docs/src/upgrade/release-3.5.x.asciidoc

////
Licensed to the Apache Software Foundation (ASF) under one or more
contributor license agreements.  See the NOTICE file distributed with
this work for additional information regarding copyright ownership.
The ASF licenses this file to You under the Apache License, Version 2.0
(the "License"); you may not use this file except in compliance with
the License.  You may obtain a copy of the License at

  http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
////

= TinkerPop 3.5.0

image::https://raw.githubusercontent.com/apache/tinkerpop/master/docs/static/images/gremlin-sleeping-beauty.png[width=225]

*The Sleeping Gremlin: No. 18 Entr'acte symphonique*

== TinkerPop 3.5.0

*Release Date: NOT OFFICIALLY RELEASED YET*

Please see the link:https://github.com/apache/tinkerpop/blob/3.5.0/CHANGELOG.asciidoc#release-3-5-0[changelog] for a complete list of all the modifications that are part of this release.

=== Upgrading for Users

==== Gryo Usage

Since the first release of TinkerPop 3.x, Gryo has been the default serialization format for Gremlin Server and
Java Driver. It was also used as the default serialization format for Gremlin Console remote connectivity to Gremlin
Server. As of this release, Gryo has been replaced as the default by GraphBinary. All packaged configuration files
and programmatic defaults have been modified as such.

It is still possible to utilize Gryo as a message serialization format by modifying Gremlin Server configuration files
to include the appropriate Gryo configurations. If using Gryo, do not user earlier versions of the driver and server
with 3.5.0. Use a 3.5.0 client to connect to a 3.5.0 server. Generally speaking, mixed version combinations will
appear to work properly, but problems will likely occur during general course of usage and it is therefore not
advisable to take this approach.

For best compatibility between 3.4.x and 3.5.x, please use GraphBinary.

link:https://issues.apache.org/jira/browse/TINKERPOP-2259[TINKERPOP-2259]

==== Configuration Upgrade

There is a major breaking change in the use of `Configuration` objects. Prior to 3.5.0, `Configuration` objects were
from the Apache Commons `commons-configuration` library, but in this version, they are of `commons-configuration2`.
While this is a breaking change, the fix for most implementations will be quite simple, Simply, change the import
statements from:

[source,text]
----
org.apache.commons.configuration.*
----

to

[source,text]
----
org.apache.commons.configuration2.*
----

It is also worth noting that default list handling in configurations is treated differently. TinkerPop largely
disabled the default list handling approach in `Configuration` 1.x, but if that functionality is still needed, it can
be reclaimed by setting the `LegacyListDelimiterHandler` - details for doing taking this step and other relevant
upgrade information can be found in the link:https://commons.apache.org/proper/commons-configuration/userguide/upgradeto2_0.html[2.x Upgrade Documentation].

See: link:https://issues.apache.org/jira/browse/TINKERPOP-2185[TINKERPOP-2185]

==== Use of null

Gremlin has traditionally disallowed `null` as a value in traversals and not always in consistent ways:

[source,text]
----
gremlin> g.inject(1, null, null, 2, null)
java.lang.NullPointerException
Type ':help' or ':h' for help.
Display stack trace? [yN]n
gremlin> g.V().has('person','name','marko').property('age', null)
The AddPropertyStep does not have a provided value: AddPropertyStep({key=[age]})
Type ':help' or ':h' for help.
Display stack trace? [yN]
gremlin> g.addV("person").property("name", 'stephen').property("age", null)
==>v[13]
gremlin> g.V().has('person','name','stephen').elementMap()
==>[id:13,label:person,name:stephen]
gremlin> g.V().constant(null)
gremlin>
----

Note how `null` can produce exception behavior or act as a filter. For 3.5.0, TinkerPop has not only made `null` usage
consistent, but has also made it an allowable value within a `Traversal`:

[source,text]
----
gremlin> g.inject(1, null, null, 2, null)
==>1
==>null
==>null
==>null
==>2
gremlin> g.V().constant(null)
==>null
==>null
==>null
==>null
==>null
==>null
----

TinkerGraph can be configured to support `null` as a property value and all graphs may not support this feature (for
example, Neo4j does not). Please be sure to check the new `supportsNullPropertyValues()` feature (or the documentation
of the graph provider) to determine if the `Graph` implementation allows `null` as a property value.

