[
https://issues.apache.org/jira/browse/TINKERPOP-2747?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17700030#comment-17700030
]
ASF GitHub Bot commented on TINKERPOP-2747:
-------------------------------------------
disq commented on code in PR #1989:
URL: https://github.com/apache/tinkerpop/pull/1989#discussion_r1135208270
##########
gremlin-go/driver/connection.go:
##########
@@ -42,15 +42,15 @@ type connection struct {
}
type connectionSettings struct {
- authInfo *AuthInfo
- tlsConfig *tls.Config
- keepAliveInterval time.Duration
- writeDeadline time.Duration
- connectionTimeout time.Duration
- enableCompression bool
- readBufferSize int
- writeBufferSize int
- enableUserAgentOnConnect bool
+ authInfo AuthInfoProvider
Review Comment:
> There appears to be one test failing in the checks right now. The maven
logs can be a bit much to search through, I like to search for `❌` in the logs
to find the failing test quickly. It appears that
`gremlin-go/driver/gorillaTransporter.go:68` can sometimes result in a nil
pointer dereference.
Thank you for the clarification on how to run the tests locally.
The issue seems to be one of the tests setting `authInfo` to `nil`, which is
unexpected (and isn't done anywhere else in the codebase as far as I can tell)
so now `gorillaTransporter.getAuthInfo()` has a nil-check and returns
`NoopAuthInfo` if `nil`. `NoopAuthInfo` is an idempotent way of specifying a
zero-value for the `AuthInfoProvider` and can be used publicly as well. (maybe
unnecessary, as simply `nil` would do the same thing)
> Add function callback hooks for gremlin-go authentication
> ---------------------------------------------------------
>
> Key: TINKERPOP-2747
> URL: https://issues.apache.org/jira/browse/TINKERPOP-2747
> Project: TinkerPop
> Issue Type: Improvement
> Components: go
> Reporter: Lyndon Bauto
> Priority: Major
>
> Gremlin Go currently supports custom authentication headers.
>
> For long running connections that reconnect over time, these headers can
> expire.
>
> We should allow injection of a function pointer to generate the header.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
