00:02 <FromGitter> <omidathari> @watzon Interesting indeed. ⏎ The task ahead of me is I am trying to link to a C library and call a function that takes a uint8 * as argument (which is actually a string in C). And it appears that the pointer I provide to the function is causing issues. Have you ever attempted this before?

00:14 <FromGitter> <watzon> I definitely haven't

00:24 <FromGitter> <omidathari> Thanks @watzon

00:25 <FromGitter> <watzon> @omidathari wait, so since it takes a UInt8* you're trying to pass bytes in? You know you can still pass it a string right?

03:28 <FromGitter> <watzon> (https://files.gitter.im/crystal-lang/crystal/0tko/image.png)

03:28 <FromGitter> <watzon> Multi-threading ftw

03:39 <FromGitter> <watzon> Also, my laptop sucks

05:01 <FromGitter> <bararchy> @watzon hmmmmm interesting

05:01 <FromGitter> <bararchy> how much

05:01 <FromGitter> <bararchy> how come*

05:05 <FromGitter> <watzon> Just not enough ram and not a new enough processor. Not good enough for training neural networks anyway

05:53 <FromGitter> <bararchy> Hahahaja story of my life

05:53 <FromGitter> <bararchy> What neural net are you building?

05:54 <FromGitter> <watzon> I'm actually working on an implementation of GLOVE for Cadmium. It's like Word2Vec, but better for multithreaded workloads

05:54 <FromGitter> <watzon> And I've actually got it working too which is the exciting part

05:55 <FromGitter> <bararchy> Ohhhhh interesting , using multiple processes ?

05:56 <FromGitter> <watzon> Yup!

05:56 <FromGitter> <watzon> Still playing around and trying to make sure it's working right

05:56 <FromGitter> <watzon> But so far it seems to be

05:58 <FromGitter> <bararchy> Cool! Will explore once its in ;

09:05 <FromGitter> <yxhuvud> @watzon btw, did you see https://github.com/bollu/bollu.github.io?fbclid=IwAR3KNBAXO82OU0nvhJ2RxBVS2ZXjyVfdrebLieEjsUqDi2j2CxqFr3PSJpo#everything-you-know-about-word2vec-is-wrong ?

09:31 <FromGitter> <watzon> You know, that explains a lot

09:36 <FromGitter> <yxhuvud> I take that as a "no" :P

11:08 <FromGitter> <bararchy> @z64 I saw you took the discordcr under your wings ?

12:48 <z64> @bararchy in a way, yeah. i've been contributing to the library and working with its users for a long time; since i dont have push access atm, its been more convenient for me to maintain a fork since the other maintainers have been busy

13:20 <FromGitter> <j8r> mps we have a problem with the recent update of shards 0.9.0

13:23 <mps> j8r: didn't know. can you tell what is the problem

13:24 <FromGitter> <j8r> for example, take this shards.yml: https://github.com/DFabric/dppm/blob/master/shard.yml

13:24 <FromGitter> <j8r> `shards install` seems to be stuck if there are too many dependencies

13:25 <FromGitter> <j8r> oh, may be https://github.com/crystal-lang/shards/issues/282

13:26 <FromGitter> <j8r> sorry mps, that's probably a shards bug - unrelated to your work

13:28 <mps> np, I suspect that there is 'heisenbug' somewhere in crystal, so maybe also in shards

13:29 <FromGitter> <j8r> heisenbug?

13:34 <mps> Heisenberg's principle of indeterminacy (translation to english from head)

13:34 <FromGitter> <j8r> i see

13:35 <mps> bug which appears randomly in strictly same environment and context

13:36 <FromGitter> <j8r> this one seems to be linked to performance

13:40 <mps> I've seen bugs (segfault, iirc) when compiling crystal binary (apk) and just repeating build bug doesn't appears. sometimes have to repeat two times. and mostly on aarch64

13:41 <FromGitter> <j8r> weird :/

13:42 <mps> I can post log with this bug later. I'm out right now

14:08 <FromGitter> <j8r> I have git bisected it, it succeded

14:10 <FromGitter> <j8r> the commit found is https://github.com/crystal-lang/shards/commit/6698da36c78268ff8dc1798c9c9048641d7814ac

