- frontend: use e.message instead of String(e) in toErrorResult to avoid
"Error: " prefix for HTTP-mode errors, matching connectionErrorMessage
- frontend: apply same fix to MongoDocBrowser error display
- backend(postgres): use pg_error_to_string across all query execution
paths (execute, execute_batch, copy_out, copy_in) to extract DbError
details instead of generic transport-layer errors
Other databases (MySQL, SQL Server, SQLite, DuckDB) already include
server-level error details in their native Error::to_string().
DuckDB opens database files with exclusive locks on Windows. Previously
connections were only released via Drop, which may not release the file
handle synchronously, causing "file already in use" errors when
reconnecting or testing connections in quick succession.
BIT columns arrive as raw bytes from mysql_async. Without specialized
handling, byte 0x01 was treated as a valid UTF-8 control character and
byte 0x00 was replaced with U+FFFD. Now BIT values are interpreted as
big-endian integers, so BIT(1) correctly displays "0" or "1".
Previously the connection timeout was hardcoded to 5s via
tokio::time::timeout, ignoring any connect_timeout in the URL.
Also mysql_async rejected camelCase connectTimeout as unknown.
Now parse_connect_timeout extracts connect_timeout/connectTimeout from
the URL (case-insensitive, 1-300s range), removes them before passing
to mysql_async, and feeds the value to with_connection_timeout.
Newer StarRocks versions reject the @@socket system variable that
mysql_async queries during connection initialization when prefer_socket
is enabled (default on non-Windows). Since we always use TCP connections,
Unix socket detection is unnecessary.
Split the object browser query into separate TABLES and ROUTINES
fetches. Some MySQL-compatible servers (sharding proxies, restricted
accounts, OceanBase/TiDB variants) reject information_schema.ROUTINES
with ER_UNKNOWN_ERROR (1105 HY000), which previously took down the
entire object browser. Now ROUTINES failures are logged and the user
still gets tables and views.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Previously connect() always overwrote the timezone with the local machine's
timezone, even when the user had explicitly set one via connection parameters
(e.g. ?options=-c timezone=Asia/Shanghai). Now we check the URL for a
timezone setting and skip the automatic SET timezone if one is present.
SET search_path in PostgreSQL is not transactional — it survives rollback
and persists on the connection. Previously, connections returned to the pool
with a modified search_path, causing subsequent queries on the same
connection to run against the wrong schema. Now RESET search_path is always
called after schema-scoped queries and transactions.
- Use client.transaction() for safer transaction handling
- Add prepare_cached for all system queries to enable statement caching
- Use query_raw streaming to avoid buffering large result sets
- Add batch_execute support for bulk DDL scripts
- Add COPY protocol support (copy_in / copy_out) for fast data transfer
- Add pg_quote_ident to prevent SQL injection in SET search_path
- Increase connection pool max_size from 5 to 10
- Fix list_indexes bounds check for expression indexes
Add mysql_async with client_ed25519 support for MySQL and MariaDB connections.\nSwitch MySQL transaction control and schema metadata reads away from sqlx MySQL APIs.\nPreserve connection compatibility by stripping unsupported URL params and handling nullable metadata safely.
Object browser now defaults to showing tables when tables exist.
Table/view context menus match sidebar tree with structure editing,
export, import, diagram, data compare, truncate, empty, duplicate,
and copy name support.
Keep the server session timezone for standard MySQL connections unless the connection URL explicitly provides a timezone override. This avoids TIMESTAMP shifts during SQL import while preserving user-specified timezone parameters.
BEGIN TRANSACTION / COMMIT / ROLLBACK must not go through sp_executesql
(client.execute) because SQL Server raises error 266 when @@TRANCOUNT
changes inside sp_executesql. Route them through simple_query instead.
Add a 512MB maximum response size limit to read_agent_line to prevent
unbounded memory allocation that could cause the process to SIGABRT
when the agent returns an unexpectedly large response.
Ref: #345
Previously, aggregation queries (date_histogram, terms, etc.) with
`size: 0` returned empty results because the empty hits array was
matched but contained no documents, ignoring the aggregations data.
OceanBase MySQL mode uses different collations for information_schema
tables, causing "Illegal mix of collations" when joining COLUMNS with
KEY_COLUMN_USAGE. Explicit COLLATE utf8mb4_general_ci resolves this.
Closes#337
t8y2