<FromGitter>
<anamba> hmm. anyone else seen this? sometimes, if i compile with --error-trace, it crashes. if i remove --error-trace, it completes just fine.
<FromGitter>
<anamba> i guess it's not so terrible, ideally you wouldn't need --error-trace when building with --release
<FromGitter>
<Blacksmoke16> try with `--no-debug`
<FromGitter>
<anamba> with --error-trace AND --no-debug?
<FromGitter>
<Blacksmoke16> yea
<FromGitter>
<anamba> hmm... hard to tell. it doesn't seem to crash every time :) i guess that's par for the course with weird memory errors. i'll try to test some more later...
alex`` has quit [Ping timeout: 265 seconds]
<FromGitter>
<Blacksmoke16> is what you get for using experimental features 😉
ua has quit [Read error: Connection timed out]
blassin3 has joined #crystal-lang
blassin3 has quit [Client Quit]
blassin has quit [Ping timeout: 268 seconds]
<FromGitter>
<tenebrousedge> hey @Blacksmoke16
<FromGitter>
<Blacksmoke16> o/
<FromGitter>
<tenebrousedge> hope things are well in your world
<FromGitter>
<Blacksmoke16> yea its going well
<FromGitter>
<tenebrousedge> if I want to poke at a property dynamically then I need a macro for that, yeah?
<FromGitter>
<bew> @zsxawerdu looks good!! I'm a fan of plain terminal neovim, what crystal tool did you use for method autocompletion? Scry?
<FromGitter>
<bew> Nice perf' @Blacksmoke16 ❤️
<FromGitter>
<bew> Can you link the commit where you did the change? (_wants to see it in my own eyes_)
<FromGitter>
<bew> *with* *
<FromGitter>
<asterite> @Blacksmoke16 Nice!
alex`` has quit [Ping timeout: 276 seconds]
<FromGitter>
<zsxawerdu> @bew scry. I been playing with Neovim + SpaceVim. really want a good front end to Neovim . Oni2 isn't out yet
alex`` has joined #crystal-lang
<FromGitter>
<wsdjeg> I recommand to use neovim-qt + spacevim
<FromGitter>
<tenebrousedge> spacevim <3
<FromGitter>
<Blacksmoke16> @bew will do when i commit it, prob later today/this weekend
* FromGitter
* c-cube uses plain old vim8
<FromGitter>
<naqvis> @j8r Just published your long awaited Crystal bindings to the (XZ)[https://github.com/naqvis/xz.cr] to the library 😆
rohitpaulk has joined #crystal-lang
rohitpaulk has quit [Ping timeout: 268 seconds]
rohitpaulk has joined #crystal-lang
<FromGitter>
<zsxawerdu> I second neovim-qt + spacevim
<FromGitter>
<zsxawerdu> just installed it on my linux box... amazingly fast
<FromGitter>
<zsxawerdu> neovim-qt + spacevim hits the sweet spot for me.
<FromGitter>
<bararchy> I've wasted something like 2 days of work to find out a change to the specs (how specs works in crystal) has made the tests fail
<FromGitter>
<bararchy> 😢
hightower2 has joined #crystal-lang
dannyAAM has quit [Quit: znc.saru.moe : ZNC 1.6.2 - http://znc.in]
dannyAAM has joined #crystal-lang
rohitpaulk has quit [Ping timeout: 240 seconds]
rohitpaulk has joined #crystal-lang
alex`` has quit [Quit: WeeChat 2.6]
alex`` has joined #crystal-lang
rohitpaulk has quit [Ping timeout: 240 seconds]
rohitpaulk has joined #crystal-lang
alex`` has quit [Ping timeout: 240 seconds]
alex`` has joined #crystal-lang
<FromGitter>
<sam0x17> does anyone know where `Spec::RootContext.check_nesting_spec` went with 0.31.0? I need it :/
<FromGitter>
<sam0x17> @naqvis does that lzma library support incremental addition of data i.e. adding 45 bytes, then 3 bytes, then 26 bytes, all under the same compression context? A lot of compression bindings fail to capture this but it is essential for a lot of use cases that involve streaming
rohitpaulk has quit [Ping timeout: 268 seconds]
rohitpaulk has joined #crystal-lang
<FromGitter>
<asterite> check_nesting_spec was internal details
<jhass>
how about just writing a wrapper for `it`? something like `def tx_it(desc) it(desc) do db.transaction do yield end end end` (plus forwaring the line metadata)
