<oriansj>who is excited about ./bin/mes-m2 -e main -s foo 1 2 3 4 5 working? <fossy>oriansj: what does that do? Lol <oriansj>fossy: it demonstrates that mes-m2 now supports the exact same shell interface as guile and the one previous demonstrates an infinite loop with internal garbage collection cycling without issue, now that sufficient amount of state is being stored in the scheme stack <fossy>i have decided to pause further work on my mescc-tools-seed work to make mes->gcc because a) mes is rather annoying to work with and b) there are other improvements in other areas we need before we can do that (other than mes-m2), especially with kaem <fossy>the only work i plan to do related to that is make TCC compiliable from kaem <fossy>I mean make TCC 0.9.27 work fully from gash <oriansj>fossy: ok, make sure to let me know when your branch of mescc-tools is ready for merge ***ng0_ is now known as ng0
***janneke_ is now known as janneke
***Server sets mode: +cnt
<oriansj>since the memories of the bugs that appeared yesterday are still fresh in my head, I am going to add a shitload of comments into mes-m2; so that other people can avoid my pain. <oriansj>first batch of improvements comments are up (lots more to come) <oriansj>I guess mihi didn't feel like discuss anything <oriansj>although I find it weird that guile believes (cond ((cons 1 2) #f) (#t #t)) should return #f <oriansj>mes-m2 question for you janneke about quasiquote support: if correct behavior is to revert to unquote behavior (what guile does) Or restrict to just proper lists as the spec (r7rs) requires eg. `(foo bar ,@(+ 4 5)) <NieDzejkob>oriansj: why is that weird? That would just mean that (cons 1 2) is truthy, as normal, right? <oriansj>NieDzejkob: because (cond ((display #\newline) #f) (#t #t)) unspecified should not evaluate to TRUE!