With respect to `null` in relation to properties, there was a bit of inconsistency in the handling of `null` in calls
to `property()` depending on the type of mutation being executed demonstrated as follows in earlier versions:

[source,text]
----
gremlin> g.V(1).property("x", 1).property("y", null).property("z", 2)
The AddPropertyStep does not have a provided value: AddPropertyStep({key=[y]})
Type ':help' or ':h' for help.
Display stack trace? [yN]N
gremlin> g.addV("test").property("x", 1).property("y", null).property("z", 2)
==>v[13]
gremlin> g.V(13).properties()
==>vp[x->1]
==>vp[z->2]
----

This behavior has been altered to become consistent. First, assuming `null` is not supported as a property value, the
setting of a property to `null` should have the behavior of removing the property in the same way in which you might
do `g.V().properties().drop()`:

[source,text]
----
gremlin> g.V(1).property("x", 1).property("y", null).property("z", 2)
==>v[1]
gremlin> g.V(1).elementMap()
==>[id:1,label:person,name:marko,x:1,z:2,age:29]
gremlin> g.V().hasLabel('person').property('age',null).iterate()
gremlin> g.V().hasLabel('person').elementMap()
==>[id:1,label:person,name:marko]
==>[id:2,label:person,name:vadas]
==>[id:4,label:person,name:josh]
==>[id:6,label:person,name:peter]
----

Then, assuming `null` is supported as a property value:

[source,text]
----
gremlin> g.addV("person").property("name", 'stephen').property("age", null)
==>v[13]
gremlin> g.V().has('person','name','stephen').elementMap()
==>[id:13,label:person,name:stephen,age:null]
gremlin> g.V().has('person','age',null)
==>v[13]
----

In conclusion, this change in greater support of `null` may affect the behavior of existing traversals written in past
versions of TinkerPop as it is no longer possible to rely on `null` to expect a filtering action for traversers.
Please review existing Gremlin carefully to ensure that there are no unintended consequences of this change and that
there are no opportunities to improve existing logic to take greater advantage of this expansion of `null` semantics.

See: link:https://issues.apache.org/jira/browse/TINKERPOP-2235[TINKERPOP-2235],
link:https://issues.apache.org/jira/browse/TINKERPOP-2099[TINKERPOP-2099]

==== Remote SideEffects

Remote traversals no longer support the retrieval of remote side-effects. Users must therefore directly return
side-effects as part of their query if they need that data. Note that server settings for `TraversalOpProcessor`, which
formerly held the cache for these side-effects, no longer have any effect and can be removed.

See: link:https://issues.apache.org/jira/browse/TINKERPOP-2269[TINKERPOP-2269]

==== Deprecation Removal

The following deprecated classes, methods or fields have been removed in this version:

