[ 
https://issues.apache.org/jira/browse/TINKERPOP-967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16510966#comment-16510966
 ] 

ASF GitHub Bot commented on TINKERPOP-967:
------------------------------------------

Github user GCHQResearcher1337 commented on a diff in the pull request:

    https://github.com/apache/tinkerpop/pull/876#discussion_r195046349
  
    --- Diff: 
gremlin-test/src/main/java/org/apache/tinkerpop/gremlin/process/traversal/step/branch/RepeatTest.java
 ---
    @@ -386,5 +456,26 @@ public void 
g_V_hasXname_markoX_repeatXoutE_inV_simplePathX_untilXhasXname_rippl
             public Traversal<Vertex, Path> 
get_g_V_hasXloop_name_loopX_repeatXinX_timesX5X_path_by_name() {
                 return 
g.V().has("loops","name","loop").repeat(__.in()).times(5).path().by("name");
             }
    +
    +        @Override
    +        public Traversal<Vertex, Path> 
get_g_V_repeatXout_repeatXoutX_timesX1XX_timesX1X_limitX1X_path_by_name() {
    +            // NB We need to prevent the RepeatUnrollStrategy from 
applying to properly exercise this test as this traversal can be simplified
    --- End diff --
    
    Thanks @spmallette I wanted to do something like this but hadn't spotted 
these tests. While writing some tests I've noticed that when 
RepeatUnrollStrategy is applied from the inside out, adding a barrier step onto 
the end: 
    repeat(a.repeat(b).times(1)).times(1)  -> repeat(a.b.barrier).times(1)
    repeat(a.b.barrier).times(1) -> a.b.barrier.barrier
    
    eg
    ```
    gremlin> g.V().repeat(out().repeat(out()).times(1)).times(1).explain()
    ==>Traversal Explanation
    
===========================================================================================================================================================================================================================
    Original Traversal                 [GraphStep(vertex,[]), 
RepeatStep([VertexStep(OUT,vertex), RepeatStep([VertexStep(OUT,vertex), 
RepeatEndStep],until(loops(1)),emit(false)), 
RepeatEndStep],until(loops(1)),emit(false))]
    
