IRC channel logs
2025-04-28.log
back to list of logs
<matrix_bridge><cosinusoidally> I suppose the current CI has the benefit that it will also test the "mirror.sh" script periodically (when the cache expires). <fossy>reason #1 is we want to test the mirror workflow as well <fossy>reason #2 is (especially since stikonas and I are both developers of live-bootstrap) it's trivially easy for us to backdoor CI if we use our mirrors <fossy>reason #3 is i don't want my mirror getting hammered bandwidth wise :D <fossy>"would it be frowned upon by github if someone just checked the tarballs into a repo and used that as a mirror?" <-- probably not, but that runs into reasons #1 and #2 as well <matrix_bridge><cosinusoidally> fair enough, that makes sense. Presumably not every tarball has an upstream gpg signature from the original developers? eg iirc your gnulib checkouts don't? If everything was signed by an upstream then in theory we wouldn't have to trust that you have not backdoored the tarballs on your mirror. <lanodan>Yeah upstream signatures are sadly still pretty rare, but I guess it could make sense to add support for gpg (and maybe signify, at least OpenBSD software typically uses it) <aggi>in the meantime, it's possible to compile/link perl-5.36.0 with tinycc too, glady, because autotools <aggi>so it's two versions verified, perl-5.8.6 which could be hooked into the bootstrappable dependency chain early, but does not support cross-compile <aggi>and perl-5.36, which supports cross-compile <aggi>since a complete posix base sytem profile passed with tinycc already on BUILD=TARGET=x86 with i386-tcc <aggi>currently, i try to push this profile from some BUILDHOST=aarch32 towards x86 with tinycc <aggi>would like to talk a little more about this, but don'w want to flood this channel <aggi>idea with this, any system on any ARCH which got a tinycc-toolchain ready could cross-compile a complete posix development host with it