* `gremlin-core`
** `org.apache.tinkerpop.gremlin.process.computer.bulkdumping.BulkDumperVertexProgram`
** `org.apache.tinkerpop.gremlin.process.computer.bulkloading.BulkLoader`
** `org.apache.tinkerpop.gremlin.process.computer.bulkloading.BulkLoaderVertexProgram`
** `org.apache.tinkerpop.gremlin.process.computer.bulkloading.IncrementalBulkLoader`
** `org.apache.tinkerpop.gremlin.process.computer.bulkloading.OneTimeBulkLoader`
** `org.apache.tinkerpop.gremlin.process.computer.clustering.peerpressure.PeerPressureVertexProgram.Builder#traversal(*)`
** `org.apache.tinkerpop.gremlin.process.computer.ranking.pagerank.PageRankVertexProgram.Builder#traversal(*)`
** `org.apache.tinkerpop.gremlin.process.computer.ranking.pagerank.PageRankVertexProgram.Builder#vertexCount()`
** `org.apache.tinkerpop.gremlin.process.computer.traversal.step.map.PageRankVertexProgramStep.modulateBy(*)`
** `org.apache.tinkerpop.gremlin.process.computer.traversal.step.map.PageRankVertexProgramStep.modulateTimes()`
** `org.apache.tinkerpop.gremlin.process.computer.traversal.step.map.PeerPressureVertexProgramStep.modulateBy(*)`
** `org.apache.tinkerpop.gremlin.process.computer.traversal.step.map.PeerPressureVertexProgramStep.modulateTimes()`
** `org.apache.tinkerpop.gremlin.process.remote.traversal.AbstractRemoteTraversalSideEffects`
** `org.apache.tinkerpop.gremlin.process.remote.traversal.EmbeddedRemoteTraversalSideEffects`
** `org.apache.tinkerpop.gremlin.process.remote.traversal.RemoteTraversalSideEffects`
** `org.apache.tinkerpop.gremlin.process.remote.traversal.RemoteTraversal#getSideEffects()`
** `org.apache.tinkerpop.gremlin.process.traversal.Order.decr`
** `org.apache.tinkerpop.gremlin.process.traversal.Order.incr`
** `org.apache.tinkerpop.gremlin.process.traversal.TraversalSource#withRemote(*)`
** `org.apache.tinkerpop.gremlin.process.traversal.dsl.graph.GraphTraversalSource#withRemote(*)`
** `org.apache.tinkerpop.gremlin.process.traversal.step.map.PropertyMapStep(Traversal.Admin, boolean, PropertyType, String...)`
** `org.apache.tinkerpop.gremlin.process.traversal.step.map.PropertyMapStep#isIncludeTokens()`
** `org.apache.tinkerpop.gremlin.structure.util.star.StarGraph#builder()`
** `org.apache.tinkerpop.gremlin.structure.util.star.StarGraph.Builder#create()`
* `gremlin-driver`
** `org.apache.tinkerpop.gremlin.driver.Tokens#ARGS_SCRIPT_EVAL_TIMEOUT`
** `org.apache.tinkerpop.gremlin.driver.Cluster.Builder#keyCertChainFile(String)`
** `org.apache.tinkerpop.gremlin.driver.Cluster.Builder#keyFile(String)`
** `org.apache.tinkerpop.gremlin.driver.Cluster.Builder#keyPassword(String)`
** `org.apache.tinkerpop.gremlin.driver.Cluster.Builder#trustCertificateChainFile(String)`
** `org.apache.tinkerpop.gremlin.driver.remote.DriverRemoteTraversalSideEffects`
** `org.apache.tinkerpop.gremlin.driver.remote.DriverRemoteTraversal#getSideEffects()`
* `gremlin-server`
** `org.apache.tinkerpop.gremlin.server.Settings.scriptEvaluationTimeout`
** `org.apache.tinkerpop.gremlin.server.Settings.SslSettings.keyCertChainFile`
** `org.apache.tinkerpop.gremlin.server.Settings.SslSettings.keyFile`
** `org.apache.tinkerpop.gremlin.server.Settings.SslSettings.keyPassword`
** `org.apache.tinkerpop.gremlin.server.Settings.SslSettings.trustCertificateChainFile`
** `org.apache.tinkerpop.gremlin.server.ResponseHandlerContext`
** `org.apache.tinkerpop.gremlin.server.op.AbstractEvalOpProcessor.evalOpInternal(ResponseHandlerContext, Supplier, BindingSupplier)`
** `org.apache.tinkerpop.gremlin.server.op.AbstractOpProcessor.generateMetaData(ChannelHandlerContext, RequestMessage, ResponseStatusCode, Iterator)`
** `org.apache.tinkerpop.gremlin.server.op.AbstractOpProcessor.handleIterator(ResponseHandlerContext, Iterator)`
** `org.apache.tinkerpop.gremlin.server.op.AbstractOpProcessor.makeFrame(ChannelHandlerContext, RequestMessage, MessageSerializer, boolean, List, ResponseStatusCode, Map)`
** `org.apache.tinkerpop.gremlin.server.op.AbstractOpProcessor.makeFrame(Context, RequestMessage, MessageSerializer, boolean, List, ResponseStatusCode, Map)`
** `org.apache.tinkerpop.gremlin.server.op.AbstractOpProcessor.makeFrame(ResponseHandlerContext, RequestMessage, MessageSerializer, boolean, List, ResponseStatusCode, Map)`
** `org.apache.tinkerpop.gremlin.server.op.AbstractOpProcessor.makeFrame(ResponseHandlerContext, RequestMessage, MessageSerializer, boolean, List, ResponseStatusCode, Map, Map)`
** `org.apache.tinkerpop.gremlin.server.util.SideEffectIterator`
* `neo4j-gremlin`
** `org.apache.tinkerpop.gremlin.neo4j.structure.Neo4jGraph#getTrait()`
** `org.apache.tinkerpop.gremlin.neo4j.structure.Neo4jGraph#CONFIG_META_PROPERTIES`
** `org.apache.tinkerpop.gremlin.neo4j.structure.Neo4jGraph#CONFIG_MULTI_PROPERTIES `
** `org.apache.tinkerpop.gremlin.neo4j.structure.trait.MultiMetaNeo4jTrait`
** `org.apache.tinkerpop.gremlin.neo4j.structure.trait.NoMultiNoMetaNeo4jTrait`
** `org.apache.tinkerpop.gremlin.neo4j.structure.trait.Neo4jTrait`

