03:37 <intelligent_boat> I remember trying it on some previous years and it didn't make a difference, but it's quite good this time

03:38 <intelligent_boat> without yjit, day 22 is the slowest day, taking 1 second to compute. with yjit, it takes 300ms, pretty nice speedup then

03:39 <intelligent_boat> day 20 is the second slowest at 450 ms without, and it also gets brought down to about 300ms with

07:38 <nakilon> o0x1eef: there was a tool years ago that was supposed to draw a function call tree but it couldn't parse my code when I tried; there are new parsers but I doubt anyone has reworked it; but in theory a custom parser with logic that I used to make that graph , can be made

07:38 <nakilon> and it's literal drawing of course, it's graphviz dot

07:40 <nakilon> but dot need human assistance anyway because it can't reorder lists on its own, such those pink blocks with gems -- you do it manually to reduce edge intersections

07:41 <nakilon> my graph basically shows just require and require_relative directives

07:41 <nakilon> and the folder location

07:53 <nakilon> do parsers know the difference between locally defined constants and globals? then that custom tool could give advices like "you forgot to require something, this constant is undefined", or "why did you require this if you don't use constants with such name?" -- assuming there is an accepted codestyle of putting the modules in the files with the

07:53 <nakilon> same name (that I don't fully follow though)

08:00 <nakilon> would be cool if when I do require some module M in a file describing module A, that then within A::B it would be accessible as A::M

08:01 <nakilon> then instead of three edges A->M, B->A, B->M there would be only two A->M, B->A

08:04 <nakilon> like "include" but for modules, not methods

09:46 <nakilon> weird rubyzip gem; the repo samples directory includes this https://github.com/rubyzip/rubyzip/blob/v2.0.0/samples/example_recursive.rb but in the published gem has different set of files, this one is missing, another one has slightly different name...

10:36 <nakilon> figured it out; the third time I fell for that trap that I have to you "rubyzip", not "zip" gem

16:39 <o0x1eef> > do parsers know the difference between locally defined constants and globals?

16:40 <o0x1eef> I think that's state the parser would have to maintain - right ? So unless it is designed that way, probably not. That's not to say existing tools or parsers couldn't be updated to maintain state like that