Scala3 upgrade #488

seroperson · 2025-07-06T10:36:02Z

Continuing #383 (comment) and #485 discussion:

What I did here:

e1325b4:
- -rewrite -source:3.7-migration.
- Some explicit Decoder definitions, as they weren't resolved otherwise.
- Moved src-jvm to src-jvm2, because it also depends on scalaj-http.
Trying to workaround a bug I mentioned here (Scala 3 #383 (comment)), 3045a65:
- Splitting CirceDecoders, moving Decoder definitions to their models
After I splitted CirceDecoders, now the same StackOverflowException were occuring on CirceEncoders 🤡, so I splitted it the same way too, bf55b75
Then Scala3 finally compiled, but Scala2 build broke for some reason with unresolved implicits in RequestHandler.scala. I guess that's because of different implicit resolution mechanism on Scala2 and Scala3, and Scala2 isn't clever enough to actually match R with Request's type parameter and find implicit in its' object. That's why I've rewritten it using dependent types, b9bd2d2 (see Request, RequestHandler and implementations):
With dependent types instead of type parameter approach it now looks cleaner and everything compiles both for Scala2 and Scala3. I had to broke compatibility a little bit, but I hope it's excusable.

I haven't tested it in real application yet, don't have anything suitable at the moment, but at least tests are green.

Thank you for your attempt @viperey and @ex0ns for your great library 👍

ex0ns · 2025-07-06T12:07:47Z

Thanks a lot @seroperson
I don't really mind the breaking changes if that means that we can move this library to scala 3. And the dependent type instead of it being a type parameter looks very equivalent to me (even though I don't understand why this was working before and it does not anymore, maybe you have a guess about this ?)

I will try to upgrade some small project of mine with this version (without migrating them to scala 3), see how painful/not painful the process is. Hopefully I should be able to do that today.

Let me know when you have the chance to test it on a real app, that might also be of interest for @viperey as they had the need to a version that support scala 3.

ex0ns · 2025-07-06T12:19:07Z

@seroperson the CI is failing because of mill version, I will upgrade that right now

Edit: now the CI has the same stack overflow error you had before I believe. I don't have them locally

seroperson · 2025-07-06T13:16:37Z

@seroperson the CI is failing because of mill version, I will upgrade that right now

Edit: now the CI has the same stack overflow error you had before I believe. I don't have them locally

Well, that's weird, I don't have it locally too. Before CirceDecoders + CirceEncoders splitting, it occurred locally too, I guess this bug somehow depends on computing resources and amount of auto-generated things. I see now it fails on Message generating, which is heavy enough. I can try to implement the decoder manually.

And the dependent type instead of it being a type parameter looks very equivalent to me (even though I don't understand why this was working before and it does not anymore, maybe you have a guess about this ?)

I guess that's because previously all decoders were in scope explicitly because they were defined right in the single CirceDecoders and were explicitly imported. After I moved them to corresponding models, I guess it wasn't smart enough to find them in companion object of R. Actually it sounds fair.

ex0ns · 2025-07-06T17:02:17Z

Great ! CI is now green. I still need to find the time to upgrade my projects to test now

seroperson · 2025-07-06T17:04:09Z

Great ! CI is now green. I still need to find the time to upgrade my projects to test now

ofc, it's better to wait than release bad package. This week I'll try too I think.

core/src/com/bot4s/telegram/future/Polling.scala

ex0ns · 2025-07-06T21:09:41Z

I added a comment based on a test I just did.

The good news is that I upgraded one of my simple project and I had to change exactly 0 lines (public API seems to be untouched), I believe that I will still release it as 6.0.0 when we are sure that it works.