Certain elements of the API were not or could not be deprecated in prior versions and were simply renamed for this
release:

* `org.apache.tinkerpop.gremlin.driver.message.ResponseStatusCode#SERVER_ERROR_SCRIPT_EVALUATION` became `SERVER_ERROR_EVALUATION`

See: link:https://issues.apache.org/jira/browse/TINKERPOP-2080[TINKERPOP-2080],
link:https://issues.apache.org/jira/browse/TINKERPOP-2231[TINKERPOP-2231],
link:https://issues.apache.org/jira/browse/TINKERPOP-2233[TINKERPOP-2233],
link:https://issues.apache.org/jira/browse/TINKERPOP-2239[TINKERPOP-2239],
link:https://issues.apache.org/jira/browse/TINKERPOP-2269[TINKERPOP-2269],
link:https://issues.apache.org/jira/browse/TINKERPOP-2273[TINKERPOP-2273],
link:http://tinkerpop.apache.org/docs/3.5.0/upgrade/#_ssl_security[3.2.10 Upgrade Documentation for SSL]

=== Upgrading for Provider

==== Graph System Providers

===== ScalarMapStep

`MapStep` had a single abstract method that needed to be implemented:

[source,java]
----
protected abstract E map(final Traverser.Admin<S> traverser);
----

This method made it easy to implement new implementations because it hid certain processing logic and made it so that
the implementer only had to reason about how to take the current object from the `Traverser` and transform it to a
new value. As 3.5.0 changed semantics around how `null` is processed, this method became a bit of a hindrance to the
more complex logic which those semantics entailed. Specifically, this method could not easily communicate to underlying
processing what a `null` might mean - is the `null` the end of the traversal stream or should the `null` be promoted
down the stream as a value to be processed.

Interestingly, the method that enabled the handling of this more complex decision making already existed in
`AbstractStep`:

[source,java]
----
protected Traverser.Admin<E> processNextStart()
----

It returns a whole `Traverser` object and forces manual retrieval of the "next" `Traverser`. At this level it becomes
possible to make choices on `null` and return it if it should be propagated or dismiss it and return an
`EmptyTraverser`. To better accommodate the `MapStep` which provides the nice helper `map(Traverser)` method as well
as the more flexible version that doesn't need that infrastructure, `ScalarMapStep` was added to extend `MapStep`. The
`map(Traverser)` was then moved to `ScalarMapStep` and those steps that could rely on that helper method now extend
from it. All other steps of this sort still extend `MapStep` and directly implement `processNextStart()`.

Providers will get compile errors if they extended `MapStep`. The easy solution will be to simply modify that code so
that their step instead extends `ScalarMapStep`. As a secondary task, providers should then examine their step
implementation to ensure that `null` semantics as presented in 3.5.0 apply properly. If they do not, then it is likely
that the step should simply implement `MapStep` directly and former `map(Traverser)` logic should be migrated to
`processNextStart()`.

See: link:https://issues.apache.org/jira/browse/TINKERPOP-2235[TINKERPOP-2235],
link:https://issues.apache.org/jira/browse/TINKERPOP-2099[TINKERPOP-2099]

===== TraversalStrategy Application

The methodology for strategy application has been altered and the change is most easily described by example. Given a
traversal with the structure:

[source,text]
----
a(b(),c(d()))
----

Strategies were formerly applied in the following order:

[source,text]
----
StrategyA on a
StrategyB on a
StrategyA on b
StrategyB on b
StrategyA on c
StrategyB on c
StrategyA on d
StrategyB on d
----

This approach has always prevented strategies from performing global operations across the traversal and all decedents
effectively as children will not have been processed by preceding strategies yet. As of this release, the approach
has been altered to apply strategies as follows:

[source,text]
----
StrategyA on a
StrategyA on b
StrategyA on c
StrategyA on d
StrategyB on a
StrategyB on b
StrategyB on c
StrategyB on d
----

In this way, strategy B can check if it is being applied to the root traversal and if it is it knows that A has been
applied globally.

This revised methodology could represent a breaking change for `TraversalStrategy` implementations if they somehow
relied on the old ordering of application. It may also present an opportunity to revise how a `TraversalStrategy` is
written to gain some processing benefit to the new order. Please be sure to review any custom strategies carefully
when upgrading to this version.

