scala/scala

Discord is now Scala’s main chat platform. Please join us at https://discord.com/invite/scala

nafg

@nafg

I'm sure you are but there's still a lot that I'm missing

Gavin Bisesi

@Daenyth

JVM backend, js (ts) frontend. JSON between.
An api didn't work right because the frontend was rounding certain identifier values, which were Long on the jvm and serialized as json numbers

that's all there is to it

nafg

@nafg

Ah yes, I had that recently

Mainly in a property test using Nyaya

Now I have

  val genJsSafeLong = Gen.chooseLong(-9007199254740991L, 9007199254740991L)

Rob Norris

@tpolecat

We ran into it with timestamps.

GraphQL has real numeric types, which is nice.

nafg

@nafg

I mean in my case it's only an issue because Circe for scala.js relies on JSON.parse

IIRC upickle doesn't have that issue

i.e. there's nothing stopping you from parsing JSON in a scala-friendly way in scala.js. It will probably be slower, of course

Oh right -- upickle avoids it by representing Longs as a string

it doesn't use json number for longs

You can customize circe to do that also

Rob Norris

@tpolecat

:musical_note: string, string, string, string, everybody loves string

nafg

@nafg

It's better than bytes...

sometimes, anyway

Rob Norris

@tpolecat

could be worse

nafg

@nafg

I mean JSON is a string representation

Li Haoyi

@lihaoyi

you can actually tell uPickle to use numbers to represent longs, nowadays we have our own JSON parser written in scala.js that is competitive with JSON.parse

not necessarily a good idea though; who knows what other non-uPickle JSON libraries you may be interacting with that will truncate your longs to 64bit doubles

M Karthick

@mkartic

I've recently run into a file at work which was just a C struct instance dumped into a file. How would I go about parsing this in scala?

Something like, struct {char a; char b[5];} will have 'ABB$$$' (where $ denotes NUL).

I can probably read the file into a byte array, and convert it by hand. Is there a better way?

nafg

@nafg

Better by what metric?

How much of this do you need to do?

M Karthick

@mkartic

@nafg Better as in, whatever I've come up with sounds like the most naive approach possible. Wondering if there might be a more elegant way to do this.

Fairly often, as we have to parse this byte file on a regular basis.

nafg

@nafg

Right but is it just this one file and this one struct format? And how complex is it?

M Karthick

@mkartic

Its just this one struct format. But there are about 10 fields of varying lengths.

All the fields are primitives. Either strings or ints.

nafg

@nafg

The simplest thing that comes to mind that might make it a bit higher-level is DataInputStream but I'm not sure if it will necessarily do the right thing.

The full-on high level solution would probably be scodec

I'm just not sure if bringing it in would be overkill for your scenario

Also you can just make your own code nicer, that can go a long way in scala

Do you want to share it?

One technique is to use recursion and pattern matching, easiest if you make it into a List[Byte]

Then you can use :: or your own extractors

ByteBuffer is another JDK class that can help with pulling primitives out of bytes

https://stackoverflow.com/questions/31317196/how-to-decode-c-stuct-like-packets-with-java-socket-programming seems related

M Karthick

@mkartic

That one seems promising!

I'll share my code once I've written it. :)

nafg

@nafg

oh sorry misunderstood. ByteBuffer sounds like a good first thing to try

Rob Norris

@tpolecat

The right way to parse and produce binary data is a library called scodec.

Oh @nafg already mentioned that.

Soren

@srnb_gitlab

@Ichoran because

the people in this group have this weird notion that llvm is trash
they make it so the lib literally only works with gnu compilers
they say "just use C++ we already have a prebuilt Makefile"
the project was already in C++ by the person who started it, I just made PRs 1-3

felher

@felher

Hey folks. Any tips on how to deal with large scale refactorings? I normaly have my sbt on ~compile to see errors while I'm coding. But once in a while, a big, large refactoring comes along with no real way to do it incrementally. Now, every compile run takes forever, since the whole project needs to be recompiled, and I'm drowning in error messages. Maybe there is a way to tell sbt "Just compile file X.scala and all its dependencies", so I can go file by file?

felher

@felher

Just opened it in Intellij to see if it's easier there, but Intellij marks everything red before I even start the refactoring, so that's probably not a good sign :D
Intellij works pretty well if everything is in a final tagless style, though. :)

felher

@felher

I'm currently doing the same as here: https://stackoverflow.com/questions/14160677/how-to-compile-single-file-in-sbt , but that's cumbersome as well

Rob Norris

@tpolecat

vs-code with metals works well for me.

Where communities thrive

People

Repo info

Activity