IRC channel logs
2023-05-23.log
back to list of logs
<stikonas>janneke: sigh, so far fairly negative progress today... <stikonas>after some rebasing I can't even get riscv64 version of mes-m2 to work (it runs and exits with 0 status without doing much...) <stikonas>ok, found one issue, slightly better now but there is probably another <stikonas>janneke: anyway, for now I've replaced those ld !-0x10 with ld !-16... <stikonas>but somehow mes-m2 enters interactive mode after running script <stikonas>janneke: and also note that I've completely unified stage0-posix-riscv64 and mes defines <stikonas>but that means that you need the very latest M2-Planet <stikonas>hmm, probably something still broken with exit call... <theruran>is amd64 on stage0-posix finished? I am running Release_1.5.0 and it is stuck on: <theruran>+> ./AMD64/artifact/cc_amd64 ./AMD64/artifact/M2-0.c ./AMD64/artifact/M2-0.M1 <theruran>on my other machine which is musl-based, it actually completes but gives me mismatching hashes for several binaries <stikonas>and none of the musl stuff should matter <stikonas>theruran: maybe submodules are out of date? <stikonas>theruran: and in general I suggest running the latest git <stikonas>still it's a bit strange that it's stuck... <theruran>lemme try master HEAD again. yeah, i made sure to update the submodules <theruran>iunno. now it's stuck at +> ./AMD64/artifact/M0 ./AMD64/cc_amd64.M1 ./AMD64/artifact/cc_amd64.hex2 <oriansj>theruran: well that step is known to be slow (as it is only reading 1 byte per syscall) <theruran>OK it has moved on to +> ./AMD64/artifact/cc_amd64 ./AMD64/artifact/M2-0.c ./AMD64/artifact/M2-0.M1 <janneke>stikonas[m]: so, some good progress after all, nice <stikonas[m]>janneke: well, at least I should have umblocked mes-m2 build... <stikonas[m]>Though for mescc devel, running x86 version is faster <oriansj>theruran: honestly with just a handful of extra functions all of those early programs could be made much faster. But I traded speed for simplicity as auditing another hundred bytes of hex0 is just no fun. <stikonas[m]>Well, hex2 is written in hex1, so it's a bit better but still... <stikonas[m]>Still it's strange that it takes a long time. On my laptop native (non qemu'd) stage0-posix takes a minute or two, and early stages are just seconds... <theruran>it's still moving along, at least. but I swear last time it was stuck overnight <theruran>stikonas[m]: looks like it is write()'ing 1 byte at a time <theruran>it was much faster on my Void musl machine. I may have some kernel settings enabled here, in addition to disk encryption and hardened toolchain that may slow it down? <janneke>so, on a completetly (?) unrelated note... <janneke>rumpdisk is a server for the hurd, based on the netbsd rumpkernel <janneke>it brings modern disk support to the hurd <muurkha>"The figurative sense of "small remnant, tail-end of anything" derives from the notion of "tail" and is recorded from 1640s in reference to the English Rump Parliament (December 1648-April 1653), so called from sitting after the expulsion of the majority of members of the Long Parliament in Pride's Purge. As an adjective from c. 1600." <muurkha>thus we have rump sessions at conferences and rump kernels, I suppose?