14:45 <Julian> rpython refuses to compile this line of code unless I give the wrong type: https://github.com/Julian/rpylean/blob/d688df7a8f7790a071f05e6b7e2cb36404046ffe/rpylean/objects.py#L1200-L1205 (the right one is the commented out line) -- I'm rereading https://rpython.readthedocs.io/en/latest/rtyper.html to try to re-learn what this hop business is, but if anyone has a more targeted hint on what I should look at

14:45 <Julian> lemme know

14:45 <Julian> When I say "wrong" -- all uses of that object pass a `Name` as the first object, not an `Expr` which is what the compiling version does (and of course then segfaults at runtime)

14:47 <Julian> (I guess the kind of hint that ideally would help is if I'm sitting in pdb at the assert hop.args_r[i].lowleveltype == new_args_r[i].lowleveltype error, is enough information maintained to see *why* rpython thinks new_args_r[i].lowleveltype is what it is?)

14:48 <Julian> or even what function is being compiled, or is there no function at this point and whatever block is not necessarily correlated to a specific function

14:48 <Julian> s/compiled/typed

16:58 <Julian> OK I did random stuff and it compiles, hooray?

16:59 <korvo> Hooray!

17:06 <Julian> https://github.com/Julian/rpylean/commit/18576f2ba9b7a5c2e63492f0ac24f6ab17debcc3 is said random flailing