As part of this change, there have been some adjustments to the `Traversal` and `Traversal.Admin` interfaces which have
helped to clarify coding intent. There is now an `isRoot()` method which determines whether or not the traversal has a
parent or not. Under revised semantics for 3.5.0, a traversal's parent must be an `EmptyStep` instance and should not
be `null`. With this change, provider `TraversalStrategy` implementations should be reviewed to evaluate if `isRoot()`
semantics cause any breaks in logic to existing code.

In addition, `TraversalStrategies` now implements `Iterable` and exposes an `iterator()` method which may be preferred
over the old `toList()` style construction for getting the list of configured strategies.

See: link:https://issues.apache.org/jira/browse/TINKERPOP-1568[TINKERPOP-1568],
link:https://issues.apache.org/jira/browse/TINKERPOP-2310[TINKERPOP-2310],
link:https://issues.apache.org/jira/browse/TINKERPOP-2311[TINKERPOP-2311]

===== Null Semantics

Graph providers should take note of the changes to `null` semantics described in the "users" section of these upgrade
notes. As `null` is now acceptable as a `Traverser` object, this change may affect custom steps. Further note that
`null` now works more consistently with mutation steps and graph providers may need to include additional logic to
deal with those possible conditions. Please see the console sessions below which uses TinkerGraph to demonstrate the
current behavioral expectations.

[source,text]
----
gremlin> g.getGraph().features().vertex().supportsNullPropertyValues()
==>false
gremlin> g.addV(null).property(id, null).property('name',null)
==>v[0]
gremlin> g.V().elementMap()
==>[id:0,label:vertex]
...
gremlin> g.getGraph().features().vertex().supportsNullPropertyValues()
==>true
gremlin> g.addV(null).property(id, null).property('name',null)
==>v[0]
gremlin> g.V().elementMap()
==>[id:0,label:vertex,name:null]
----

In the above example, `addV()` defaults to `Vertex.DEFAULT_LABEL`, the `id` is generated and setting the "name"
property to `null` results in the value not being set. If the property value is set to an actual value and then set
to `null` TinkerGraph will remove the property key all together:

[source,text]
----
gremlin> g.getGraph().features().vertex().supportsNullPropertyValues()
==>false
gremlin> g.addV().property('name','stephen')
==>v[0]
gremlin> g.V().elementMap()
==>[id:0,label:vertex,name:stephen]
gremlin> g.V().has('vertex','name','stephen').property('name',null)
==>v[0]
gremlin> g.V().elementMap()
==>[id:0,label:vertex]
...
gremlin> g.getGraph().features().vertex().supportsNullPropertyValues()
==>true
gremlin> g.addV().property('name','stephen')
==>v[2]
gremlin> g.V().has('vertex','name','stephen').property('name',null)
==>v[2]
gremlin> g.V().elementMap()
==>[id:2,label:vertex,name:null]
----

The above examples point out the default operations of TinkerGraph, but it can be configured to actually accept the
`null` as a property value and it is up to graph providers to decided how they wish to treat a `null` property value.
Providers should use the new `supportsNullPropertyValues()` feature to indicate to users how `null` is handled.

For edges, the `label` still cannot be defaulted and must be specified, therefore:

[source,text]
----
gremlin> g.V(0L).as('a').addE(null).to('a')
Label can not be null
Type ':help' or ':h' for help.
Display stack trace? [yN]n
gremlin> g.V(0L).as('a').addE(constant(null)).to('a')
Label can not be null
Type ':help' or ':h' for help.
Display stack trace? [yN]
----

Also, edges have similar behavior to vertices when it comes to setting properties (again, the default configuration for
TinkerGraph is being used here):

