« first day (1135 days earlier)      last day (640 days later) » 

00:00
Lyxal enqueued pull request #1767 (Add unevaluated input element) in Vyxal/Vyxal
Lyxal pushed a commit to Vyxal/main: Merge pull request #1767 from Vyxal/v2-add-don't-eval-input
Lyxal deleted branch v2-add-don't-eval-input in Vyxal/Vyxal
Lyxal closed issue #1763 (Add Raw Input Element) in Vyxal/Vyxal
Lyxal merged pull request #1767 (Add unevaluated input element) in Vyxal/Vyxal
Lyxal deleted branch gh-readonly-queue/main/pr-1767-571746a93b6b4a698301b13458acd668b577648c in Vyxal/Vyxal
iirc akka had licencing issues or something that made everyone drop it
like it probably wouldn't apply to us, but support/usage of akka has dropped greatly recently
it was a major thing on the scala subreddit after all :p
00:49
Oh dear
Like I get they have to make money somehow, but apparently the way they did it wasn't very much appreciated by companies
> The BSL is a source available license that freely allows usage of the code for development and other non-production work such as testing. Production use of the software requires a commercial license. The commercial license will be available for early stage companies (less than US $25m in annual revenue) at no charge.
:(
(it's free for us obviously)
(and OSS gets a free licence grant anyway)
Hmm
Might still be better to use something else
the problem with it was that companies making $25m would have substantial costs involved with using akka
00:53
I'm not very knowledgeable about web stuff: just to confirm, we want a library to make an HTTP server, right?
like it wouldn't hurt big players like Microsoft, apple, google, et al so much
but companies just above the mark would be impacted
And companies making less wouldn't be impacted
That doesn't sound so bad
correct
@user well, apparently it was done in a sudden way
meaning one day you don't have any extra operating costs, the next day you're explaining to your manager that you either have to pay licencing or switch to another library
@user How about http4s?
@lyxal Huh
I'd expect the announcement to say "in <x> months, this will be implemented"
at least, that's what i remember reading
00:56
Where's the announcement btw?
a fair few people have abandoned akka for leaving OSS
like they've mentioned they won't be updating their libraries using it any more
rip
gosh dang, $1995 per core
> cross compiles to Scala.js and Scala Native (waiting for Scala 3 support in SN)
Wait I thought scala-yaml supported Scala 3 Native
that's taken straight from the readme
01:09
I think that might be outdated
circe-scala-yaml seems to be working for me
ysthakur pushed a commit to Vyxal/v3-momentum-branch-1: Use circe-scala-yaml
I love GitHub Codespaces
I'm looking forward to if/when we can have copilot powered PRs
although that'd probably require organisation licencing rather than just having at least one user licenced to copilot
Wait how did that fail
Or rather, how did that work in my codespace
did you commit everything?
because it seems to be an import error
 [error] -- [E008] Not Found Error: /home/runner/work/Vyxal/Vyxal/shared/src/test/scala/YamlTests.scala:7:17
[error] 7 |import io.circe.{yaml, Decoder, HCursor, Json}
[error]   |                 ^^^^
[error]   |                 value yaml is not a member of io.circe
01:14
Yeah no the error makes sense because I didn't port it to use the new lib
But it somehow worked in my codespace???
I think I ran ./mill native.test before I saved changes
01:25
And the fact that native worked there probably means the mill native test module is using the JVM?
ysthakur pushed a commit to Vyxal/v3-momentum-branch-1: Actually use circe-scala-yaml in YamlTests
[info] - Execute `+` on inputs 1, 2 *** FAILED ***
[info]   Expected 12, but got 3 (in Checkpoint) at YamlTests.scala:105 (YamlTests.scala:130)
w...
oh
{ in: ["1", 2], out: "12" }
hm looks like it's evaluating the input
Yeah, not sure why, though
I haven't changed anything major
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: Does wrapping in quotes help?
You know, this was going to be a temporary fix anyway, because that circe-scala-yaml library isn't being maintained
How about we just use VirtusLab/scala-yaml directly?
I can do that tomorrow
01:36
I mean, if it compiles to native 3, sure
wrapping in quotes did not help
[info] - Execute `+` on inputs "1", 2 *** FAILED ***
[info]   Expected 12, but got ""1"2" (in Checkpoint) at YamlTests.scala:105 (YamlTests.scala:130)
that's arguably worse
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: It did not
@lyxal It does, it's what this circe-scala-yaml library uses
02:38
ysthakur pushed a commit to Vyxal/v3-momentum-branch-1: Use VirtusLab/scala-yaml
Oh I forgot to test Native before pushing
02:53
ysthakur pushed a commit to Vyxal/v3-momentum-branch-1: Use getClass.getResourceAsStream
Aaaaand I just realized there's no way this'll work for JS because you can't load resources in Scala.JS
argh
Is that an issue though
Because chances are that if it works in jvm + native, it'll work on js
And if it's something complicated like reading files, then a) that feature will have a compilation error and b) we'll have discussed a workaround anyway
@lyxal It won't
Because Scalajs doesn't bundle resources into it (yet, at least)
No I mean the tests
Even with Native, only getClass.getResourceAsStream works (it's implemented with some hacks)
Oh
Yeah I guess
02:56
If something works in jvm and native tests it's probably gonna work on js
But Scala.js does have different semantics in some places for performance and stuff
Because it's already such a limited subset of functionality
@user but in the test cases performance doesn't matter
No I mean Scala.js itself has different semantics
Not our stuff
We can't change it even if we don't care about performance
I think it does some number stuff differently
Oh yeah that's right
It says 0.1 + 0.2 = 0.30000004
Well, at least it did at one point
@user but if a test passes in jvm and native, then chances are it'll pass js, right?
03:17
ye
btw, do we want to keep supporting mill and sbt both?
Yes
 
