jaeger changed the topic of #crux to: CRUX 3.7 | Homepage: https://crux.nu/ | Ports: https://crux.nu/portdb/ https://crux.ninja/portdb/ | Logs: https://libera.irclog.whitequark.org/crux/
ukky has quit [Quit: leaving]
ukky has joined #crux
ukky has quit [Quit: leaving]
ukky has joined #crux
ukky has quit [Client Quit]
ukky has joined #crux
woland has quit [Quit: Leaving]
hestia has quit [Quit: Quit]
guido_rokepo has joined #crux
lavaball has joined #crux
lavaball has quit [Remote host closed the connection]
lavaball has joined #crux
stoffepojken has quit [Ping timeout: 246 seconds]
dim44 has quit [Quit: Leaving]
<SiFuh_> farkuhar: The address is very much not normal. The only one they can find is possibly here 5°25'47.0"N 100°18'33.0"E. Google maps says 19 but it isn't 19, it is 15. (19, Berjaya Hwy, Pulau Tikus, 10350 George Town, Penang)
<SiFuh_> If it is this area, even though that house looks like crap. The house prices average 3.3 million MYR so just over 700K USD.
<farkuhar> SiFuh_: thanks for the update. If the lawyer was just as careless typing the address, as she was in composing the Malay legalese, then I'm not surprised she got the location wrong.
<SiFuh_> farkuhar: I still find it interesting she used lot numbers
<SiFuh_> Still 1/12 of 700K is 59K USD ;-)
<farkuhar> Well, the selection of this lawyer wasn't up to me. The other inheritors living in Penang chose to employ her services, so I just went along with their decision.
dim44 has joined #crux
<SiFuh_> I am assuming they are not Malaysian?
<dim44> Is there a command to clean up older version port builds?
<dim44> *as in CRUX older version
<jaeger> Take a look at prtsweep or prtwash
dim44 has quit [Remote host closed the connection]
dim44 has joined #crux
dim44 has left #crux [#crux]
dim44 has joined #crux
<dim44> jaeger:Thanks, will do, still ironing out the system
sajcho has joined #crux
<jaeger> No problem
<jaeger> Another thing you could do is change PKGMK_SOURCE_DIR and PKGMK_PACKAGE_DIR in /etc/pkgmk.conf to locations outside the ports tree which are easy to prune later, etc.
sajcho has quit [Quit: Client closed]
<dim44> How do the prtsweep and prtwash behave for port directories that don't have an rsync/httpup? As in it's just a folder where I have my custom ports.
<jaeger> They have dry-run or test modes, try it and see :)
<dim44> I did run it, just making sure
<jaeger> I suspect they don't care where the ports dir is but I'm not 100% sure
<dim44> okay I'll some more automated testing just in case
<dim44> Ah, okay they are bash scripts as usual and as always they are pretty readable. nice!
sajcho has joined #crux
sajcho has quit [Quit: Client closed]
<cruxbot> [contrib.git/3.7]: python3-cleo: updated outdated build logic
<cruxbot> [core.git/3.7]: libmpfr: update to 4.2.1
<cruxbot> [core.git/3.7]: ca-certificates: update to 20230822
<ukky> What/who schedules/runs pwck in the system? I just got local email to root@localhost that user 'avahi' does not exists in the system. avahi was installed, then removed yesterday.
<jaeger> It's part of shadow and in /etc/cron/daily
<jaeger> That is to say it's a daily cron job installed by the shadow port
<ukky> thanks. Never heard of this functionality.
<ukky> jaeger: how is your RISC motherboard doing? I don't remember whether it was SPARC or MIPS.
<jaeger> It's RISC-V rather than SPARC... and I haven't had much time to mess with it yet
<jaeger> SPARC/MIPS
<farkuhar> hestia isn't here today, but jaeger might be interested in this follow-up to yesterday's discussion: https://git.sdf.org/jmq/prt-get/raw/branch/mixed-upinst/doc/i3-softdeps.test
<ukky> did you plan to port CRUX onto MIPS arch?
<jaeger> I don't have any MIPS hardware
<jaeger> At least I don't think so
<jaeger> For clarification, SPARC and MIPS are RISC architectures... and so is RISC-V
<ukky> That is why my mind spins... I know MIPS (and I have one device with MIPS CPU), I have heard SPARC, but RISC to me was just a generic term, which includes even MicroBlaze CPU.
