Apply fix_potential_double_encoding() to table/object comment fields
returned by list_tables(), list_table_status_show(), and row_to_object(),
consistent with the existing fix for column comments (c21b5d0).
The sidebar tree displays TABLE_COMMENT directly from these functions,
so when MySQL returns Latin1-decoded bytes for UTF-8 content, the
comment shows as garbled text like 'OA部门åŒæ¥è¡¨' instead of
'OA部门同步表'.
Fix UTF-8 boundary issues in mysql_top_level_limit, parse_usize_token,
and mysql_keyword_at functions that caused crashes when SQL contains
Chinese characters in string literals.
The issue was introduced in commit 021e1682 which added byte-level
indexing without proper UTF-8 boundary checks. When parsing SQL with
Chinese text (e.g., \"WHERE name LIKE '%批量保存%'\"), the byte index
could point to the middle of a multi-byte UTF-8 character, causing
Rust to panic when slicing the string.
Changes:
- parse_usize_token: Use std::str::from_utf8 to validate slice before parsing
- mysql_keyword_at: Add is_char_boundary checks for safe string slicing
- mysql_top_level_limit: Skip non-UTF-8-boundary bytes after incrementing index
This ensures all byte-level indexing operations stay on valid UTF-8 boundaries,
preventing panics while maintaining the original functionality of detecting
LIMIT clauses for bounded result set collection.
Fixes crash reported by users when executing queries with Chinese conditions.
* fix(redis):format INFO output as plain text in cluster mode
* fix(redis): unify INFO category command output in cluster mode
In Redis cluster mode, the INFO command (with category parameters) returns aggregated results from multiple primaries. This fix formats the output as human-readable plain text, consistent with standalone mode, to improve readability.
Closes#1326
Redis write commands (SET/DEL/...) ran from the query editor or the
key-browser terminal did not update the dbN (count) labels in the left
tree until a manual refresh. Add refreshRedisDbKeyCounts(), which re-fetches
authoritative per-db counts via INFO keyspace and patches the labels in place
without rebuilding the tree (preserving expanded key nodes), then trigger it
fire-and-forget after any mutating command in both execution paths.
* fix(grid): normalize smart quotes in JSON input to prevent parse errors
When manually typing JSON values in the data grid, input methods (especially
Chinese IME and macOS smart punctuation) can automatically convert standard
ASCII quotes to smart quotes (U+201C, U+201D, etc.), causing JSON parse failures.
Changes:
- Remove incorrect logic that tried to parse original value before normalization
- Add hasSmartQuotes() function to detect 7 types of smart quotes:
* Chinese quotes: U+201C, U+201D
* German quotes: U+201E, U+201F
* Smart single quotes: U+2018, U+2019
* Fullwidth quote: U+FF02
- Rewrite normalizeSmartQuotes() using charCodeAt to avoid oxc compiler issues
- Add comprehensive tests for MySQL JSON field quote normalization
All existing tests remain passing.
* fix(grid): preserve valid JSON while normalizing smart quotes
The Redis key-name completion cache was only cleared on connection
lifecycle events, so SET/DEL/EXPIRE/... left autocomplete showing
stale keys. Drop the cache for the db each write command ran on:
- add isRedisMutatingCommand() based on the command-table safety field
(confirm = write; destructive blocked cmds minus read-only/admin ones)
- invalidate after each mutating command in queryStore batch execution
and in RedisKeyBrowser.runRedisCommand
* feat: highlight selected SQL matches
* fix(core): detect column extra changes (auto_increment, on_update) in table structure SQL generation
- Add has_column_extra_change check to build_column_sql skip condition
- Add is_column_extra_empty to handle None/Some(false)/empty ColumnExtra
- Fix false positives when extra is parsed but no effective change
* feat(generate): fix date range persistence, add i18n for data generation
* fix:Remove invalid references
* fix(generate): ensure config save/load covers all generator params
- Remove auto-reset watch in GeneratorParamsPanel that cleared params
when loading a profile with different generatorKey
- Extend defaultGeneratorParams() to provide explicit defaults for all
35+ generators (full_name, phone, email, barcode, url, file_path,
foreign_key, image, regex, product_name, etc.) so profiles persist
and restore correctly through JSON serialize/deserialize round-trip
---------
Co-authored-by: t8y2 <1156263951@qq.com>
The result used filter:false + validFor, which CodeMirror treats as
mutually exclusive — validFor is ignored, so the list froze and the
selection did not move while typing. Build the result with the default
built-in filter so typing narrows the list and moves the selection.
- add completion module: command names, two-stage subcommands, and key names
- key-name completion is async with a per-db cache capped at 1000 entries, cleared on disconnect
- gate key completion by argument position: stop after a single-key command is filled, keep suggesting for variadic key-list commands
- reuse the diagnostics command table as the data source, showing group/arity/safety
* feat(mq): implement Apache Pulsar message queue admin
- Add complete MQ admin adapter layer with version detection
* Pulsar adapter with multi-version API support
* OAuth2 token caching and authentication
* JWT token signing for Pulsar clients
* Transport layer (SSH/Proxy) integration
- Implement comprehensive admin operations
* Tenant, namespace, topic management
* Subscription and consumer/producer monitoring
* Rate limits, backlog quotas, retention policies
* Permissions and role-based access control
* Raw admin API for advanced operations
- Add desktop UI components
* MQ admin console with tabbed interface
* Tenant, namespace, topic panels
* Subscription management and cursor operations
* Monitoring panel with stats visualization
* Policy configuration UI
- Add web routes for MQ admin operations
- Add secret store support for MQ auth credentials
- Add comprehensive test coverage
* fix: del local test files
* fix(mq): mark message queue as unsupported mcpMode in driver manifest
MQ admin connections do not use the MCP query bridge - they use a
separate REST admin API. Mark mcpMode as 'unsupported' to align with
the test expectations and reflect the actual architecture.
Fixes CI test: driver-manifest.test.ts
* fix(mq): fix TypeScript errors in MQ components
- Fix typo in SubscriptionsPanel.vue: messagformatError(e)Id -> message.messageId
- Use formatError() helper in RawApiPanel.vue instead of accessing e?.message
- Fix undefined 'tab' reference in ConnectionTree.vue: use config.db_type instead
Fixes frontend CI TypeScript check errors.
* fix(mcp): add 'mq' to MCP server connection type description
The MCP server's connection type description must include all database
types from the driver manifest. Add 'mq' to the list to fix the test:
'add connection database type description includes every manifest database type'
Fixes: packages/mcp-server/tests/driver-manifest.test.ts
* fix(mq): add missing ConnectionConfig fields for Rust compilation
- Add gbase_server and keepalive_interval_secs fields to all ConnectionConfig
test constructors in mq/config.rs, mq/service.rs, and storage.rs
- Import uses_bare_mysql_pool in connection.rs tests module
Fixes compilation errors from upstream ConnectionConfig changes.
* chore(ui): simplify Pulsar icon to logo-only version
Remove text from Pulsar icon (127x30 -> 35x30), keeping only
the wave logo symbol for a cleaner, more consistent appearance
with other database icons.
* fix(mq): return removed pools from drain_connection_pools
---------
Co-authored-by: t8y2 <1156263951@qq.com>
t8y2
leic4u
xcstatus
至高吾尚
bigliangzao
林剑凛
github-actions[bot]
dalew
luoianun
lexmin0412
Spencer.Chang
Spencer.Chang
haipengno1
miracle
dbin0123