<FromGitter>
<maxbertinetti> > So I realized that I have to have crystal to compile the project ⏎ ⏎ lol @eliasjpr
sorcus has quit [Quit: WeeChat 2.9]
sorcus has joined #crystal-lang
woodruffw has quit [Ping timeout: 264 seconds]
woodruffw has joined #crystal-lang
woodruffw has quit [Ping timeout: 256 seconds]
woodruffw has joined #crystal-lang
postmodern has quit [Quit: Leaving]
woodruffw has quit [Ping timeout: 246 seconds]
woodruffw has joined #crystal-lang
woodruffw has quit [Ping timeout: 256 seconds]
repo has quit [Remote host closed the connection]
Human_G33k has joined #crystal-lang
HumanGeek has quit [Ping timeout: 272 seconds]
<raz>
blacksmoke16: is the error always the same when it strikes? (i.e. always regex as the last mention)
<raz>
could be a breadcrumb... something in regex land not getting cleaned up, trying to apply regex to something that turned nil, or something like that
<FromGitter>
<Blacksmoke16> yea, if i remove the gsub it goes to something else. Pretty sure there is just something that is borked, like a dangling pointer in a hash or something
<raz>
well, if you at least have something that fails every time you run it, that's good
<raz>
the worst bugs are intermittent ones
<FromGitter>
<Blacksmoke16> indeed
<raz>
maybe just throw it at ast when he's around, he'll take his usual 3,75 seconds to think about it, mumble something indecipherable, then tell you the 1-line fix
<FromGitter>
<Blacksmoke16> its reproducible but only when like in spec land, when these two specific specs exist
<raz>
ah ok... if you need a trillion LOC to reproduce it, that might not be throwable so well
<FromGitter>
<Blacksmoke16> yea :/
<raz>
but it still breaks if you run *only* these two specs?
<FromGitter>
<Blacksmoke16> im also not convinced its an actual problem, or just something related to how im testing something, like doing something unsafe etc
<raz>
that might still be a way to bisect it down (find the sequence that breaks it)
<FromGitter>
<Blacksmoke16> going to try and reduce it more tonight
<raz>
well, yea, if unsafe calls are involved i'm out. i don't touch them because the docs say they are unsafe :p
<FromGitter>
<Blacksmoke16> not really "unsafe" stuff, just "hacky" :P
<raz>
well, hacky should not cause segfaults, unless explicit unsafe calls are involved, i would think
<FromGitter>
<Blacksmoke16> i have a theory its going to be something with "using a hash that is in a closured proc when something something struct"
<FromGitter>
<Blacksmoke16> :S
<raz>
oh yea... i know these something something moments...
<raz>
but yea, just switching some struct's to classes (if feasible) might also be worth a try
<FromGitter>
<Blacksmoke16> also tried switching everything to a class and didnt help
<raz>
ok, so at least one option eliminated
<FromGitter>
<Blacksmoke16> oh lovely it doesnt happen on my mac? Not sure if i just didnt commit everything or what...
<FromGitter>
<Blacksmoke16> lol it doesnt happen anymore
<FromGitter>
<naqvis> turned into “it works on my machine” bug 😄
<FromGitter>
<naqvis> might be some libraries version?
<FromGitter>
<Blacksmoke16> had it not happen, then happen, then now not happening
<FromGitter>
<Blacksmoke16> ah rip, happens like 1/5 times :/
<FromGitter>
<Blacksmoke16> 2/5
<FromGitter>
<naqvis> kinda hidden bug, might be sth happening under typical conditions
<FromGitter>
<naqvis> what if you run with gc disabled?
<FromGitter>
<Blacksmoke16> ooo good point, i was running this with `--order random`
<FromGitter>
<Blacksmoke16> yea im just going to say its because of some less than ideal things i was doing. Going to refactor that out go from there
<FromGitter>
<naqvis> 👍
<FromGitter>
<Blacksmoke16> was storing a proc in a class within a struct class var that gets cached, then reseting data in a `Spec.before_each`
<FromGitter>
<Blacksmoke16> 🙈
<FromGitter>
<naqvis> :P
postmodern has joined #crystal-lang
<FromGitter>
<asterite> what does "resetting" mean here? In any case it should never segfault
<FromGitter>
<Blacksmoke16> ill try and reduce it to something more manageable and create an issue if i can still reproduce
<z64>
hi all. hope this is cool to share here; the team i'm working with is looking to hire another crystal developer. you can find the details here, please use my listed email to follow up. thanks!<3 https://paste.sr.ht/~z64/9eccf27a54093ffaec2e70a689b849ba11174f06
<FromGitter>
<Blacksmoke16> got it narrowed down to non spec code, but now need to remove my implementations to make it simpler for an issue