Changing Future.unit will null made the example (RandomBot) work

seroperson · 2025-07-07T16:59:12Z

I've migrated my bots to Scala3 and they seem run okay too 🎉At least locally.

viperey · 2025-07-07T17:14:53Z

core/src/com/bot4s/telegram/clients/SttpClient.scala

  private val apiBaseUrl = s"https://$telegramHost/bot$token/"

-  override def sendRequest[R, T <: BotRequest[_]](request: T)(implicit encT: Encoder[T], decR: Decoder[R]): F[R] = {
+  override def sendRequest[T <: BotRequest: Encoder](


@ex0ns out of curiosity, did your test hit this area of the code?

I would have to double check as I only did a very small set of tests, but I would say yes as RandomBot is based on ExampleBot which should be using the SttpClient.

Let me make sure of it 👌

Okay, that line change broke my private code, but, minor adjustments apart, it works.
In general the new version seems to be smoothly working, great job.

cc/ @seroperson

yea, it broke mine too just a little. Well, anyway I guess we're going to up major version.

Would you be able to share what you changed? I will include it in the release notes

Like this?:

- Circe Codecs from `com.bot4s.telegram.marshalling.CirceDecoders` and `com.bot4s.telegram.marshalling.CirceEncoders` were moved to corresponding models. - `com.bot4s.telegram.methods.Request`, `com.bot4s.telegram.methods.JsonRequest` and `com.bot4s.telegram.methods.MultipartRequest` aren't type-parameterized anymore, `R` type parameter was replaced with a dependent type `Response`.

I was more talking about what you had to change in the code of your bots to align with the new version, but I guess it was mostly removing type parameters ?

Oh, yea, well, just aligned sendRequest signatures with new changes. It'll only affect you in case of custom RequestHandler or if you're using Request type somewhere.

For example, I have a custom ZIORequestHandler, which I had to change like this:

ex0ns · 2025-07-10T21:10:53Z

Thanks a lot for the work ! I'm going to merge that. Due to some changes to the publication of the artifact, we will not have a snapshot release for this one yet, but I will make sure to publish version 6.0.0 before the end of the week

It will be hard to test/cover each an every scenario that this lib cover in the current state, but I updated a few of my bot, and ran the example available in this repo and they all worked as expected

ex0ns · 2025-07-14T19:58:15Z

I upgraded to mill 1.0 in the meantime to have working snapshot, I will now merge this on main and leave the snapshot to be tested for a few days before doing the proper release.

ex0ns · 2025-07-14T20:07:25Z

5.8.4-65f7d08f-SNAPSHOT is now available

seroperson · 2025-07-25T13:45:44Z

5.8.4-65f7d08f-SNAPSHOT is now available

Is it really available? Where can I find it?

[error]   not found: https://repo1.maven.org/maven2/com/bot4s/telegram-core_3/5.8.4-65f7d08f-SNAPSHOT/telegram-core_3-5.8.4-65f7d08f-SNAPSHOT.pom
[error]   not found: https://central.sonatype.com/repository/maven-snapshots/com/bot4s/telegram-core_3/5.8.4-65f7d08f-SNAPSHOT/telegram-core_3-5.8.4-65f7d08f-SNAPSHOT.pom
[error]   not found: https://oss.sonatype.org/content/repositories/snapshots/com/bot4s/telegram-core_3/5.8.4-65f7d08f-SNAPSHOT/telegram-core_3-5.8.4-65f7d08f-SNAPSHOT.pom

ex0ns · 2025-07-25T14:39:24Z

I'm sorry it's not, sonatype changed recently and it failed silently.
I'm getting The version cannot be a SNAPSHOT on maven publish, which I don't understand and it's not very documented... will investigate

https://mill-build.org/mill/scalalib/publishing.html#_snapshot_versions

ex0ns · 2025-08-09T14:23:53Z

For future reference:

This might be coming from mill

(in the 1.0.1 release). I will upgrade mill

ex0ns · 2025-08-09T15:04:17Z

@seroperson the CI ran:

