CSharpDB v2.6.0

v2.6.0

Multilingual Collations and Ordered Text Indexes

• Expanded SQL and collection collation support with explicit BINARY, NOCASE, NOCASE_AI, and built-in ICU:<locale> options for comparisons, ordering, unique enforcement, and path indexes.
• Added ordered SQL text index options and planning support, including collation-aware metadata, guardrails, and benchmark coverage for text-heavy index shapes.
• Stabilized ordered-text overflow handling so large duplicate or prefix-heavy text indexes can spill safely instead of overloading a single index page.

Maintenance, Diagnostics, and Admin Hardening

• Hardened reindex rebuilds, WAL open failure cleanup, and advisory statistics freshness after reopen so maintenance and recovery paths behave more predictably after index or storage edge cases.
• Improved the Admin experience for large datasets by reusing table-style pagination and filters in SQL and designer results, making welcome-page row counts on-demand, and clarifying storage-maintenance failures.
• Tightened storage diagnostics and repair coverage around corrupt indexes, live-database inspection, and maintenance workflows.

Reusable Benchmark Data and Release Validation

• Added tests/CSharpDB.DataGen, a JSON-spec-driven data generator for relational, document, and time-series workloads with CSV, JSONL, and direct binary-load output modes.
• Added focused release benchmark mode plus refreshed benchmark docs and baseline tooling so release validation is easier to reproduce on a clean machine.
• Reorganized sample and tutorial project paths, refreshed NuGet search metadata, and updated docs and site content to match the current shipped surface.