[
https://issues.apache.org/jira/browse/TINKERPOP-1844?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16310760#comment-16310760
]
David M. Brown commented on TINKERPOP-1844:
-------------------------------------------
Well Stephen I don't see any easy way around this. Off the top of my head all I
can think of is either returning a list with Gremlin-Python even when a set is
required by the traversal, or implementing custom types. Both have their
disadvantages, though I dislike the idea of custom types most. We do already
have a couple custom types, like {statics.long}, but they are just dummy types
that make sure values are properly serialized. Going any further than that is
going to start to get a little contrived in my opinion. If we disallowed sets,
users who need to do set operations could manually convert the list to a set,
and since they are Python users, they won't expect a int/long or float/double
distinction. Not ideal, but better than a maze of types that don't exist in
Python. Maybe there is a better approach, but I don't know what it
is...suggestions anyone?
> Python GLV test should run for GraphSON 3.0
> -------------------------------------------
>
> Key: TINKERPOP-1844
> URL: https://issues.apache.org/jira/browse/TINKERPOP-1844
> Project: TinkerPop
> Issue Type: Improvement
> Components: python
> Affects Versions: 3.3.0
> Reporter: stephen mallette
> Priority: Critical
>
> The GLV tests are running for python on GraphSON 2.0. When I ran them for
> GraphSON 3.0 I received some strange errors that looked like they might have
> been related to bulking, but I can't say for sure. Anyway, we should
> definitely be running these tests against 3.0 and potentially 2.0 as well
> (2.0 can be a separate ticket and is less of a priority).
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
