Remove Binary format and related members.

[carlosalberto]

As part of 0.3.0 we will be removing Binary (and later, under 0.5.0, putting it back once the format gets updated).

This is part of #720

[codecov-io]

Codecov Report

❗ No coverage uploaded for pull request base (master@9aa6d1e). Click here to learn what that means.
The diff coverage is n/a.

@@            Coverage Diff            @@
##             master     #891   +/-   ##
=========================================
  Coverage          ?   82.82%           
  Complexity        ?      830           
=========================================
  Files             ?      110           
  Lines             ?     2870           
  Branches          ?      249           
=========================================
  Hits              ?     2377           
  Misses            ?      382           
  Partials          ?      111

Impacted Files	Coverage Δ	Complexity Δ
...a/io/opentelemetry/opentracingshim/TracerShim.java	62.96% <ø> (ø)	11 <0> (?)
.../io/opentelemetry/opentracingshim/Propagation.java	100% <ø> (ø)	5 <0> (?)
...ationcontext/DefaultCorrelationContextManager.java	69.23% <ø> (ø)	7 <0> (?)
...rrelationcontext/CorrelationContextManagerSdk.java	100% <ø> (ø)	5 <0> (?)
...ain/java/io/opentelemetry/trace/DefaultTracer.java	95.23% <ø> (ø)	6 <0> (?)
...ain/java/io/opentelemetry/sdk/trace/TracerSdk.java	100% <ø> (ø)	8 <0> (?)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9aa6d1e...4873ed4. Read the comment docs.

[Oberon00]

Would it be easier to leave BinaryContext in a non-working form here? If we temporarily remove it, we break users now and later they have to put it back. If we don't remove it, we break them later and they can change their usage to the new interface right away.

[carlosalberto]

I'm not sure we will be breaking users, because I don't think anybody has been using this (for example, we don't have binary support for CorrelationContext and the one for tracing is rather experimental).

In JS they decided to keep it because there were pieces already using it, so it makes sense for them to keep things around, but that's not our case. I can keep them around, but it feels weird to me to keep around 'dead-for-now' code ;)

[Oberon00]

We plan to use the BinaryFormat interface at Dynatrace, with our own implementation. But if you think no one else is using it, we don't really require having that interface in the API.

[carlosalberto]

I don't think anybody else than you guys is using it at this time. I imagine users relying on that once we have a full implementation though (not a partial one like the one we have right now).

I'm adding then the "Holding On" label, in case we don't move fast enough on the OTEP 66 changes ;)

[carlosalberto]

Ooops, was expected to put the label to #914

@bogdandrutu Please provide your review/opinion on this one ;)

[bogdandrutu]

Please rebase, also I am not sure what was the decision.

[carlosalberto]

Rebased and updated. The decision from the last SIG call was to temporary remove it.

@open-telemetry/java-approvers please review & approve.