<     May 2017     >
Su Mo Tu We Th Fr Sa  
    1  2  3  4  5  6  
 7  8  9 10 11 12 13  
14 15 16 17 18 19 20  
21 22 23 24 25 26 27  
_2_8 29 30 31
02:07 Sgeo joined
06:12 _whitelogger joined
08:59 prophile joined
10:26 mylesborins joined
16:05 <ec> fuckkkk
16:06 <ec> I need to Not Program, tho
16:06 <ec> I'm such a fucking addict
16:06 <ec> *days* of my life disappear down the gullet of some hungry shell-script that never even sees the light of day
16:06 <ec> I'm such a fucking addictive personality; but drugs, alcohol, Millennial Gambling¹, partying, none of those things hit me too hard
16:07 <ec> *just* programming and MMos
16:09 <jfhbrook> programming's more satisfying when you ship
16:09 <jfhbrook> last time I did programming at home I whipped up the thing but crapped out around packaging and tests
16:09 <jfhbrook> https://github.com/jfhbrook/coolbot98/blob/master/coolbot98/services/registry.py
16:09 <jfhbrook> this is the kinda thing I geek out about
16:53 <ec> I'm the opposite
16:53 <ec> I *care about* other stuff more, but I tend to *obsessively attack* packaging, management, testing, documentation tasks
18:42 <swart> I like getting distracted by random problems that have nothing to do with what I'm trying to test, and wasting hours of my day tracking down what turns out to be something stupid* [* still haven't figured out what this morning's stupid problem is]
18:43 <pikajude> something stupid*
18:43 <pikajude> (*note: all problems are stupid)
18:43 <swart> well only if they're my own fault. which is a lot of them
18:46 <pikajude> today i'm working with a class that has 33 nullable arguments in its constructor
18:47 <swart> that sounds unpleasant
18:47 <pikajude> :^)
18:48 <pikajude> oh, sorry, 27 nullable arguments
18:48 <pikajude> 6 of them are non-null
18:48 <swart> I've just spent 3 hours trying to find out why I can't get the same results to appear on two different machines
18:49 <swart> complaining doesn't help
19:46 <jfhbrook> ec: it doesn't help that packaging in python is kinda shitty
19:46 <jfhbrook> ec: and, like I get caught up in details--I wanna do it right y'know?
19:50 <jfhbrook> in other news though, I finished porting a spark job to python today
19:50 <jfhbrook> like put it into review and only got minor comments
20:17 <swart> well that was interesting. bug was my bug, but it had been there from the beginning and only showed up with test data
20:18 <pikajude> :O
20:19 <swart> that's what we get for loading array-valued attributes from relational databases and csv files
20:19 <pikajude> is this the beginning of a hex
20:21 <swart> dunno. it's now after 1:20 on Friday, I missed lunch and I'm exactly where I was at 4:20 last night
20:21 <pikajude> 4:20?? nice
20:21 <pikajude> i just received the new jeans i bought in the mail
20:21 <pikajude> but they're not the right ones, so i'm depressed
20:21 <swart> that's brave
20:22 <pikajude> being depressed?
20:22 <pikajude> oh
20:22 <swart> no ordering clothes in the mail
20:22 <pikajude> i mean
20:22 <swart> return them I guess
20:22 <swart> it's easy to do these days
20:22 <pikajude> i thought i was buying some that i had gotten from the same place before
20:22 <pikajude> but they are different
20:22 <pikajude> the only reason i bought them online is because they weren't in store
20:24 <swart> time for lunch
20:29 <jfhbrook> yo ljharb who the fuck is this james kyle dude
20:29 <jfhbrook> and why do yarn people seem to be so bad at dependencies
20:29 <pikajude> yo, yo, yo
20:29 <jfhbrook> like, lockfiles shouldn't apply for packages but you should have them anyway
20:29 <jfhbrook> subpackages*
20:30 <pikajude> i never used yarn
20:30 <pikajude> i hope it's good
20:31 <ljharb> jfhbrook: james was the second core babel dev, and he helped create yarn at fb
20:33 <jfhbrook> so babel is fine
20:34 <jfhbrook> yarn sounds like a turd though
20:34 <ljharb> it's not all bad
20:34 <ljharb> james is a delightful human tho
20:35 <jfhbrook> I mean
20:35 <jfhbrook> if yarn was a sports they'd be like
20:35 <jfhbrook> the pats, or worse
20:35 <jfhbrook> they'd be the sf giants >:6
20:36 <jfhbrook> seriously, fuck the giants
20:36 <jfhbrook> anyway, I'll always root for npm in a brawl
20:36 <ljharb> fa sho
20:41 <jfhbrook> I had a pretty good conversation with wycats though that night we got yelly about subdep lockfiles in #node.js and he noticed
20:43 <ljharb> anything come out of it?