<FromGitter>
<sam0x17> does `it` actually execute though? I thought it just puts a block somewhere
<FromGitter>
<sam0x17> at the time it runs
<FromGitter>
<sam0x17> I know in the ruby world examples dont actually run when they are defined
<FromGitter>
<sam0x17> like in `rspec`
<FromGitter>
<sam0x17> if so, then I could easily do it that way
<FromGitter>
<sam0x17> I might just submit a PR that adds `around_each` and random sorting of specs -- I need both desperately on this project -- way too many accidental spec run order sensitivities
<FromGitter>
<Blacksmoke16> the PR that implemented focus made all the it blocks run at program exit
<FromGitter>
<sam0x17> I literally thought it was an example of using Spec
<FromGitter>
<sam0x17> so I disregarded
<FromGitter>
<Blacksmoke16> xD
<FromGitter>
<sam0x17> well, that fixed it
<FromGitter>
<sam0x17> so yeah, I think I'll submit a PR for `Spec.around_each` if people are interested -- it's very useful if you are doing anything transactional and want to wrap specs in a block
ht_ has quit [Quit: ht_]
<FromGitter>
<rmarronnier> I have a Type problem (crazy, right?)
<FromGitter>
<rmarronnier> `Error: instance variable '@observation_space' of Cadmium::Classifier::Viterbi must be Set(String), not (Apatite::Matrix(Float64) | Set(String) | Nil)`
<FromGitter>
<rmarronnier> with ` observation_space = model[0] if model[0].is_a? Set(String) ⏎ `
<FromGitter>
<rmarronnier> I thought it would work
<FromGitter>
<Blacksmoke16> this way you're only setting it if there is a model and its a `Set(String)`
<FromGitter>
<rmarronnier> yeah, it just did it ! Thanks
<FromGitter>
<Blacksmoke16> np
<FromGitter>
<rmarronnier> But I thought the if suffix would do it
<FromGitter>
<rmarronnier> It's weird it's not equivalent
<FromGitter>
<Daniel-Worrall> Iirc the suffix works differently
<FromGitter>
<Blacksmoke16> part of the problem is `observation_space` would have been left nil if the type was not `Set(String)` which isnt a valid type for it
<FromGitter>
<Blacksmoke16> hence the `| Nil` in the type mismatch error
<FromGitter>
<Daniel-Worrall> I've had if suffix problems before
<FromGitter>
<rmarronnier> Yeah I got it for the `nil` problem. But I guess the key part of the solution is the `.as(Set(String))`
<FromGitter>
<Blacksmoke16> that would have helped yea
<FromGitter>
<Blacksmoke16> cept it would fail if you try to cast nil to a set
<FromGitter>
<rmarronnier> This works so that's not a suffix problem :-)
<FromGitter>
<Blacksmoke16> the problem with the suffix is it doesnt restrict that `model[0]` would be a `Set(String)` like it does in the normal block version
<FromGitter>
<rmarronnier> Yeah, getting rid of the proxy model variable gets rid of the possibility of getting a nil value
<FromGitter>
<Blacksmoke16> i.e. why you have to use the `.as`
<FromGitter>
<rmarronnier> Thanks. It's very helpful
<FromGitter>
<jwoertink> Is there a way to tell the memory size of a hash?
<FromGitter>
<jwoertink> I tried using `Benchmark.memory`, but it returns 0
<FromGitter>
<jwoertink> `Benchmark.memory` is a bit confusing
<FromGitter>
<jwoertink> so I guess it's looking at the memory of operations happening and not the objects in the yield?
<FromGitter>
<asterite> It returns the number of bytes that get allocated in the block. Referencing a variable doesn't allocate memory. Creating a Hash does.
<FromGitter>
<jwoertink> Ah. I thought it was just returning how big the block was in memory.
<FromGitter>
<zsxawerdu> I never learned ruby, other than hack a script together once or twice
<FromGitter>
<zsxawerdu> safari books online has this effective ruby course , so I hit play. and all the mistakes they tell you to avoid is caught by the compiler in crystal