malina> ye mdev is 
01:11:56 <malina> ye, sounds good
01:12:01 <malina> is that working ok then?
01:12:09 <malina> if so, what is the issue? 
01:12:12 <malina> or not?
01:12:19 <zedmatrix> yea had no issue building musl+gcc15
01:12:23 <malina> (a tad semi afk because I am trading right now)
01:12:32 <malina> ye, but init wise
01:12:33 <malina> I mean
01:12:59 <zedmatrix> yea some simple funcs https://github.com/zedmatrix/ZBuild-System/tree/main/LFS-gcc15-musl/stage4/sysv-init
01:13:06 <malina> erm, what abt lfs, it has a init script, once can adjust no?
01:13:34 <zedmatrix> i wanted to trim it
01:13:57 <malina> yu0p, so looks like something i did too initially
01:13:58 <zedmatrix> kinda took some from the lfs sysv book 
01:14:11 <malina> I kinda used lfs scripts when i first started off, then used superman, my init, 
01:14:33 <malina> and kinda somehting in between etc. to be honest, when id di the musl I redid a loty but my muisl died out some years ago and so it's still a kinda WIP
01:14:35 <malina> not entirely done
01:14:55 <malina> and since then I haven't been working on those things becaue ye some months ago id di the bootstrap from the small asm file
01:15:03 <zedmatrix> there wasnt much needed to patch
01:15:07 <malina> and been mostly buy then to do marv 2
01:15:37 <malina> ye init isn't hard to do, i mean it's mainly to bring up the physdevs and that's about it
01:16:05 <malina> have you checked out suckless.org zed?
01:16:17 <malina> I mean, it is a trove of info for making things small, succint and easy
01:16:26 <malina> any lfser should really peruse that too
01:16:35 <malina> less is more
01:16:42 <zedmatrix> i remember looking at it briefly
01:17:11 <malina> dwm is a very nice simple gui env
01:17:16 <malina> mdev might be from it?
01:17:25 <malina> I mean my old musl used toybox as the userland
01:17:40 <zedmatrix> i wrote this http://0x0.st/8nsm.sh to simplify LFS 
01:17:41 <malina> which itself is quite challenging due to so much non posix commands out there
01:20:02 <zedmatrix> but its fun moving a rootfs onto a usb to boot :)
01:20:37 <malina> ye vm,s usb, anything is easy with linux 
01:21:17 <zedmatrix> especially if you start using uuids
01:22:01 <zedmatrix> but without the testsuites lfs can be done in an afternoon :)
01:31:02 <nsITobin> zedmatrix: wish me luck!
01:32:25 <zedmatrix> rpmbuild?
01:32:34 <nsITobin> no stage1 with changes
01:32:40 <nsITobin> if i am wrong.. my system dies
01:32:43 <nsITobin> :P
01:33:03 <nsITobin> .. let me make sure i still have my backup i took a week ago
01:33:35 <nsITobin> cd lfs
01:35:56 <zedmatrix> aren't in a vm ?
01:36:17 <nsITobin> why should I be
01:36:20 <nsITobin> i have a backup
01:36:52 <nsITobin> and fedora cinnamon on a flash drive which can install wimlib and restore if it goes bad but i can do one more test before i pull the trigger
01:38:18 <zedmatrix> i also have extra pc's around to play with :)
01:38:43 <nsITobin> zedmatrix: looks correct to me https://dpaste.org/ieJLT/raw
01:38:46 <nsITobin> I am poor
01:38:58 <nsITobin> send me money.. paypal.me/mattatobin
01:39:01 <nsITobin> :P
01:40:04 <zedmatrix> i dont have money :)
01:41:24 <zedmatrix> as long as mosy things are prefix'd /mnt/lfs and $LFS is properly export'd things for stage1 should work
02:51:14 <nsITobin> zedmatrix: https://gitlab.com/binaryoutcast/snowball-os/baseos/-/commit/f94c8b4493e7db25e3c0844e629658632731f235
02:55:38 <nsITobin> specifically my chroot script is slightly improved and primed to accept path and command as arguments so i can start it doing something like stage2 and stage3 after stage1 from ./mach build.. later
02:55:44 <nsITobin> https://gitlab.com/binaryoutcast/snowball-os/baseos/-/blob/TRUNK/build-aux/lfschroot?ref_type=heads
02:59:01 <zedmatrix> at the chroot_pre you can just "mkdir -pv $LFS/{dev,proc,sys,run} " its smart enough to make or not :)
02:59:56 <zedmatrix> or in your case "${TARGET_SYSROOT}" nsITobin
03:01:59 <nsITobin> if they don't exist it is an error as it would be created by the filesystem script
03:08:10 <zedmatrix> you can just mkdir -p 
03:08:36 <nsITobin> and if the thing ain't mounted
03:08:40 <nsITobin> then i have to clean it up
03:08:44 <nsITobin> and try again
03:08:54 <nsITobin> or if i enter the wrong path
03:09:30 <zedmatrix> you can check with mountpoint
03:09:56 <nsITobin> for empty directories?