9 hours later…
12:14
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: More tests for add
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: Add a lot of tests for *
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: Add tests for <
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: Add tests for >
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: Add yaml tests for A
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: Add tests for C
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: Yaml tests for D, E, G and H
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: Add yaml tests for I, J, K, L and M
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: yaml tests for O and N
12:50
The magic of copilot is how I converted that many tests :p
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: YAML tests for P, R, S and T
Lyxal pushed a commit to Vyxal/v3-momentum-branch-1: Last few yaml tests
anyhow, that's the existing element tests ported to yaml
but it's somehow it's made the YAML error
java.lang.Error: Test groups cannot be scalars
 
3 hours later…
16:18
ysthakur requested changes on #1765 (Version 3 Momentum PR #1)
How does content of global array interact with lambda map?
This works well:
⟨⟨1 | 2 | 3 ⟩ | 3⟩ ÷(:⅛Ǔ)¾ → ⟨ ⟨ 1 | 2 | 3 ⟩ | ⟨ 2 | 3 | 1 ⟩ | ⟨ 3 | 1 | 2 ⟩ ⟩

This gives nothing:
⟨⟨⟨ 1 | 2 | 3 ⟩ | 3 ⟩ | ⟨ ⟨ 5 | 6 | 7 ⟩ | 3 ⟩⟩ ƛ÷(:⅛Ǔ); ¾ → ⟨ ⟩

This gives something strange:
⟨⟨⟨ 1 | 2 | 3 ⟩ | 3 ⟩ | ⟨ ⟨ 5 | 6 | 7 ⟩ | 3 ⟩⟩ ƛ÷(:⅛Ǔ)¾; → ⟨ ⟨ ⟨ 1 | 2 | 3 ⟩ | ⟨ 2 | 3 | 1 ⟩ | ⟨ 3 | 1 | 2 ⟩ ⟩ | ⟨ ⟨ 1 | 2 | 3 ⟩ | ⟨ 2 | 3 | 1 ⟩ | ⟨ 3 | 1 | 2 ⟩ | ⟨ 5 | 6 | 7 ⟩ | ⟨ 6 | 7 | 5 ⟩ | ⟨ 7 | 5 | 6 ⟩ ⟩ ⟩
16:57
The reason why the second one gives nothing is because maps are evaluated lazily. You can use ⟑ to evaluate it immediately and then it works as you want it to.
The third one gives exactly what you would expect because the global array is indeed *global*.
@AndrovT Thanks! I start experiments (^_^)
nayakrujul pushed a commit to Vyxal/v3-momentum-branch-1: Indent tests for *
@user ^ is what you meant by "indent", right?

« first day (1135 days earlier)      last day (640 days later) »