15:05 <FromGitter> <Blacksmoke16> @bararchy https://github.com/kemalcr/kemal/pull/538#issuecomment-503176523

15:49 <FromGitter> <alex-lairan> > Crystal is a programming language with the following goals: ⏎ > - Have a syntax similar to Ruby (but compatibility with it is not a goal) ⏎ ⏎ Is this line should stay ? Because the syntax moved a lot between the two languages ⏎ Ruby : `5 |> to_s` ... [https://gitter.im/crystal-lang/crystal?at=5d0907eda8d9871b32b88263]

15:49 <FromGitter> <Blacksmoke16> there was a discussion about the pipe operator

15:50 <FromGitter> <Blacksmoke16> https://github.com/crystal-lang/crystal/issues/1388

15:50 <FromGitter> <alex-lairan> Hope we don't implement the same as the ruby one

15:51 <FromGitter> <Blacksmoke16> its prob not going to happen

15:56 <FromGitter> <wontruefree> I totally dont get the ruby one

15:57 <FromGitter> <wontruefree> it is just dot syntax

15:58 <Yxhuvud> I suppose it might happen IF it turns out to be reasonably common and ideomatic in Ruby-land.

16:01 <FromGitter> <Blacksmoke16> read thru that issue, unless someone has a good use case of them being useful that cant be done already without them as nicely, then no point in having them

16:32 <FromGitter> <watzon> Is there a good way to test for memory leaks in Crystal

16:32 <FromGitter> <watzon> I'm pretty sure I have a pretty massive one

16:59 <Yxhuvud> valgrind? but it probably won't catch references that you hold but don't want to

17:05 <FromGitter> <watzon> Thanks I've never used valgrind before but I've heard of it

17:05 <FromGitter> <watzon> Trying it now

17:25 <FromGitter> <asterite> the way I test memory leaks is looping and then trying to reduce the code (sorry, my tools are always very primitive)

17:33 <FromGitter> <watzon> I think I found the issue. My Matrix class in my linear algebra library is supposed to be immutable, so every time you retrieve a row or column from the Matrix it creates a new vector and returns that. I think that's causing issues.

17:34 <oprypin> it really shouldn't by itself

17:34 <oprypin> but also: don't make it immutable

17:34 <FromGitter> <watzon> Yeah I'm going to change that

17:43 <FromGitter> <watzon> Hmm is there a reason this ⏎ ⏎ ```@text = @text.gsub(/([^[:alpha:]]+)|((?=\w*[a-z])(?=\w*[0-9])\w+)/, " ")``` ⏎ ⏎ would be giving me an `Uninitialized Read` error? [https://gitter.im/crystal-lang/crystal?at=5d0922d1faf70031f944eca0]

17:43 <FromGitter> <watzon> With Dr. Memory and Valgrind

17:44 <FromGitter> <watzon> https://hastebin.com/luquwanoyi.pl

18:12 <FromGitter> <Blacksmoke16> https://crystal-lang.org/api/master/String.html#at%28index%3AInt%29-instance-method shouldnt this have a `DEPRECATED` tag on the docs since its annotated as such? i remember that being a feature of the annotation

18:31 <oprypin> Blacksmoke16, bring it up regardless

18:31 <oprypin> (as in, an issue)

18:32 <FromGitter> <Blacksmoke16> 👍

19:04 <FromGitter> <Blacksmoke16> https://github.com/crystal-lang/crystal/issues/7895

19:04 <FromGitter> <Blacksmoke16> also made one for the UUID slice issue discussed a bit yesterday https://github.com/crystal-lang/crystal/issues/7894

19:18 <oprypin> 👍👍

19:45 <z64> :)

20:05 <FromGitter> <Blacksmoke16> https://play.crystal-lang.org/#/r/7375 camel case should prob default to lower = true

20:05 <FromGitter> <Blacksmoke16> pretty sure thats pascalcase :S

21:28 <FromGitter> <j8r> I hope shards 0.9.1 will be out soon : /

21:29 <FromGitter> <j8r> The new SAT solver has issues with long dependencies lists