<jaeger> RISC covers a lot of architectures, easy to get mixed up
<ukky> Does RISC-V gcc toolchain exist? Wiki says gcc supports OpenRISC architecture.
<joacim> been thinking about getting one of the sifive boards as a toy
<joacim> but, expensive toy
<ukky> jaeger: so, you are pretty much setup to install CRUX on your device :-)
<jaeger> The hardest part of getting a single board computer up and running is usually figuring out the bootloader, heh
<ukky> That is my main field of work currently
<jaeger> In general or with specific hardware?
<ukky> Specific hardware. But it already includes i386, MicroBlaze and Arm architectures.
<ukky> We use Coreboot/uBoot to boot Linux
<ukky> But if OS is not required, we directly load firmware
<ukky> I am soooo interested in porting CRUX to Arm devices. But current crux-arm approach does not work for me. I would like to create a build system on X86-platform to cross-compile CRUX-ARM.
guido_rokepo has quit [Quit: guido_rokepo]
<jaeger> You should talk to pitillo and see if maybe you can share work
<jaeger> I admit I don't have a great understanding of SBC boot processes, and it seems hard to find that info in some cases
<jaeger> I also find cross-compiling to be rather dense and hard to understand, past simple things like kernels
<ukky> Devices I have to work with are not that powerful, so cross-compiling is a must-have. Sometimes it takes up to 6-8 hours on powerful build system to cross-compile full root-FS.
<jaeger> When I've tried to cross-compile systems I always run into issues with figuring out how to get the proper libraries and locations set up
<ukky> Regarding CRUX for ARM, I have seen all git repos available on crux.nu. CRUX assumes you are compiling on actual CPU, besides initial ARM toolchain.
<ukky> All cross-compilers setup private include and library path. I can build software using multiple cross-compilers simultaneously. Except when using cmake, which can only use single cross-compiler.
<cruxbot> [compat-32.git/3.7]: gst-plugins-base-32: updated depends on line
<cruxbot> [compat-32.git/3.7]: pipewire-32: 0.3.77 -> 0.3.78
<cruxbot> [contrib.git/3.7]: p5-business-isbn-data: 20230811.001 -> 20230822.001
<cruxbot> [contrib.git/3.7]: dysk: 2.7.2 -> 2.8.0
<cruxbot> [contrib.git/3.7]: inotify-tools: 3.22.6.0 -> 4.23.8.0
<cruxbot> [contrib.git/3.7]: p5-mozilla-ca: 20230807 -> 20230821
<cruxbot> [contrib.git/3.7]: pipewire: 0.3.77 -> 0.3.78
<cruxbot> [contrib.git/3.7]: python3-poetry: 1.6.0 -> 1.6.1
<cruxbot> [contrib.git/3.7]: python3-flask: 2.3.2 -> 2.3.3
<cruxbot> [contrib.git/3.7]: python3-phonenumbers: 8.13.18 -> 8.13.19
<cruxbot> [contrib.git/3.7]: python3-tox: 4.9.0 -> 4.10.0
<cruxbot> [contrib.git/3.7]: python3-websocket-client: 1.6.1 -> 1.6.2
<cruxbot> [contrib.git/3.7]: weechat: 4.0.3 -> 4.0.4
<cruxbot> [opt.git/3.7]: python3-wheel: 0.41.1 -> 0.41.2
<cruxbot> [opt.git/3.7]: orc: build orc-test as its needed for gst-plugins-base
<jaeger> There's also https://github.com/crux-arm/crux-arm-release which I believe is the current "way to go"
<jaeger> How do you check/know where a particular cross toolchain's library path is? And is that a sysroot or something else?
<ukky> One of the ways: x86_64-pc-linux-gnu-gcc --print-file-name=libgcc.a
<ukky> This prints: /usr/lib/gcc/x86_64-pc-linux-gnu/12/libgcc.a
<ukky> And this: aarch64-pc-linux-gnu-gcc --print-file-name=libgcc.a
<ukky> Prints: /usr/lib/gcc/aarch64-pc-linux-gnu/12/libgcc.a
<ukky> I might be wrong, but the location where cross-toolchain will look for include and library files must be specified at the time cross-toolchain is built. That is how Linux-From-Scratch builds initial (clean) LFS toolchain on the host distro.
<ukky> On my main development system I have 14 gcc compilers installed, which can be used simultaneously (using appropriate prefix).
