IRC channel logs

2021-12-20.log

back to list of logs

<oriansj>stikonas: well yes; simplification of mes.c certainly would be the easier route but growth in M2-Planet is reasonable to a degree
<jackhill>Blog post about something else, but I took notice of this parenthetical at the end: "(For the record, back in 2013, Ken told me that the backdoored compiler did in fact exist and was deployed, but it had a subtle bug that exposed the backdoor: the reproduction code added an extra \0 to the string each time it copied itself. This made the compiler binary grow by one byte every time it rebuilt
<jackhill>itself. Eventually someone noticed and debugged what was going on, aided by the fact that the “print assembly” compilation mode did not insert the backdoor at all, and all traces were cleaned away. Or so we believe.)" https://research.swtch.com/trojan
<muurkha>jackhill: thank you very much for finding that
<jackhill>:)
***saksophony_ is now known as saksophony
<Hagfish> https://justine.lol/sectorlisp2/
<Hagfish>comparing it to the genetic code of a phage is an intriguing benchmark
<mala>Is this work relevant to the bootstrappable project? https://justine.lol/sectorlisp2/
<muurkha>oh hi Danny
<muurkha>ltns
<mala>muurkha, heya!
<muurkha>(kragen)
<muurkha>Hagfish had just mentioned it
<mala>yeah i thought i was taking a risk just whaling into the channel without context, but in my defence i did check the logs :)
<muurkha>glad to see you!
<stikonas[m]>Somebody mentioned sectorlisp before if I remember correctly
<stikonas[m]>The problem so far with similar projects is that all them are individual pieces rather than something you can use to bootstrap GCC
<stikonas[m]>One could in principle write stuff in lisp, but nobody turned up to do that
<stikonas[m]>Well, there is mes but I guess sectorlisp is nowhere as advanced
<muurkha>aye