    ConnectiveStrategy           [D]   [GraphStep(vertex,[]), 
RepeatStep([VertexStep(OUT,vertex), RepeatStep([VertexStep(OUT,vertex), 
RepeatEndStep],until(loops(1)),emit(false)), 
RepeatEndStep],until(loops(1)),emit(false))]
    MatchPredicateStrategy       [O]   [GraphStep(vertex,[]), 
RepeatStep([VertexStep(OUT,vertex), RepeatStep([VertexStep(OUT,vertex), 
RepeatEndStep],until(loops(1)),emit(false)), 
RepeatEndStep],until(loops(1)),emit(false))]
    FilterRankingStrategy        [O]   [GraphStep(vertex,[]), 
RepeatStep([VertexStep(OUT,vertex), RepeatStep([VertexStep(OUT,vertex), 
RepeatEndStep],until(loops(1)),emit(false)), 
RepeatEndStep],until(loops(1)),emit(false))]
    InlineFilterStrategy         [O]   [GraphStep(vertex,[]), 
RepeatStep([VertexStep(OUT,vertex), RepeatStep([VertexStep(OUT,vertex), 
RepeatEndStep],until(loops(1)),emit(false)), 
RepeatEndStep],until(loops(1)),emit(false))]
    IncidentToAdjacentStrategy   [O]   [GraphStep(vertex,[]), 
RepeatStep([VertexStep(OUT,vertex), RepeatStep([VertexStep(OUT,vertex), 
RepeatEndStep],until(loops(1)),emit(false)), 
RepeatEndStep],until(loops(1)),emit(false))]
    AdjacentToIncidentStrategy   [O]   [GraphStep(vertex,[]), 
RepeatStep([VertexStep(OUT,vertex), RepeatStep([VertexStep(OUT,vertex), 
RepeatEndStep],until(loops(1)),emit(false)), 
RepeatEndStep],until(loops(1)),emit(false))]
    RepeatUnrollStrategy         [O]   [GraphStep(vertex,[]), 
VertexStep(OUT,vertex), VertexStep(OUT,vertex), NoOpBarrierStep(2500), 
NoOpBarrierStep(2500)]
    CountStrategy                [O]   [GraphStep(vertex,[]), 
VertexStep(OUT,vertex), VertexStep(OUT,vertex), NoOpBarrierStep(2500), 
NoOpBarrierStep(2500)]
    PathRetractionStrategy       [O]   [GraphStep(vertex,[]), 
VertexStep(OUT,vertex), VertexStep(OUT,vertex), NoOpBarrierStep(2500), 
NoOpBarrierStep(2500)]
    LazyBarrierStrategy          [O]   [GraphStep(vertex,[]), 
VertexStep(OUT,vertex), NoOpBarrierStep(2500), VertexStep(OUT,vertex), 
NoOpBarrierStep(2500), NoOpBarrierStep(2500)]
    TinkerGraphCountStrategy     [P]   [GraphStep(vertex,[]), 
VertexStep(OUT,vertex), NoOpBarrierStep(2500), VertexStep(OUT,vertex), 
NoOpBarrierStep(2500), NoOpBarrierStep(2500)]
    TinkerGraphStepStrategy      [P]   [TinkerGraphStep(vertex,[]), 
VertexStep(OUT,vertex), NoOpBarrierStep(2500), VertexStep(OUT,vertex), 
NoOpBarrierStep(2500), NoOpBarrierStep(2500)]
    ProfileStrategy              [F]   [TinkerGraphStep(vertex,[]), 
VertexStep(OUT,vertex), NoOpBarrierStep(2500), VertexStep(OUT,vertex), 
NoOpBarrierStep(2500), NoOpBarrierStep(2500)]
    StandardVerificationStrategy [V]   [TinkerGraphStep(vertex,[]), 
VertexStep(OUT,vertex), NoOpBarrierStep(2500), VertexStep(OUT,vertex), 
NoOpBarrierStep(2500), NoOpBarrierStep(2500)]
    
    Final Traversal                    [TinkerGraphStep(vertex,[]), 
VertexStep(OUT,vertex), NoOpBarrierStep(2500), VertexStep(OUT,vertex), 
NoOpBarrierStep(2500), NoOpBarrierStep(2500)]
    ```
    
    Do you think it's ok to leave multiple barrier steps at the end? I could 
add an extra check that we aren't inserting a barrier after a barrier here - 
but this could interfere with a barrier added by a user?
    
https://github.com/apache/tinkerpop/blob/d3b6a340dbb9c89355c40089a5c083dc0dbd4c5b/gremlin-core/src/main/java/org/apache/tinkerpop/gremlin/process/traversal/strategy/optimization/RepeatUnrollStrategy.java#L72


> Support nested-repeat() structures
> ----------------------------------
>
>                 Key: TINKERPOP-967
>                 URL: https://issues.apache.org/jira/browse/TINKERPOP-967
>             Project: TinkerPop
>          Issue Type: Improvement
>          Components: process
>    Affects Versions: 3.1.0-incubating
>            Reporter: Marko A. Rodriguez
>            Assignee: Marko A. Rodriguez
>            Priority: Major
>
> All the internal plumbing is staged for this to happen, we just haven't gone 
> all the way. In short, a {{NESTED_LOOP}} traverser has an internal 
> {{loopStack}} where {{repeat(repeat())}} will have a {{loopStack}} of two. 
> The {{it.loops()}} checks of the internal repeat will always check the top of 
> the stack and when its done repeating will delete its counter off the top of 
> the stack.
> [~dkuppitz]'s work on {{LoopStep}} will be backwards compatible. In 
> {{RepeatStep}} we will support:
> {code}
> repeat('a',out('knows').repeat('b',out('parent')))
> {code}
> and thus, things like {{loops('a')}} as well as {{times('a',2)}}. Note that 
> naming the loop stack will be a super rare case as most people will just 
> assume standard nested looping semantics with a push/pop stack.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to