<ukky> Gentoo uses sysroot system for cross-compiling, AFAIK.
ukky has quit [Ping timeout: 252 seconds]
ukky has joined #crux
<ukky> jaeger: https://github.com/crux-arm/crux-arm-release assumes that host system is ARM, i.e. it cannot cross-compile CRUX-ARM root-FS on a non-ARM CPU.
<jaeger> Correct
<jaeger> I just mentioned it because you said you'd seen repos on crux.nu but didn't mention the github repos, that's all
<ukky> understood
<jaeger> er, crux-arm.nu
<ukky> Yes, I've got pretty full image of crux-arm.nu. Basically, downloaded much every diff-file and recreated git repo locally, as crux-arm.nu has no git/http/https protocol enabled for the repo.
<ukky> crux-arm.nu has a Makefile which builds cross-toolchain for Arm and Aarch64 on a non-Arm CPU.
<jaeger> Yeah. I don't know if it's maintained
<ukky> There is no point in maintaining cross-toolchain if developers got initial bootable ISO for ARM. Devs just switched to native ARM gcc.
<cruxbot> [contrib.git/3.7]: docker-compose: tweaked build command to fix version output
<cruxbot> [contrib.git/3.7]: cgit: edited post-install
<cruxbot> [contrib.git/3.7]: graphicsmagick: fixed footprint
<r0ni> cross-compiling is eww
<r0ni> my solution for arm is just use M1, then i have no headaches
<jaeger> M1 is nice for ARM, yeah. I don't have anything fast for RISC-V, though
<jaeger> Got plenty of fast x86-64 hardware
<ukky> Oups, what is M1 :-)
<jaeger> apple silicon
<r0ni> oh ya, other arches are always a PITA, tbh debian has a good setup for that stuff
<ukky> Ah, never used Apple products
<ukky> Are there non-Apple, non-laptop ARM, desktop/server systems?
<jaeger> There are definitely arm server systems, don't know about desktop
<ukky> I once had to do some BIOS-level code for ARM-based server prototype with 160 CPU cores, but that was pretty expensive system.
<r0ni> ukky look up pine64
<r0ni> i have a arm64 laptop called the pinebook pro, it's a budget system similar to a rpi4 in power
<ukky> r0ni: will check pine64
<r0ni> oh you said non-laptop lmao
<r0ni> well they make other hardware as well
<r0ni> they have the rockpro series as well
<jaeger> I have a non-pro pinebook, first generation. Wanted to like it but the keyboard is a garbage fire
<r0ni> jaeger this one is a bit better, but still suffers from sticky keys occasionally. I do like it though, it feels decent to use
<r0ni> crux zooms on it at least my optimized builds do anyway lol
<ukky> SiFuh_: $1,500 USD is too expensive as development board for home.
<SiFuh_> If I had a job I'd get it just to mess around with it
<r0ni> ok i'm supposed to be meeting ppl, so i'm going to head out... cheers everyone
<SiFuh_> ukky: Wonder how the Russians have progressed. They had some pretty good chips
<SiFuh_> Russia’s Elbrus
<jaeger> I employ *zero* hyperbole when I say the OG pinebook's keyboard is the worst one I've ever used
<jaeger> I've read that the pinebook pro's is far better
<ukky> There is Pine SBC with Arm-A53 (4-cores) for $39 USD.
<SiFuh_> Always liked the old USSR Cold War and Russian tech
<ukky> SiFuh_: I prefer older products too, but only if made in USA or Western Europe.
<SiFuh_> German engineering ;-)
<ukky> German products are good. Not sure about enrineering :-)
<SiFuh_> I like their heavy machinery for industrial and farm use
<SiFuh_> But I am talking about old products.
<SiFuh_> Here in Asia it is all crap. Spend little money, you buy a little turd. Spend big money, you get a big turd. Hahaha
<ukky> Recently bought small vacuum made circa 1980, just because it looks good.
<SiFuh_> Most of my equipment in Australia pre-dates the 50's. Still going strong
<farkuhar> Didn't Toyota teach us that inviting the folks on the manufacturing floor into the management discussions leads to improved engineering practices?
<SiFuh_> Not really, because they still put the diff breathers on the top of the diff.
<SiFuh_> If they listened to the folks they'd put it a position a lot higher
lavaball has quit [Remote host closed the connection]
j0hnny has joined #crux
cruxbot has quit [Ping timeout: 240 seconds]
cruxbot has joined #crux
Stx_ has joined #crux
Stx has quit [Ping timeout: 612 seconds]
Stx_ is now known as Stx
tilman has quit [Ping timeout: 246 seconds]
tilman has joined #crux