[source,text]
----
gremlin> g.getGraph().features().vertex().supportsNullPropertyValues()
==>false
gremlin> g.addV().property('name','stephen')
==>v[0]
gremlin> g.V().has('vertex','name','stephen').as('a').addE('knows').to('a').property(id,null).property('weight',null)
==>e[2][0-knows->0]
gremlin> g.E().elementMap()
==>[id:2,label:knows,IN:[id:0,label:vertex],OUT:[id:0,label:vertex]]
gremlin> g.E().property('weight',0.5)
==>e[2][0-knows->0]
gremlin> g.E().elementMap()
==>[id:2,label:knows,IN:[id:0,label:vertex],OUT:[id:0,label:vertex],weight:0.5]
gremlin> g.E().property('weight',null)
==>e[2][0-knows->0]
gremlin> g.E().elementMap()
==>[id:2,label:knows,IN:[id:0,label:vertex],OUT:[id:0,label:vertex]]
...
gremlin> g.getGraph().features().vertex().supportsNullPropertyValues()
==>true
gremlin> g.addV().property('name','stephen')
==>v[8]
gremlin> g.V().has('vertex','name','stephen').as('a').addE('knows').to('a').property(id,null).property('weight',null)
==>e[10][8-knows->8]
gremlin> g.E().elementMap()
==>[id:10,label:knows,IN:[id:8,label:vertex],OUT:[id:8,label:vertex],weight:null]
gremlin> g.E().property('weight',0.5)
==>e[10][8-knows->8]
gremlin> g.E().elementMap()
==>[id:10,label:knows,IN:[id:8,label:vertex],OUT:[id:8,label:vertex],weight:0.5]
gremlin> g.E().property('weight',null)
==>e[10][8-knows->8]
gremlin> g.E().elementMap()
==>[id:10,label:knows,IN:[id:8,label:vertex],OUT:[id:8,label:vertex],weight:null]
----

Graphs that support multi/meta-properties have some issues to consider as well as demonstrated with TinkerGraph:

[source,text]
----
gremlin> g.getGraph().features().vertex().supportsNullPropertyValues()
==>false
gremlin> g.addV().property(list,'foo',"x").property(list,"foo", null).property(list,'foo','bar')
==>v[0]
gremlin> g.V().elementMap()
==>[id:0,label:vertex,foo:bar]
gremlin> g.V().valueMap()
==>[foo:[x,bar]]
gremlin> g.V().property('foo',null)
==>v[0]
gremlin> g.V().valueMap(true)
==>[id:0,label:vertex]
...
gremlin> g.addV().property(list,'foo','bar','x',1,'y',null)
==>v[0]
gremlin> g.V().properties('foo').valueMap(true)
==>[id:1,key:foo,value:bar,x:1]
gremlin> g.V().properties('foo').property('x',null)
==>vp[foo->bar]
gremlin> g.V().properties('foo').valueMap(true)
==>[id:1,key:foo,value:bar]
...
gremlin> g.getGraph().features().vertex().supportsNullPropertyValues()
==>false
gremlin> g.addV().property(list,'foo',"x").property(list,"foo", null).property(list,'foo','bar')
==>v[11]
gremlin> g.V().elementMap()
==>[id:11,label:vertex,foo:bar]
gremlin> g.V().valueMap()
==>[foo:[x,null,bar]]
...
gremlin> g.addV().property(list,'foo','bar','x',1,'y',null)
==>v[0]
gremlin> g.V().properties('foo').valueMap(true)
==>[id:1,key:foo,value:bar,x:1,y:null]
gremlin> g.V().properties('foo').property('x',null)
==>vp[foo->bar]
gremlin> g.V().properties('foo').valueMap(true)
==>[id:1,key:foo,value:bar,x:null,y:null]
----

See: link:https://issues.apache.org/jira/browse/TINKERPOP-2235[TINKERPOP-2235],
link:https://issues.apache.org/jira/browse/TINKERPOP-2099[TINKERPOP-2099]

===== AbstractOpProcessor API Change

The `generateMetaData()` method was removed as it was deprecated in a previous version. There already was a preferred
method called `generateResultMetaData()` that took an extra `Settings` parameter. To fix compilation issues simply
replace implementations of the `generateMetaData()` method with `generateResultMetaData()`. Gremlin Server has
only been calling `generateResultMetaData()` since the deprecation, so this correction should be straightforward.

===== StoreStep and AggregateStep

Note that `StoreStep` has been renamed to `AggregateLocalStep` and `AggregateStep` has been renamed to
`AggregateGlobalStep`. The renaming is important to consider if any custom `TraversalStrategies` have been written
that rely on the old step names.

See: link:https://issues.apache.org/jira/browse/TINKERPOP-2254[TINKERPOP-2254]

==== Graph Driver Providers

===== TraversalOpProcessor Side-effects

`TraversalOpProcessor` no longer holds a cache of side-effects and more generally the entire side-effect protocol has
been removed and is no longer supported in the server or drivers.

See: link:https://issues.apache.org/jira/browse/TINKERPOP-2269[TINKERPOP-2269]