<oriansj>stikonas: if anything is wrong please let us know because I love learning out my mistakes ^_^ <stikonas>no, it was just me not too used to git submodules :D <stikonas>I thought git pull does git submodule update too <oriansj>I probably should add that detail into the notes so that other people don't make that same mistake <oriansj>It is a good example of a detail we could have done better and let us; fix the README so that other people can avoid that problem <stikonas>well, I guess in this case they are more likely to read README first <oriansj>stikonas: we always assume in good faith that people want to do the right thing <oriansj>especially since only people who care would dare bother to deal with bootstrapping <oriansj>and those sorts of people deserve care and consideration; as they are important to our future. <fossy>I swear I included that in the spanky new README. <oriansj>fossy: you did but not the command for updating the submodules after they have been updated. <oriansj>easy mistake to make, in fact I didn't spot it myself until stikonas pointed it out <oriansj>it happens, we fix and move on (I just hope my volume of spelling mistakes reduce over time) <dddddd>OK, so now I know what's really missing in M1.scm of slow-utils. Three functions indeed (and their unspoken dependencies). A couple notes: the constant for AArch64 is a bit weird (same as the mescc-tools C version, by the way). I'd change if you think is a good idea (AARM64 -> AARCH64). Also, default-toggle seems unbound and it's unused anywhere else so far. I don't see what's supposed to do looking at the C version. Is it a leftover? <oriansj>as we haven't done an official M2-Planet release with AArch64, the rename is fine. It is a leftover that can be pulled <dddddd>The external behaviour won't change. The command-line flag has the normal spelling. <dddddd>print-hex is the easiest one and I already started implementing it. <dddddd>In any case, if anyone else wants to work in M1.scm I'll step aside (I have other priorities right now and this is just for small breaks). <dddddd>Time for me to go. Happy coding! <dddddd>Be seeing you (I'm not a number! and all that... if you know the reference :D) <fossy>I found the issue... my mes-seed works... but the mes I recompiled refuses to work. ***ng0_ is now known as ng0
<dddddd>Another productive day documenting. Today about function call convention implementation with a focus on and stack usage. <fossy>wow TinyCC is a dumpster fire <dddddd>fails to work? difficult to hack? ...? <fossy>1. building it in a portable way is very difficult <fossy>2. it dosen't like x86_64 asm registers afaict <fossy>i'm intersted to know why we are bothering with doing things like porting M2-Planet to ARM when we can just build up on x86_64 and cross-compile from there <fossy>anyway we need to work on mes libc and tinycc support for x86_64 <dddddd>AFAIU to bootstrap in the platform itself <fossy>ah janneke your bootstrap.sh is x86 only (for tinycc) ***ChanServ sets mode: +o rekado
<rekado>stikonas: hey, thank you for the link. It’s good to know that this shorter route works. I’ll try to implement this in Guix later. <stikonas>rekado: ok. It's just a bit shorter... not by much...