Feature/interrupt with algebra

[pchlupacek]

@mpilquist @SystemFw there is the interrupt implemented with Algebra.

The idea is pretty simple, I had it on mind couple of weeks, but always failed to implement up to until now :-). Essentially the idea is that scope that is first to be interrupted is caching it`s next step in case of interruption and using that step when the interruption occurs.

Also as part of this -

• Interrupted exception is gone
• Scope now takes type O
• ++, repeat, flatMap reverted to M10 implementations

The implementation is bit noisy (if you can find a way how to make it cleaner I will really appreciate). Going forward I am thinking of introducing new type Compile[F, O] that will Contain current Scope[F, O] to cleanup algebra and also allow different compilation implementations to be plugged in (i.e. debugging compiler, javascript-optimized compiler etc..)

[pchlupacek]

FWIW I think the travis failure is unrelated

[SystemFw]

I think some instability in the retry spec is unavoidable (is testing repetitions in time), let's see if the failure is indeed a random one

[pchlupacek]

@SystemFw I think the issue is that delays is mutable unguarded ListBuffer. I think that spec needs to be made stronger, as the delays are accesses potentially from multiple threads.

[mpilquist]

Can the TODO be resolved by specifying UO as type param to getScope (instead of Nothing)?

[pchlupacek]

Unfortunatelly the result of unconsAsync is Pull[F, Nothing, AsyncPull[F, Option[(Segment[O, Unit], Stream[F, O])]]] so I would need then typecase UO back to Nothing, at least that's what compiler is asking me to do.
I have plan to address in #1056.

[pchlupacek]

@mpilquist @SystemFw pls let me know when you will be done reviewing this. I have ready merge pull request that uses new interruption, and can just make PR once this is merged.

[SystemFw]

I'm in airports all day today, so it's more likely going to be tomorrow. However, feel free to go ahead with only Michael's approval, I'm still catching up on the core interpreter stuff

…

On 7 Jan 2018 17:20, "Pavel Chlupacek" ***@***.***> wrote: @mpilquist <https://github.com/mpilquist> @SystemFw <https://github.com/systemfw> pls let me know when you will be done reviewing this. I have ready merge pull request that uses new interruption, and can just make PR once this is merged. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#1059 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AHaN7rj5zGU-iiUhstc3bi03pEbrHOXzks5tIPzRgaJpZM4RVX30> .