00:02 <FromGitter> <thedracle> So, I'm trying to make a websocket server using wss, and I'm getting the following OpenSSL error when connecting to it via a (browser based) client: ⏎ ⏎ ```code paste, see link``` ⏎ ⏎ It looks really similar to this error @nsuchy reported on Apr 14th regarding a HTTPS client connection from crystal, and omitting the 'Content-Type' header seemed to be the recommended solution. ...

00:02 <FromGitter> ... [https://gitter.im/crystal-lang/crystal?at=5e0fd61ca769bf30fd32ba41]

00:03 <FromGitter> <thedracle> Which obviously doesn't make any sense in the context of a server.

00:03 <FromGitter> <thedracle> I'm using libssl1.0.0--- which maybe isn't compatible with Crystal?

00:10 <FromGitter> <thedracle> The actual server code is as follows: ⏎ ⏎ ```code paste, see link``` [https://gitter.im/crystal-lang/crystal?at=5e0fd800c23234403cc4aa77]

00:32 <FromGitter> <stnluu_twitter> The code looks right. Any reason you're on v0.27?

00:32 <FromGitter> <stnluu_twitter> Does it work with v0.32?

00:35 <FromGitter> <stnluu_twitter> Also, try absolute path for the cert & the key file and see if that works for you

00:35 <FromGitter> <stnluu_twitter> depending on how you run it, the current directory might be different from what you're expecting

01:59 <FromGitter> <watzon> Yeah you should definitely update your Crystal version and try again. Try updating openssl too.

02:35 <FromGitter> <thedracle> Let me try upgrading to a newer crystal, I just used what was installed via brew.

02:35 <FromGitter> <thedracle> Thanks Chris.

02:35 <FromGitter> <thedracle> And Steven.

02:41 <FromGitter> <stnluu_twitter> brew install should give you 0.32 now

02:43 <FromGitter> <thedracle> Yeah, just updated, I must have installed it some time ago.

03:04 <FromGitter> <thedracle> Upgrading fixes the OpenSSL issue-- now the client is complaining about Sec-WebSocket-Protocol, which it looks like WebSocketHandler doesn't support.

04:58 <FromGitter> <stnluu_twitter> i had some luck with the kemal shard (had not tried WS)

04:58 <FromGitter> <stnluu_twitter> have you seen how it does things?

08:07 <FromGitter> <watzon> @thedracle tbh I think we've got some work to do with WebSockets

08:58 <FromGitter> <btihen> @thedracle - cool thanks for posting that - It looks interesting.

11:52 <FromGitter> <bew> @PenguinOwl try `make clean_cache` as well

19:05 <FromGitter> <smprather> Hi. First post. I'm getting this: ⏎ ` 394 | @log_fh = File.open(@log_file, "w") ⏎ ⏎ ``` ^------``` ⏎ ... [https://gitter.im/crystal-lang/crystal?at=5e10e1fef6ee9c3e8e079e28]

19:06 <FromGitter> <smprather> Hmm... my attempt at proper mark-down didn't go as planned, but you get the idea.

19:06 <FromGitter> <tenebrousedge> nothing in particular, but you'll need to set the type of that ivar

19:07 <FromGitter> <smprather> Yeh, I'm pretty sure explicit setting will work. I just though this was a case where inference was supposed to work. I haven't checked the source, but I assume the return type of File.open is annotated.

19:09 <FromGitter> <tenebrousedge> it is, but having a return type doesn't always ensure type inference

19:10 <FromGitter> <smprather> 1) `@log_fh = Type.method`, where `method` has a return type ⏎ ⏎ ``` annotation, type is inferred from it``` [https://gitter.im/crystal-lang/crystal?at=5e10e3411c3a592052cfbc9d]

19:11 <FromGitter> <smprather> (Sorry, from the list of "The assignments must look like this:")

19:11 <FromGitter> <tenebrousedge> yeah, it says that. I don't know how true that is

19:13 <FromGitter> <smprather> As expected, this in the class header works: @log_fh : File. I'm just wondering if this is a reportable inference bug (or enhancement).

19:15 <FromGitter> <smprather> Anyway, thanks @tenebrousedge !

19:16 <FromGitter> <tenebrousedge> @smprather this might be relevant https://github.com/crystal-lang/crystal/issues/4681#issuecomment-313417536

19:24 <FromGitter> <smprather> Indeed, the compiler would have to be able to detect if an expression contains a reference to another instance (or local) variable. I admit, it could be tricky. It seems simple cases of a function call(s) with literal arguments could be handled, but that's more of a version 2.x enhancement I suppose. I believe Scala has conquered tricky inference like this.

19:25 <FromGitter> <smprather> Whoops, drop the literal arguments requirement. It shouldn't go into checking method arg types. (and I meant method when I said "function" above)

19:26 <FromGitter> <smprather> In a nutshell, any human could infer this type w/o exception, so it seems the compiler could be taught.

19:27 <FromGitter> <tenebrousedge> perhaps, but if it carries a substantial speed penalty it might not get implemented

19:40 <FromGitter> <smprather> Doh, the type of args would have to be checked to determine the proper overload. That could get into inf loop. So there would have to be a special case of all method overloads returning the same type.

20:20 <FromGitter> <christopherzimmerman> Anyone heard of a language called V? I keep coming across modules directly ported from crystal.

20:31 <FromGitter> <smprather> The C -> V translator looks promising. It can presently translate Doom to V.

22:14 <FromGitter> <drum445> Hello, has anyone else seen issues with `Invalid memory access (signal 11) at address 0x7fae00000004` when running your app under --release mode?

22:24 <raz> nope, never.

22:24 <raz> can you reliably reproduce it?

22:25 <FromGitter> <tenebrousedge> usually that's a compiler bug

22:25 <FromGitter> <drum445> I can

22:25 <FromGitter> <drum445> I've had to chagne my code to this to avoid it `Array(Article).from_json(docs).dup # dup to avoid invalid mem errors`

22:25 <FromGitter> <tenebrousedge> hmmmmm

22:26 <FromGitter> <drum445> Printing the array also fixes the problem, seems like unless I use the object it falls out of memory

22:38 <FromGitter> <Blacksmoke16> Did you make a property on Article not nilable but it's missing from the json

22:39 <FromGitter> <Blacksmoke16> Or something along those lines

22:47 <FromGitter> <drum445> Afraid not, no difference in the docs json object between the times it crashses (--release) and the times it doesn't

22:48 <FromGitter> <drum445> the json object is coming from mongo fyi

22:50 <FromGitter> <tenebrousedge> can you throw the object and deserializer into a Crystal Play?

23:38 <FromGitter> <drum445> Sorry, can't get it to replicate at all