<fossy>stikonas: ? did you test using chroot or qemu mode <fossy>but for chroot mode it should discount the /usr/bin/guile checksum <fossy>because its not deterministic in chroot mode (some kernel difference) <stikonas>but we still need to calculate the correct one now <fossy>the match code has not been committed.. <stikonas>but I think that should now let us do things like if match $ARCH x86; then <oriansj>stikonas: your mescc-tools-extra patch has been merged. <oriansj>stikonas, fossy: are there any cases where we actually do kaem -f file1 -f file2 -f file 3 ? ***ericonr- is now known as ericonr
***theruran_ is now known as theruran
<oriansj>vagrantc: thanks you for the patch for mescc-tools. sorry for taking so long to notice it <vagrantc>i'm used to things taking 3-12 months or more in other projects, this seems like fine turnaround :) <oriansj>vagrantc: messaging me here, would have gotten same day review and merge ^_^ <stikonas[m]>Kaem has a return code bug... Will make PR later too <stikonas[m]>But it's less critical there since we always want 0 return code there <stikonas[m]>Well, full kaem now has conditionals, so return code became more important ***stikonas_ is now known as stikonas
<stikonas>with updated checksum for /usr/bin/guile <Hagfish>is the intention for live-bootstrap to continue by preparing the other requirements for guix, or should that be a separate project/module? <stikonas>any distro should be able to get those built binaries and use them to bootstrap itself <stikonas>just as live-bootstrap takes over after stage0-posix <stikonas>so you can have different distro plugins <stikonas>(not sure if they should live in the same repo) <fossy>so the answer to that Hagfish is no, first i want a complete GCC toolchain, and then we can build Guix (and Debian, and Fedora, etc) off of that <stikonas>well, we just need to get newer GCC toolchain <fossy>s/complete/complete and up-to-date <stikonas>fossy: by the way, kernel is not checksumed? <stikonas>I guess it includes initramfs and timestamps <stikonas>yeah, you didn't add any checksum file where linux is built <stikonas>although, getting checksum is a bit annoying <fossy>i think i'll just sha256sum it and just take the output of that in the console.. <fossy>like just temp add it to the script <stikonas>yeah, but if initramfs is checksumed, that will fail <fossy>i don't think we want initramfs to be checksummed <stikonas>hmm, kernel does not include initramfs in it's image? <fossy>how did you figure out how to fix that <stikonas>cause I wrote it correctly in kaem-minimal for risc-v <stikonas>so when I saw this, I realized that it's wrong <fossy>how do you know that its & 0xff00 and >> 8? <stikonas>mixed it up and combined with WIFEXITED... <fossy>also, we probably do want glibc in live-bootstrap before we are "done" <fossy>will make it much easier to bootstrap most distributions which will need glibc <stikonas>hmm, possibly, although it shouldn't really matter <fossy>yes, but bootstrapping glibc distributions from musl will be more of a pain than it is worth <fossy>oh, this would not be too painful to move to glibc, because we are still using static binaries <fossy>i think i will look at autogen for now <stikonas>oh also those bits and shifts for exit status are defined in man 2 waitpid <stikonas>not sure how we can run it with guile 3... <fossy>stikonas: i'm just looking at any potential bootstrap path <oriansj>stikonas: merged your mescc-tools patch for kaem <stikonas>oriansj: thanks, it should now work with at least match executable <stikonas>I just realised that other exit codes might still be broken when kaem is built with M2-Planet <stikonas>but I think "match ${ARCH} ..." might be the main use-case <stikonas>I might update stage0-posix with those new checkouts...