Sutido 0.10: ClickHouse, Real Abort, and a Security Audit

0.10 is the biggest release in Sutido's history: a fifth engine, an abort button that actually aborts, the results of a full security audit — and, one day later in 0.10.1, a results grid that behaves like a spreadsheet.

ClickHouse Is Engine Number Five

Sutido now speaks native ClickHouse SQL over HTTP — DESCRIBE, PREWHERE, aggregations over millions of rows — with schema browsing from the system tables and sensible default limits so SELECT * on a huge table stays browsable.

Abort That Actually Aborts

We tested this on a table with millions of rows: hit abort on a runaway SELECT * and the server kept scanning — closing the socket wasn't enough. Now cancellation is server-side everywhere: KILL QUERY on ClickHouse, pg_cancel_backend on PostgreSQL, request cancellation on MSSQL, plus real timeouts. When you stop a query, it stops.

The Dual MongoDB Driver

The legacy-server story got its proper ending: Sutido ships two driver generations and negotiates the right one per server. MongoDB 3.6 through current, one connection form.

The Security Audit

We ran a structured audit across the whole app and fixed what it found, including:

• Renderer sandbox enabled with a bundled preload; navigation and window.open blocked everywhere; a narrowed CSP
• SQL identifiers escaped in metadata clients; comment-safe statement classification
• Update feed URLs restricted to https; host-key confirmation dialog when an SSH server's key changes
• Storage hardening: WAL mode, corruption recovery, and graceful handling of undecryptable secrets

Quality of Life

• Server version badges in the sidebar per connection
• Grouped cluster view in the error log
• Streamed collection export/import for large MongoDB collections
• A theme-aware icon set — and a ninth theme, Kawaii

…and 0.10.1: The Grid Grows Up

One day later, the results grid learned to behave like a spreadsheet:

• Select columns (and several at once) by clicking headers; Ctrl+A selects every row — scoped to the grid, not the app
• Ctrl+C copies one column as a value list, several as TSV with a header row — paste straight into Excel
• Ctrl+Shift+C copies rows as JSON; the context menu adds SQL IN lists and Markdown tables
• Sort arrows on every column — stable, type-aware, nulls last
• Filter by a cell's value into a new tab, and compare two rows field by field