03:10:37 <nsITobin> i am just not understanding why i would want to make more work for my self in case of erRor
03:12:07 <nsITobin> i just like how the amount of hardcoded paths is greatly reduced
03:12:41 <nsITobin> right now the lfschroot script will not proceed unless the target sysroot is properly prepped and the target command exists
03:12:53 <nsITobin> previously i just let it error out
03:12:57 <nsITobin> naturally
03:13:06 <zedmatrix> i was talking about this https://gitlab.com/binaryoutcast/snowball-os/baseos/-/blob/TRUNK/build-aux/lfschroot?ref_type=heads#L24
03:13:47 <nsITobin> what about it.. if those directories do not exist it should not continue
03:14:30 <nsITobin> Check whether a directory or file is a mountpoint.
03:14:40 <nsITobin> how does that help me determine existance
03:15:37 <zedmatrix> i just mean pre create them 
03:15:47 <nsITobin> i do..
03:15:56 <nsITobin> as the second script in stage1
03:16:03 <nsITobin> long before it would be invoked
03:16:05 <zedmatrix> when i finish option 1 of my build, i just chroot
03:16:19 <nsITobin> and i want to use this script for more than one thing
03:16:20 <zedmatrix> it creates all vfs 
03:16:34 <nsITobin> and i want it NOT to create things unless it is either existing or prepped
03:17:38 <nsITobin> https://gitlab.com/binaryoutcast/snowball-os/baseos/-/blob/TRUNK/stage1/tasks/1500-filesystem?ref_type=heads#L37
03:19:35 <zedmatrix> i do that as part of my chapter 7
03:19:51 <zedmatrix> including the files
03:20:01 <nsITobin> yeah i found it redundant to create directories twice for a system that won't even exist after the rpms are built
03:20:39 <nsITobin> stage1 and stage2 have no existance after the last rpm is built
03:20:46 <nsITobin> just gets deleted
03:22:13 <nsITobin> that is why I can cheat so flagrantly.. exclude buildability exclude like a dozen packages
03:22:55 <zedmatrix> finished building the laptop to the fast 24gb ssd
03:24:49 <zedmatrix> and tested my packBuild for the grub-uefi files
03:33:50 <zedmatrix> hmm looks like i can build lua earlyer like right after readline, nsITobin
03:36:58 <nsITobin> the only reason I am building so much in stage2 is because i haven't the time to work out how low it can go
03:37:25 <nsITobin> i decided a full buildchain rpm deps then rpm was a fine starting point
03:37:50 <zedmatrix> well most of those will be needed for whatever post-LFS you build
03:38:34 <nsITobin> most of those are needed for building ANYTHING AT ALL WITH RPMBUILD
03:39:36 <zedmatrix> i am still looking at not building rust, sofar Mesa's nouveau seems to be the only one
12:04:00 <nsITobin> zedmatrix: well i majorly fucked up my stage1 last night .. not hostwise just exec wise.. hopefully NOW i have fixed it correctly
12:04:27 <nsITobin> this makefile is a piece of crap tho i am getting sick of it
12:05:35 <nsITobin> until i replace it it will remain fragile
13:30:02 <nsITobin> zedmatrix: works again https://ibb.co/B2nhHJ4Y
14:15:24 <nsITobin> ./mach build
14:15:25 <nsITobin> https://gitlab.com/binaryoutcast/snowball-os/baseos/-/blob/TRUNK/mach?ref_type=heads#L115
14:16:29 <nsITobin> who needs fedora meta-tools
14:16:39 <nsITobin> i have rpm bash and mah brain
14:34:43 <zedmatrix> nsITobin: eventually you'll come to the dark side :)
14:35:11 <nsITobin> zedmatrix: been there and back
14:35:21 <nsITobin> I have a better way.. or at least a simpler one
14:58:27 <nsITobin> zedmatrix: how different ARE the instructions between sysv and systemd
14:58:40 <nsITobin> aside from configuration files and bootscripts
14:58:50 <nsITobin> and udev being included of course
14:58:59 <nsITobin> like do packages get built differently?
14:59:08 <nsITobin> cause I'd love to simply make init system a choice
15:11:06 <nsITobin> zedmatrix: how different ARE the instructions between sysv and systemd
15:11:09 <nsITobin> aside from configuration files and bootscripts
15:11:13 <nsITobin> like do packages get built differently?
