SecureRandom.javaSecuritySecureRandom for two effect types

[dpevunov-cp]

Im not sure this is the right and idiomatic way to implement this, but this greatly helps in real app wiring when you have to create instance for some effect G(for example ConnectioIO) when the wiring is done in F(IO for example)

This is my first PR here, please let me know if this should be done somehow differently

[dpevunov-cp]

@armanbilge @djspiewak Could you have a look please?

[BalmungSan]

Maybe add some docs to both versions?
Similar to Ref.in, also I think that in name is semi standard for this kind of F & G API: https://typelevel.org/cats-effect/api/3.x/cats/effect/kernel/Ref$.html#in[F[_],G[_],A](a:A)(implicitF:cats.effect.kernel.Sync[F],implicitG:cats.effect.kernel.Sync[G]):F[cats.effect.kernel.Ref[G,A]]

[dpevunov-cp]

Maybe add some docs to both versions? Similar to Ref.in, also I think that in name is semi standard for this kind of F & G

I have added docs. But thinking of .in Im not sure it fits here moreover it would be incompatible and need to release in new MINOR version. I would like to have it in new PATCH

[armanbilge]

it would be incompatible and need to release in new MINOR version. I would like to have it in new PATCH

Unfortunately, adding a new API always breaks forward compatibility, no matter what it's called. We won't be able to get this into a patch, but I've added it to the milestone for the next minor.

---

Similar to Ref.in, also I think that in name is semi standard for this kind of F & G API

Yes, this is a good point. Let's follow this convention.

[armanbilge]

but this greatly helps in real app wiring when you have to create instance for some effect G(for example ConnectioIO) when the wiring is done in F(IO for example)

Btw, there is probably a workaround: create a SecureRandom in ConnectionIO and then mapK it to IO (or vice versa).

[geny200]

@dpevunov-cp Do you mind if I take PR and finish it? (as a part of the #4410 )

[dpevunov-cp]

@dpevunov-cp Do you mind if I take PR and finish it? (as a part of the #4410 )

I don't mind but I can try to update PR today

[dpevunov-cp]

@dpevunov-cp Do you mind if I take PR and finish it? (as a part of the #4410 )

@geny200 I have updated PR but for some reason some ci build tasks failed and I could not reproduce it locally. Could you take a look please?

[geny200]

I tried to reproduce it, and and I couldn't either. It looks like a flaky test to me, because this functionality has nothing to do with your changes.

Failed job - https://github.com/typelevel/cats-effect/actions/runs/15239205886/job/42857139080?pr=4338
Failed test - cats.effect.std.SupervisorSuite.async - restart / cancel race
Error - java.util.concurrent.TimeoutException: test timed out after 20 seconds