IRC channel logs

<matrix_bridge><Jeremiah Orians> Well yes, at the end of the day. People are just building paths from wherever they want to start to wherever they want to go and sharing them with others. There is no such thing as a single true bootstrap. Just collections of ways that we know work and the lessons that we collectively learned from finding/making them

<matrix_bridge><Jeremiah Orians> The journeys are more important than the outcomes; as if any of them disappeared. Making a replacement becomes easier over time as various people carve out their own paths and ideas.

<matrix_bridge><Jeremiah Orians> It is only a matter of time before someone forks gcc 4.8 and runs with it

<matrix_bridge><Jeremiah Orians> Or tcc or etc

<aggi>it's several been decades until a complete system bootstrap was recovered

<aggi>and some of it's crucial pieces can't easily be replaced (builder0/stage0/M2/mes-libc/pnut/tcc)

<aggi>most important, bootstrapping criteria aren't arbitrary

<aggi>i am a little worried still, because the size and complexity of live-bootstrap is overwhelming

<aggi>which is related mainly to WHAT is bootstrapped there (gcc, latest kernel, autotools)

<aggi>meanwhile it is possible already to keep running a complete GNUish/POSIX development host with linux-tcc kernel

<aggi>which is the opportunity to scale down a fully bootstrapped development host orders of magnitude

<aggi>in essence, that is a linux2 + busybox + tcc-toolchain (and a few other pieces: yasm, cdrtools)

<aggi>probably the most boring *nix ever

<aggi>with the latest tcc-0.9.27/pnut-cc patchset there is a good chance to cleanup the tcc/musl re-compilation passes

<aggi>and rolling along with tcc mob/HEAD directly for this

<aggi>meanwhile i've received another extortion letter here, cold-appartment rent hike

<aggi>zero response to any of my inquiries at employers, universities etc. for some economic prospect to sustain a living

<aggi>it's not that much work remaining to finalize tiny-bootstrap fork (with the initial package set relevant to bootstrappable

<aggi>the other ~500 ebuilds ported and tested for linux-tcc/2.4/musl will have to wait)

<aggi>although there is a few pieces i want to ship along with this, tetex for example, some ghostscript/gnuplot (all of which was anything but trivial to see that pass compile-time and run-time again with musl/linux-tcc)

<aggi>after all, i had applied at some universities too, where one or another math-head might show interest in these things

<aggi>and i want to keep things in order, sane and simple as possible

<xentrac>different people have different bootstrapping criteria, whch indeed are not arbitrary

<xentrac>condolences on the rent hike. I just suffered one a couple of months ago

IRC channel logs

2026-08-05.log