neo4j-browser
neo4j-browser copied to clipboard
neo4j
Expanding node relationships causes `t is undefined`
Description
When exploring the graph returned from a query, for some nodes expanding their relationships will cause an error that renders the graph exploration mode unusable for the current query window and every other query window. The page has to be refreshed.
The error is Uncaught TypeError: t is undefined:
r /browser/app-1a3e9b736d418eb4284a.js:2871
groupedRelationships /browser/app-1a3e9b736d418eb4284a.js:2871
layoutRelationships /browser/app-1a3e9b736d418eb4284a.js:2871
onTick /browser/app-1a3e9b736d418eb4284a.js:2871
P browser/app-1a3e9b736d418eb4284a.js:2871
tick /browser/app-1a3e9b736d418eb4284a.js:2871
xe /browser/vendor.chunkhash.bundle.js:270
Ne /browser/vendor.chunkhash.bundle.js:270
Reproduce Steps
Expand nodes in graph explorer of the query window
Expected Behavior
No error
Current Behavior
Above.
Can you propose a solution?
No
Your Environment
| software name | version |
|---|---|
| OS | Debian 10 |
| Web browser | Firefox ESR |
| node.js | |
| npm | |
| yarn | |
| Neo4j Browser | Latest |
| Neo4j | Latest |
Thanks for reporting this! Could you help us out by writing out some "CREATE" statements that create a graph that has this problem? I'm struggling to re-create your issue
I can once I deterministically reproduce the crash and understand why the crash is happening.
@OskarDamkjaer The closest i've gotten to reproducing is double clicking a node's expand relationships button quickly
We're working to improve our automatic crash reporting, so hopefully we can find the issue that way. I'll post an update when/if I learn more about this crash
What information is collected by the 3rd party APM services you use? I am currently blocking access to canny.io, intercom.io, sentry.io, and segment.com via NoScript
Sentry.io is our crash reporting tool, it collects stats on what browser was used and what clicks led up to each crash. I've already seen someone else run into what seems like your crash. I'll post an update when I find the time to look into what causes it, I've added a card to our internal bug tracking system