20:43 <jfhbrook> I mean
20:43 <jfhbrook> I didn't convince him of much
20:44 <jfhbrook> but it did remind me that he's ultimately just some dude
20:44 <jfhbrook> and like the other time I'd talked with him was like 2011 and like
20:44 <jfhbrook> we've all done some growing up
20:44 <jfhbrook> his wife was annoyed he was yelling on the internet instead of going to bed, I thought that was funny
20:45 <jfhbrook> his thoughts on frameworks, ie the value of ember, are interesting and valid, even if I don't particularly like ember itself
20:45 <jfhbrook> like ember's fine, it does all the things, it's a viable alternative to react + all its friends
20:45 <jfhbrook> just wasn't my fave the time I worked on an ember app, that's all
20:46 <jfhbrook> it was GA event tracking, funny enough
20:46 <jfhbrook> this is likely going to be the limit of my kinja FE contributions as well
20:50 <ljharb> what's kinja?
20:51 <jfhbrook> https://kinja.com
20:51 <jfhbrook> gizmodo's cms
20:51 <jfhbrook> but also technically a medium competitor except apparently nick denton is bad at marketing?
20:51 <ljharb> ah
20:51 <pikajude> oh
20:52 <pikajude> no wonder kotaku, jezebel, jalopnik, and lifehacker all basically look exactly the same
20:52 <jfhbrook> like I honestly believe somebody fucked up kinja by not making it really really obvious that they're going toe to toe with medium
20:52 <jfhbrook> yeah pikajude it's a unified platform, it's actually pretty dope
20:52 <jfhbrook> they also have some pretty solid product people there
20:52 <jfhbrook> like the product is on point, it's easy to use, flexible in the ways that matter, rigid in the ways that keep it simple and fun to use
20:53 <jfhbrook> fully featured comments system, because that's core to the kinja platform
20:56 <pikajude> kool
20:56 <pikajude> shame about actual kotaku comments
20:56 <jfhbrook> I have some decent news ljharb
20:56 <jfhbrook> > One key detail about package-lock.json is that it cannot be published
20:57 <jfhbrook> so I can use a package-lock.json if I so wish for development without hosing downstream consumers
20:57 <pikajude> some downstream consumers are into being hosed though
20:57 <jfhbrook> well npm-shrinkwrap continues to work as before, I think
20:57 <jfhbrook> in fact I think the format is backwards compatible
20:58 <jfhbrook> yeah, in fact it appears that they're literally the same format, and npm-shrinkwrap.json is used for this latter use case
20:58 <jfhbrook> of "real lockfiles," I suppose
20:58 <jfhbrook> I guess yarn won that debate
20:58 <jfhbrook> you ever see Rat Race?
20:59 <jfhbrook> 2001 comedy
20:59 <jfhbrook> I keep thinking "maybe it's a secret publicity stunt" except "maybe it's an advisory lockfile"
20:59 <jfhbrook> http://www.imdb.com/title/tt0250687/?ref_=fn_al_tt_1 if you've never heard of it
21:04 <ljharb> jfhbrook: ok that's really nice, that npm will npmignore them no matter what
21:04 <ljharb> so at worst case, my optout will be "gitignore + postinstall script to remove it"
21:11 <jfhbrook> oh, yeah--the effect is the same but yeah I think it's actually straight ignored on upload
21:16 <ljharb> cool
21:16 <jfhbrook> yeah, new docs say that shrinkwrap is only if you're installing to deploy, as with a cli tool or maybe an app
22:20 <ec> swart!
22:20 <ec> been a while
22:22 <ec> somebody catch me up on this lockfile jazz
22:22 <swart> holaec
22:22 <ec> I barely had time to install yarn for the first time before people started yelling that it was terrible and npm appeared to have subsumed the functionality???
22:23 <ec> I'm so bad at keeping up with the JS world outside my little project-directory-bubble
22:23 <jfhbrook> yeah, that's the short answer
22:23 <jfhbrook> yeah don't feel bad about that, yarn was afaict just an annoying diversion
22:23 <* swart> is looking forward to not learning yet another package manager
22:23 <ec> well
22:23 <ec> what about the other third-party node_modules managers
22:23 <jfhbrook> yarn's the only one with serious adoption afaik
22:24 <jfhbrook> anyway, npm 5 supposedly solves the same problems yarn did as compared to 4, but better
22:24 <ec> pnpm?
22:24 <ec> and that other one?
22:24 <jfhbrook> see we can't even remember their names
22:24 <ec> lmfao
22:24 <ec> pnpm has reasonable adoption and is kept up-to-date
22:24 <ec> and there's another that is the same status of Meaningfulness but I've personally never used it
22:48 <ljharb> yarn is the only npm option that's ever been viable, altho the others tried hard
22:50 <* ec> nods
22:50 <pikajude> yarn also makes node more cat-accessible
22:56 <joepie91> ec: you're probably thinking of ied?
23:00 <ljharb> didn't ied rename, so it wasn't named after a missile?
23:01 <pikajude> that's probably a good idea
23:09 <jfhbrook> or the thing that blew up my cousin's stryker in iraq
23:09 <jfhbrook> (he was fine, like battery acid damage but no condolences necessary)
23:10 <jfhbrook> icbm would be a better name honestly
23:16 <joepie91> ljharb: I uh, can see why they would
23:16 <joepie91> lol
23:17 <joepie91> hm, doesn't seem like they actually renamed though