00:03 <FromGitter> <tenebrousedge> JS is such rancid garbage

00:21 <sorcus> No one love JS :-D

03:26 <FromGitter> <rishavs> i love js :(

03:50 <sorcus> rishavs: Don't worry, JS not bad :-)

03:51 <sorcus> rishavs: But for me JS has one thing that i little hate... HTTP requests :-D

03:53 <FromGitter> <rishavs> I love how json is a 1st class citizen in js. no other language that I have used has that simplicity

03:55 <FromGitter> <craysiii> What is a hash then?

04:01 <FromGitter> <Blacksmoke16> a pita for the most part :p

04:30 <FromGitter> <tenebrousedge> > I love how json is a 1st class citizen in js. no other language that I have used has that simplicity ⏎ ⏎ Is this sarcasm?

04:31 <FromGitter> <tenebrousedge> or some other form of parody?

08:21 <FromGitter> <grkek> Death penalty for the JS pleb

10:35 <FromGitter> <Daniel-Worrall> Please be respectful to people.

11:15 <FromGitter> <rurounijones_gitlab> Ahoi all. I have a Channel setup as a queue. It accepts different event types (Update, Delete, SetProperty). Each event is a NamedTuple alias with different fields. In the reader of this channel I am trying to direct the events to their respective by using a `case` statement based on `event.class` but the compiler is complaining about the fact that I have code, that is only called for the Update

11:15 <FromGitter> ... event, is looking for keys in the Delete event. My thought was that, since this code is never called for a Delete event, the compiler should know that and not complain. Is my thinking correct?

11:20 <FromGitter> <rurounijones_gitlab> (I am porting the code from Ruby hence the use of NamedTuples since the original was just passing hashes around. It would not be a complete pain to switch to using actual classes instead of aliases, I just want to know if I am wrong in my assumptions)

11:22 <FromGitter> <rurounijones_gitlab> Ah, reading the docs more, I don't think `alias` actually matters at compile time so the `case` statement is bunk since all it sees is `NamedTuple`, not the alias name

11:28 <FromGitter> <rurounijones_gitlab> and just to finish off for posterity. I believe it is `typeof(T)`, not `T.class` - not that it works for my use-case anyway.

12:13 <FromGitter> <rurounijones_gitlab> Ok, even switching to sending events in fully defined classes I am still getting this behavour. ⏎ ⏎ ```code paste, see link``` ⏎ ⏎ `Error: undefined method 'data' for TacScribe::ReferenceLatitudeSettingEvent (compile-time type is (TacScribe::ReferenceLatitudeSettingEvent | TacScribe::ReferenceLongitudeSettingEvent | TacScribe::UnitDeletionEvent | TacScribe::UnitUpdateEvent))`

12:13 <FromGitter> ... [https://gitter.im/crystal-lang/crystal?at=5e302564fe0e6f74e9d473ac]

12:20 <FromGitter> <rurounijones_gitlab> And again for posterity. `typeof` was not needed as `case` uses the `===` operator and from the docs "For example, Class defines case equality as when an object is an instance of that class,". No idea if this would have worked with `alias` but I am not going to rewrite the code to check.

13:51 <FromGitter> <tenebrousedge> :plus1:

13:52 <FromGitter> <tenebrousedge> @rurounijones_gitlab ⏎ ⏎ ```[1].class #=> Array ⏎ typeof([1]) #=> Array(Int32)``` [https://gitter.im/crystal-lang/crystal?at=5e303c96f6945f41ef2f952e]

15:56 <repo> there's no version of JSON::Any#dig that would accept an array as subkeys?

15:57 <repo> so if i had a runtime input like "foo.bar.baz" i'd need to manually dig?

15:58 <FromGitter> <tenebrousedge> hmmmm, maybe `reduce`

15:58 <repo> yeah

15:58 <FromGitter> <tenebrousedge> ```arr.reduce(json) {|m, e| m.dig(e) }```

15:58 <repo> yeah

16:01 <repo> or for nilable stuff arr.reduce(json) { |m, e| m.try(&.[]? e) }

16:01 <repo> or dig i guess

16:02 <repo> *dig?

16:03 <FromGitter> <tenebrousedge> and for the obnoxious null issue: `arr.reduce(json) {|m, e| m.dig?(e) rescue m }`

18:19 <FromGitter> <christopherzimmerman> For any baseball junkies out there, I'm working on a retrosheet play-by-play parser. Currently able to parse about 6300 games per second.

23:17 <FromGitter> <Daniel-Worrall> Anyone used an rss feed builder on Crystal?