<fossy>stikonas: it seems translations <fossy>or maybe I was just looking wronh <stikonas>intl probably handles translations, but I think those C files ***ChanServ sets mode: +o rekado_
<stikonas>hmm, I'm having some trouble regenerating even those automatically generated aclocal.m4... <fossy>stikonas[m]: isn't the reason we use musl 1.1.x is because thats what gios musl tcc branch is based off? <stikonas>fossy: I think we mainly used musl 1.1.x because 1.2 has more advanced C that tcc can't build <stikonas>I think tcc struggles even with GCC compiled musl 1.2, there is something in the headers there <pder>stikonas: I've been working on dropping coreutils 6.3 in favor of 6.10. I am able to build sha256sum as well as mktemp now. <stikonas>pder: I think I might also be able to build xz <pder>I am doing coreutils 6.10 in the exact spot that 6.3 is done <pder>Yes, starting with your makefile and making a few adjustments <stikonas>since coreutils xz is like 3 times smaller than gz <pder>oh I didnt realize xz was in coreutils <stikonas>I mean use xz to download coreutils*.tar.xz <stikonas>4.999.9beta might work in terms of autotools <pder>Ill try to get this cleaned up later, and hopefully the bash build will be reliable with that PIPESIZE constant <stikonas>I'm asking if they can reinstate gzip bz2 tarballs too... <pder>Also, if its any easier, xzdec might be easier to build since its smaller <stikonas>well, xzembedded is even smaller actually <stikonas>although, not compatible cmd line interface <stikonas>pder: your stderr PR is fine, but maybe rebase it once GCC is pushed, so that CI can pass <pder>ok. speaking of gcc, once the new musl headers are installed it appears to break tcc, so maybe tcc could be removed at that point? <pder>or maybe we can figure out a way to install them side by side <pder>stikonas: do you know if date was broken in coreutils-5.0? It appears that your coreutils-6.3 makefile is building date as well as sha256sum <stikonas>I think we can just remove tcc at this point... <stikonas>also there are some remnants of mes c library... <stikonas>some stuff might be easier to clean up with bash >= 4 <stikonas>since we can use it to delete everything but one thing <pder>stikonas: I pushed a branched named mktemp. Feel free to try merging with your gcc branch and see if the bash-5.1 build is consistent with checksum. I'll test as well <stikonas>ok, thanks, I'll run chroot first, then qemu <stikonas>GCC branch now passes too, but if we can merge mktemp and bash works, I'll undo that last commit... <stikonas>fossy: did you mention before that newer libtool would automatically work with autoreconf? <stikonas>pder: I'm still building with qemu, but your commit looks fine <Hagfish>"Rebuild GCC with GCC and also against the latest musl." <Hagfish>i'm so hyped to see that in the parts.rst <Hagfish>is there a plan to build a kernel to run future steps in, and get the CI to boot that? <Hagfish>i'm not sure what would be the earliest point in the chain it would be possible to build a viable kernel (and i guess there are philosophical questions about whether there is any benefit in even trying) <Hagfish>changing topic slightly, i see that the Computerphile channel has uploaded a video today about Trusting Trust <stikonas>pder: hmm, I don't get the same checksum in qemu even with mkstemp...