15:11:15 <nsITobin> cause I'd love to simply make init system a choice
15:12:20 <zedmatrix> not too diff, but i think you would have to have 2 separate rpmspecs
15:13:15 <zedmatrix> the sysv has a couple of extra packages sysklogd and sysvinit << those are handled by systemd
15:14:29 <nsITobin> well if the configuration is only different in a few other packages.. then a simple %{?_systemd} check would do
15:30:03 <zedmatrix> some of the packages have configuration changes like --with-systemd
15:32:53 <zedmatrix> cool just made a mini zget that parses my zbc files'
15:33:57 <zedmatrix> nsITobin: really just pulled the download/parser engine out of my zbuild 
15:34:46 <nsITobin> the bash scripts them selves and the crappy makefile don't REALLY matter .. but being able to do em from a single command is nice
15:34:55 <nsITobin> it is what I been working towards
16:07:57 <nsITobin> https://gitlab.com/binaryoutcast/snowball-os/baseos/-/blob/TRUNK/mach?ref_type=heads#L77
16:08:09 <nsITobin> a bit ghetto but does the jorb
16:19:24 <zedmatrix> nice 
16:27:42 <nsITobin> $ ./mach
16:27:42 <nsITobin> ./mach Usage: clobber | build [stageN] | hostsh [command] | chroot [path [command]]
16:28:12 <nsITobin> what more do you need
16:28:16 <nsITobin> :P
16:38:52 <zedmatrix> cool just made an alsa-package directory with sources
16:40:08 <zedmatrix> like this http://0x0.st/8nqQ.png nsITobin
16:54:21 <zedmatrix> nsITobin: the first test of my alsaPack.tar worked!
16:54:38 <nsITobin> and my scripting is busted again
16:55:37 <zedmatrix> i just have my packBuild source the extract .tar
16:57:10 <zedmatrix> like this https://bpa.st/7GGA
17:01:42 <zedmatrix> i just haveto make sure the stage_build list is in the right order
17:02:26 <nsITobin> there fixed it
17:02:43 <nsITobin> my bash looping and arg skills are still not that strong
17:02:51 <nsITobin> few dozen more will likely do it :P
17:03:37 <nsITobin> now trying a complete build to the prelem rpms in stage3 then will either.. fuck around with stage2 scripts or simply go back to writing spec files
17:04:23 <nsITobin> i am pretty happy with how this is shaping up
17:05:13 <zedmatrix> damn forgot to install lspci
17:13:30 <nsITobin> zedmatrix: if i -j1'd the entire build process it could slowly curn on one of my VPSes.. assuming linking anything doesn't take more than 4gbs
17:18:24 <zedmatrix> oops forgot i can just copy sources/* 
17:23:20 <zedmatrix> hmm i guess maybe a packInstall > extract the pack... call packBuild ... ?? remove tmp
18:03:22 <zedmatrix> hmm ran into a zbuild.cfg issue when i pushd with the new packInstall
18:03:24 <zedmatrix> nsITobin: 
18:05:27 <nsITobin> so what does that mean for you
18:06:54 <zedmatrix> i think i need to define a location instead of current path
18:07:55 <nsITobin> what would be better is if i could redirect all output for each bash build script to a logfile
18:09:37 <nsITobin> zedmatrix: what i really am questioning is if i need to rebuild the ch7 packages before rpm.. would cut down on time
18:11:20 <zedmatrix> i think https://archlinux.org/packages/core/x86_64/glibc/ is the reason for chapter 7 isolation
18:11:51 <zedmatrix> building a clean perl and python for glibc
18:12:31 <nsITobin> yeah
18:13:39 <zedmatrix> if you used musl you wouldn't haveto build perl and python :)
18:14:09 <zedmatrix> then you would have the only rpm-musl :)
18:15:13 <nsITobin> does bison need rebuilt before I build the final bison?
18:15:19 <nsITobin> for instance
18:15:47 <zedmatrix> maybe .. https://archlinux.org/packages/core/x86_64/bison/
18:16:08 <zedmatrix> see the dep list
18:17:15 <nsITobin> static gettext vs shared
18:17:19 <nsITobin> on final pass
18:18:25 <zedmatrix> the first ch7 gettext isnt a full package
18:20:41 <nsITobin> but it is obviously enough to get by until after sed in ch8
18:36:44 <zedmatrix> yea its https://archlinux.org/packages/core/x86_64/bison/ that needs a rebuild
19:08:51 <nsITobin> gonna take a new image of stage1 state cause i deleted the old one .. gonna work on stage2 now which is dead simple.. ./mach build stage2 it chroots and runs make
19:11:16 <zedmatrix> just rewriting my finalize before reboot https://bpa.st/FGWQ
19:11:54 <nsITobin> cd ..
19:11:56 <nsITobin> err
19:13:32 <nsITobin> likely gonna rename the repo to just.. build
19:13:34 <nsITobin> LOL
19:13:53 <nsITobin> or setup
19:13:58 <nsITobin> would make it easier to clone
19:15:54 <zedmatrix> i think your build script is smaller.... https://bpa.st/4GVQ
19:16:24 <nsITobin> not if you take out colors and usage
19:16:51 <zedmatrix> i have 6 stages
19:33:40 <zedmatrix> ok now restarting and building to the 1Tb internal, with the updates nsITobin