[
https://issues.apache.org/jira/browse/TINKERPOP-2261?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16882442#comment-16882442
]
stephen mallette commented on TINKERPOP-2261:
---------------------------------------------
> Longer term, in the spirit of being vendor agnostic, in my opinion, TinkerPop
> should define its own transaction semantics and add a reference
> implementation to test those semantics.
There has been discussion of specifying better transaction semantics in TP4.
Josh wanted it baked into the mm-ADT spec i think. Perhaps TinkerGraph would
implement such semantics then (or not...dunno). I'm curious about where that
might go, but don't feel optimistic.
> Do you think it is worth re-opening this discussion again in the community?
> Or did we have a strong conclusion against this the last time discussion
> happened?
I don't remember the last time it came up. I don't know that anyone is against
the feature itself. It's more a point of taking a simple in-memory graph which
is very easy to reason about and debug and going on to introduce a load of
complexity (e.g. transactions, persistence, etc) that has to be maintained.
That complexity also sorta mars it as a "simple" reference implementation that
folks can look at to learn. But, if you feel strongly about it, feel free to
bring it up on dev - a change of that size is something to raise there.
> Adding transactions to Tinkergraph for better testability
> ---------------------------------------------------------
>
> Key: TINKERPOP-2261
> URL: https://issues.apache.org/jira/browse/TINKERPOP-2261
> Project: TinkerPop
> Issue Type: Improvement
> Components: tinkergraph
> Affects Versions: 3.3.7, 3.4.2
> Reporter: Divij Vaidya
> Priority: Major
>
> Tinkergraph should act as a datastore which supports transactions. This would
> help define tx semantics for Gremlin and add test suites that fail when those
> semantics are violated.
> As an example,
> [https://issues.apache.org/jira/projects/TINKERPOP/issues/TINKERPOP-2257] bug
> would have been caught by the test suite if the tinkergraph was mimicking
> transaction support.
> Hence, this task breaks down into:
> 1. Mimic transaction support in TinkerGraph
> 2. Add tests that validate the transaction semantics (e.g. accessing the
> iterator after the transaction is closed will cause an error)
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
