<oriansj>rickmasters: my current task is to first create the tools for creating and reading a bootstrap filesystem in the M2-Planet C subset. Then to write a program to be the bootloader in the MBR. Then finally a kernel (assuming someone else doesn't decide to pick it up to help speed things along. <oriansj>as for why mescc; it seemed far more logical than expecting janneke to write mes.c in hex and for me to write another C compiler capable of building TCC when mescc could do the job. <oriansj>It might have not been the optimal solution and it was a great deal of work to get there but it works and until someone puts in time to do something that replaces it, it is what we have to work with. <stikonas>but right now ASM is more familiar to me <littlebobeep>Ohhh well I thought you worked on ASM code for example and you said you were not goot at scheme <stikonas>that doesn't mean easier, just that I might have invested more time in learning it <stikonas>I worked on riscv support (so asm code too) for stage0-posix <littlebobeep>ja I knew you did riscv but I didn't know if you contributed ARM or x86 work too <oriansj>and asm is very easy to build a predictive model of in your mind <stikonas>you still have "functions" just you need to add some extra stack work before and after <oriansj>littlebobeep: well how much do you know about assembly language programming? <littlebobeep>I find it quite difficult to understand compared to high-level languages <oriansj>but even without knowing that if I showed you load r0, [r1+next] could you guess what it would probably do? <achaninja>if you have programmed a lot of C you can get by in assembly by just transliterating <achaninja>give each variable a stack location, use call and leave + ret for functions <achaninja>i guess that is sort of what happened with the high level prototypes included with stage0 <achaninja>w.r.t. zstd in that thread above, I can vouch it is very fast <achaninja>far faster than xz for compression and decompression by something like an order of magnitude ***lukedashjr is now known as luke-jr
<bauen1>oriansj: being bribed sounds like getting funded :D <unmatched-paren>precisely :) we were noting that it could be incredibly easy to _look_ like you're subverting stage0, but actually you're just taking the money :D <fossy>stikonas[m]: not really any good reason for links in separate file <oriansj>bauen1: well effectively it has to be, otherwise you wouldn't have the time/effort required to get it in a form that wouldn't be immediately flagged and cause everything you have ever done to be more deeply audited. ***furrymcg1e is now known as furrymcgee
***jjardon_ is now known as jjardon
***dnm_ is now known as dnm
<dongcarl>oriansj: haha that was a great read, not a bad way to exploit the handler's relative ignorance <muurkha>achaninja: I didn't realize zstd was comparable to lz4! ***stikonas_ is now known as stikonas
<stikonas>fossy: I might have a way to sort out those symlinks in live-bootstrap packages <stikonas>or at least it worked till findutils was built and we switched to different tar command (which then crashed for me0