[
https://issues.apache.org/jira/browse/CURATOR-499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16737440#comment-16737440
]
Jordan Zimmerman edited comment on CURATOR-499 at 1/8/19 7:22 PM:
------------------------------------------------------------------
Dang - we're in a bit of a catch-22 here. It's been in it's current form long
enough that to support this behavior we'd have to consider it "new" behavior. I
can imagine we'd use some kind of special spelling, maybe
{{forPath("/path/to/node//", data)}} or something.
Or, another "verb" in the DSL: {{.withEmptyPath().forPath("/path/to/node/",
data)}}
was (Author: randgalt):
Dang - we're in a bit of a catch-22 here. It's been in it's current form long
enough that to support this behavior we'd have to consider it "new" behavior. I
can imagine we'd use some kind of special spelling, maybe
{{forPath("/path/to/node//", data)}} or something.
> ZKPaths strips trailing "/" incorrectly when creating sequential nodes
> ----------------------------------------------------------------------
>
> Key: CURATOR-499
> URL: https://issues.apache.org/jira/browse/CURATOR-499
> Project: Apache Curator
> Issue Type: Bug
> Reporter: Marquis Wang
> Assignee: Jordan Zimmerman
> Priority: Minor
>
> I think this was introduced in CURATOR-166, which changed ZKPaths to strip
> trailing slashes from child nodes.
> This is fine in most cases: it made it so
> {code:java}
> curatorFramework.create()
> .forPath("/path/to/node/", data);{code}
> would create a node at {{/path/to/node}}.
> However, if you want to create a sequential node:
> {code:java}
> curatorFramework.create()
> .withMode(CreateMode.PERSISTENT_SEQUENTIAL)
> .forPath("/path/to/node/", data);{code}
> In clients after CURATOR-166 (2.7.1 and higher), this will create a node
> {{/path/to/node00000001}}. Before, and if you call the zookeeper cli with the
> same options, it would create a node {{/path/to/node/00000001}}.
> This effectively makes it so you cannot use curator to create sequential
> nodes where the entire node name is the sequence number.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
