SiFuh has quit [Remote host closed the connection]
SiFuh has joined #crux-social
lavaball has joined #crux-social
ivandi has quit [Quit: WeeChat 4.2.2]
ivandi has joined #crux-social
<SiFuh>
remiliascarlet: Are you upset because you are a xenophobe according to that US White House Resident, Bidet?
dlcusa has quit [Remote host closed the connection]
dlcusa has joined #crux-social
<SiFuh>
ukky: You kept grub for for the ISO?
<SiFuh>
As in to boot the ISO
<ukky>
SiFuh: I did not change boot process for the ISO. Crux ISO uses syslinux for Legacy BIOS boot, and grub2-efi for UEFI BIOS boot.
<SiFuh>
ukky: I remember discussing it with jaeger a while ago. If I recall correctly SYSLINUX didn't support EFI in the old days
<ukky>
I might be wrong, but syslinux might already support UEFI boot. Though, I don't use it (yet), so cannot say for sure.
<SiFuh>
ukky: It does
<SiFuh>
However, isolinux may not.
<SiFuh>
Anyway, just curious. I have no real intentions to change the booting of the ISO
<ukky>
Switching to syslinux (from grub) in part of my very long todo list
<SiFuh>
ukky: My build was successful. Just installation had a two issues. Nothing serious. So now that I am in a full working musl environment, my plan is to reinstall CRUX musl from my ISO and reboot into CRUX musl and start the ISO rebuild again, but much cleaner.
<SiFuh>
ukky: Also, I discovered a port that I required that I never installed 'patch'
<ukky>
Probably I had to install 'patch' too. I might scan build logs to see what part of build has complained about missing 'patch'
<SiFuh>
Mine was building ports on a new system
<ukky>
In chroot environment?
<SiFuh>
It is under core in CRUX.
<SiFuh>
ukky: Oooh yeah, maybe
<SiFuh>
Must be because I am building the kernel now for new isntall
<SiFuh>
install
<ukky>
Also, I have noticed that 'stage 0' performs 'prt-get isinst <port_name>' on the build system, when it should perform that check for 'sysroot' filesystem.
<SiFuh>
Hmm
<SiFuh>
You are right
<ukky>
That is a bug in Pkgfile design
<SiFuh>
Ahh and my first ISO build failed with gcc and I remembered when I was building custom GCC on and ISO I had to set -j to 1
<ukky>
And a lot of Pkgfiles reference files on the build system to check if some package/port is installed.
<SiFuh>
But the other day I noticed the Pkgfile said only make instal to -j1 and not make
<SiFuh>
ukky: It is why I am rebuilding again on the clean musl install
<SiFuh>
I don't understand why -j1 is for install and not for bootstrap
<SiFuh>
I'll ask jaeger
<ukky>
It should be legacy 'make -j1'. We can check earliest available core/gcc/Pkgfile to make sure.
<SiFuh>
I just don't see why I'd use -j1 when I already compiled it.
<SiFuh>
make bootstrap compiled it.
<SiFuh>
make install just installs it
<SiFuh>
ukky: isinst should be removed from all ports you are building for the ISO
<ukky>
You are right about the sequence, but you are wrong that 'make install' cannot fail on a multi-threaded execution.
<SiFuh>
ukky: I know -pipe was removed
<ukky>
'prt-get isinst' should be okay if we add '-r <sysroot_path>' for ISO build and for cross-compiling.
<SiFuh>
ukky: mine failed during compilation. Was complaining that the C++ version of gcc couldn't compile. So I change it to -j1 and then it was fixed
<SiFuh>
ukky: I'd just skim through the ports we wish to build and see what they are poking around for
<farkuhar>
ukky: how about alias prt-get="prt-get --install-root=<sysroot_path>" in the stage0 environment, rather than editing the Pkgfiles themselves?
<ukky>
farkuhar: That is actually a great idea
<ukky>
Though, that alias alone will not save me from rewriting all Pkgfiles to accomodate for CBUILD!=CHOST, and for LIBC!=glibc, and INIT!=sysvinit
<ukky>
and DEVMGR!=eudev
<farkuhar>
Well, yeah, if the Pkgfiles were originally written under the assumption that eudev and glibc (and linux-pam ...) would be available, then obviously they need to be revised when swapping out (or removing) those components.