***sm2n is now known as sm2n_
***sm2n_ is now known as sm2n
<oriansj>as it has been delayed far longer than it should have been. I'm shipping it now <oriansj>stage0-posix incorporating M2libc is now up <oriansj>I'll incorporate the mescc-tools-extras being built when I get the chance. <oriansj>Which should make live-bootstrap have a much faster and cleaner starting point. <oriansj>it incorporates all of the speed enhancements and new architectures as well. ***janneke_ is now known as janneke
***rekado_ is now known as rekado
***roptat_ is now known as roptat
<oriansj>well looks like I screwed up the AArch64, I better fix that. Pity single byte writes are so slow on flash when you use emulation. (the cc_aarch64 stage is a bit painfully slow) but atleast the M2libc speed improvements make massive gains for that case. <oriansj>and the hex2+M1 hashing behavior really makes them much more zippy. <mihi>(not really morning at 3PM, but I know that timezones exist) <mihi>congrats to finishing the M2libc migration :) <oriansj>mihi: few little bits to touch up and fix but yeah the x86 and AMD64 branches are ready to go <mihi>when looking at the new tree: Is there a particular reason that Kaem and catm exist both in mescc-tools and mescc-tools-extra? I assume one of them is about to go? <oriansj>mihi: yes there is. MesCC doesn't use catm or kaem but they are essential to the bootstrap. So I wanted to get the perspective of fossy (kaem is their baby) before making that sort of change. <oriansj>thank you for pointing that out to me mihi. I guess I have been focused locally lately. <mihi>yay, after a week of almost only rain the sun is just coming out. Got to go, see y'all next time :) <stikonas>oriansj: so what's the staths of m2libc now? Can it build mes or not yet? <oriansj>stikonas: well the big unification of the M1 definitions and ELF-headers is done for all but mes-m2. which is running into a problem relating to MesCC just using file descriptors instead of file structs. <oriansj>which is what the GCC branch is trying to address <oriansj>So once the GCC branch gets MesCC running again. I'll be able to finish that M2libc conversion and do the final unification. <stikonas>on unrelated note, I'm also thinking that we should have some inftrastructure in stage0-posix to automatically launch some other kaem file, so that live-bootstrap can plug into it a bit easier than current hack... <stikonas>possibly something like if next.kaem exists, run it <stikonas>although, we don't have conditionals in kaem yet, so maybe need to outsource it to C <oriansj>stikonas: adding a single conditional and exists condition shouldn't be hard to add to kaem. <oriansj>or make a dummy after.kaem which live-bootstrap or any other project could hijack. <stikonas>now live-bootstrap has to run kaem-optional-seed twice to hijack stuff at the beginning of stage0 <oriansj>just have its contents be default just a comment of hey, replace this file and everything will go from there. <stikonas>anyway, that's something to do when we move live-bootstrap to stage0-posix with m2libc <oriansj>stikonas: details to address when we get there. <oriansj>oh and I've added the flag --blood to get_machine so that scripts no longer have to reason about what they would need to pass to blood-elf <stikonas>it's basically a somewhat stupid copy paste error <stikonas>in principle can be fixed with 1 character replacement :D <stikonas>or possibly introduced when automake was moved after gcc... <Noisytoot>Why was everyone banned from #bootstrappable on freenode? <stikonas>I guess to prevent people who missed the move to libera? <stikonas>although, that could have been done a bit better with some message <stikonas>most people already moved to here I think <drakonis>also to prevent the new owners from weaponizing the old channel <stikonas>well, I don't think it would prevent that if they really wanted... <stikonas>they probably have elevated permissions on the server anyway <vagrantc>yeah, the current freenode has demonstrated willingness to take over and repurpose channels <stikonas>fossy: any thoughts on what we should do with CI? <Melg817>Thank you guys for reviews and merge!) <stikonas>I guess the next big thing to merge is perl <Melg817>yea, that would be epic) and unlock much more stuff