[server] Fix exec_mut rejection when only insert or remove index queries are sent #1898

[michaelvlach]

No description provided.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
agdb	Ready	Preview, Comment	Jun 25, 2026 8:52pm

[Copilot]

Pull request overview

This PR fixes server-side role classification for index mutation queries so exec_mut correctly routes InsertIndex/RemoveIndex batches through the mutable execution path (instead of being treated as read-only and rejected).

Changes:

• Treat QueryType::InsertIndex as requiring DbUserRole::Write.
• Treat QueryType::RemoveIndex as requiring DbUserRole::Write.

Comments suppressed due to low confidence (1)

agdb_server/src/utilities.rs:67

• required_role drives whether exec_mut uses the mutable/cluster path; this bug fix adds InsertIndex/RemoveIndex, but there is still no regression test ensuring index-only query batches are classified as Write. Adding a unit test that builds Queries(vec![QueryType::InsertIndex(..)]) and Queries(vec![QueryType::RemoveIndex(..)]) and asserts required_role(...) == DbUserRole::Write would prevent this from reappearing.

pub(crate) fn required_role(queries: &Queries) -> DbUserRole {
    for q in &queries.0 {
        match q {
            QueryType::InsertAlias(_)
            | QueryType::InsertEdges(_)
            | QueryType::InsertIndex(_)
            | QueryType::InsertNodes(_)
            | QueryType::InsertValues(_)
            | QueryType::Remove(_)
            | QueryType::RemoveAliases(_)
            | QueryType::RemoveIndex(_)
            | QueryType::RemoveValues(_) => {
                return DbUserRole::Write;
            }
            _ => {}
        }
    }

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.