IRC channel logs
2019-10-24.log
back to list of logs
<stikonas>yeah, initially I was surprised that so many hashes have changed... <stikonas>hmm, did I actually update kaem-optional-seed hash in docs <stikonas>ok, maybe it's not mentioned anywhere in docs of mescc-tools-seed... <OriansJ>however changing comments can't change binaries in M2-Planet <OriansJ>changing error messages should change the hash (as is seen in cc_x86*, lisp.s and M2-Planet) <OriansJ>but kaem.c doesn't have it in anything other than a comment <OriansJ>and there is a 732byte difference, so it isn't a replace <OriansJ>M0 does not put anything in binaries that isn't directly in the source <OriansJ>So even with major changes in M2-Planet or M* or Hex*, the binaries must remain the same if none of the source changed <OriansJ>// Lines in M2-Planet however are code (see // CONSTANT foo 4, etc) <OriansJ>and # are line comments in M2-Planet <OriansJ>so #define and #includes are ignored <stikonas>hmm, kaem has sha256sum of here 04dd8ea2ea9eefdd7bfb2f375bd6d352cb2a2229d7efdc47c37f16f835dd5c81 <stikonas>somehow it's the same as kaem-optional-seed on my system <OriansJ>hmmm as mine is 6b29c69b626ed687d635202e9017a5b8a8959c7548c941c4255a35712385c996 <OriansJ>prior to your 676128e54d210921fd45f172e4fe0eed720b0cfc commit <OriansJ>and my c266436df65f7708e39824a1982e8d1fbfe5e62f merge <OriansJ>you should have a 001e20a2f8a9ddf4634f312a29b42a85c681c146 <OriansJ>should show a good signature from CA1132DD2BA1E81AC10C95D35410E91C14959E87 <OriansJ>Primary key fingerprint: CA11 32DD 2BA1 E81A C10C 95D3 5410 E91C 1495 9E87 <OriansJ>Do you have that in your git history? <stikonas>OriansJ: oh, I think I am one commit ahead <stikonas>I do have 001e20a2f8a9ddf4634f312a29b42a85c681c146 though <stikonas>my last commit is 8d34733436244f4d2e33bdd19f089d55ef84c51f <OriansJ>that is the one where I am trying to figure out the change in the kaem-optional-seed <stikonas>let me try to compile again before my commit, and after <stikonas>oh tyes, it was 6b29c69b626ed687d635202e9017a5b8a8959c7548c941c4255a35712385c996 <stikonas>I probably coppied it manually somehow from some older version <stikonas>and accidentally added with git add to commit <stikonas>sorry about that, I almost forgot about it myself <OriansJ>stikonas: the kaem-optional-binary was a hand checked binary; that I haven't updated in a while (Because I don't want anyone to trust it) <stikonas>I'm still unsure how kaem-optional-binary got into my commit... <OriansJ>no harm done, we caught it prior to a merge <stikonas>well, just removed that file from my commit <OriansJ>which reminds me, I still need to write kaem in assembly to reduce the optional-binary size down and simplify the audit <OriansJ>stikonas: one detail is that kaem-optional-binary isn't actually required by mescc-tools-seed as one can run kaem.run with any shell or simply execute the commands manually <stikonas>well, I was just running commands manually <stikonas>possibly later I coppied that file to kaem-optional-binary <stikonas>well, it still needs some kind OS or shell... <OriansJ>I probably should add to .gitignore so that when people build their own trusted kaem binary; it doesn't accidentially show up <OriansJ>actually no, with knight hardware in fpga and a couple paper tape readers <OriansJ>and M2-Planet + mescc-tools as submodules so people can get it all at once <stikonas>hmm, so when I played with knight VM, I was able to get cc_x86 <stikonas>how do I build M2-Planet which runs on Knight? <stikonas>hmm, I probably just need to pick the right defs file for M0, is that right? <OriansJ>actually I haven't written the cc_knight.S yet <OriansJ>it would be only a 112 line diff from the current cc_x86.S <OriansJ>simply change out some strings and a few minor differences in behavior as outlined in M2-Planet's Knight generator and boom done <OriansJ>I guess I have been putting this clean up off long enough <stikonas>well, I guess mes-m2 was higher priority... <oriansj>in fun news mescc-tools-seed has gotten a makeover <oriansj>and once I remembered that M0 in AMD64 is touchy about hex numbers (doesn't like 0x3c but does 0x3C fine) <oriansj>The upcoming patch will enable a full AMD64 bootstrap in mescc-tools-seed <oriansj>now here comes a hard style question; should I move the NASM ports of mescc-tools into mescc-tools-seed/