 [526] Detected a 'SNAPSHOT' version for Artifact(com.bot4s,telegram-core_3,6.0.0-SNAPSHOT), publishing to Sonatype Central Snapshots at 'https://central.sonatype.com/repository/maven-snapshots/'
[526] Publishing to 'https://central.sonatype.com/repository/maven-snapshots/' finished with result: DeployResult(
[526]   artifacts = [
[526]     com.bot4s:telegram-core_3:pom:6.0.0-20250809.145419-1
[526]     com.bot4s:telegram-core_3:jar:6.0.0-20250809.145419-1
[526]     com.bot4s:telegram-core_3:jar:sources:6.0.0-20250809.145419-1
[526]     com.bot4s:telegram-core_3:jar:javadoc:6.0.0-20250809.145419-1
[526]   ]
[526]   metadatas = [
[526]     com.bot4s:telegram-core_3:6.0.0-SNAPSHOT/maven-metadata.xml
[526]     com.bot4s:telegram-core_3/maven-metadata.xml
[526]   ]
[526] )

seroperson · 2025-08-09T15:17:02Z

@seroperson the CI ran:

Thank you!

ex0ns and others added 7 commits July 3, 2025 11:26

[WIP] Migration to scala 3

b76533b

Fixing migration errors

e1325b4

Splitting CirceDecoders

3045a65

Splitting CirceEncoders

bf55b75

Using dependent types for Request

b9bd2d2

Minor cleanup + scalafmt

5e3bb55

Merge branch 'main' into scala3-new

413ee93

Merge branch 'main' into scala3-new

b062e17

Defining Encoder + Decoder manually for Message

d85daee

Merge branch 'main' into scala3-new

f67afd9

ex0ns reviewed Jul 6, 2025

View reviewed changes

core/src/com/bot4s/telegram/future/Polling.scala Outdated Show resolved Hide resolved

Replace Future.unit with null in Polling

1ab0fe8

viperey reviewed Jul 7, 2025

View reviewed changes

Merge branch 'main' into scala3-new

da79edf

ex0ns merged commit 65f7d08 into bot4s:main Jul 14, 2025
1 check passed

Scala3 upgrade #488

Scala3 upgrade #488

Uh oh!

Conversation

seroperson commented Jul 6, 2025

Uh oh!

ex0ns commented Jul 6, 2025

Uh oh!

ex0ns commented Jul 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

seroperson commented Jul 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ex0ns commented Jul 6, 2025

Uh oh!

seroperson commented Jul 6, 2025

Uh oh!

Uh oh!

ex0ns commented Jul 6, 2025

Uh oh!

seroperson commented Jul 7, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

viperey Jul 7, 2025

Choose a reason for hiding this comment

Uh oh!

ex0ns Jul 7, 2025

Choose a reason for hiding this comment

Uh oh!

viperey Jul 8, 2025

Choose a reason for hiding this comment

Uh oh!

seroperson Jul 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ex0ns Jul 8, 2025

Choose a reason for hiding this comment

Uh oh!

seroperson Jul 8, 2025

Choose a reason for hiding this comment

Uh oh!

ex0ns Jul 8, 2025

Choose a reason for hiding this comment

Uh oh!

seroperson Jul 8, 2025

Choose a reason for hiding this comment

Uh oh!

ex0ns commented Jul 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ex0ns commented Jul 14, 2025

Uh oh!

Uh oh!

ex0ns commented Jul 14, 2025

Uh oh!

seroperson commented Jul 25, 2025

Uh oh!

ex0ns commented Jul 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ex0ns commented Aug 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ex0ns commented Aug 9, 2025

Uh oh!

seroperson commented Aug 9, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

ex0ns commented Jul 6, 2025 •

edited

Loading

seroperson commented Jul 6, 2025 •

edited

Loading

seroperson commented Jul 7, 2025 •

edited

Loading

seroperson Jul 8, 2025 •

edited

Loading

ex0ns commented Jul 10, 2025 •

edited

Loading

ex0ns commented Jul 25, 2025 •

edited

Loading

ex0ns commented Aug 9, 2025 •

edited

Loading