Populärt: Data uppdaterade 18 november 2026

Kategorier

Internet1 416
Kommunikation328
Ljud och video939
Grafik och design562
Kontor och produktivitet1 180
Utvecklarverktyg3 285
AI-verktyg608
Säkerhet och integritet408
Systemverktyg3 187
Spel544
Utbildning och vetenskap242
Nätverk och IT292

rudbman: äldre versioner att ladda ner

A multi-platform GUI database workbench that talks to any database with a JDBC driver.

Här listas 9 versioner av rudbman, från 26 augusti 2026 till den aktuella 0.3.9. Varje version länkar till utgivarens egna installationsfiler.

2026: 9 9 26
VersionDatumStorlekFiler
0.3.9Senaste 14 november 2026 — 64-bitars EXE (Inno Setup)
Versionsinformation
rudbman 0.3.9 — schemas restored from a script, statement bodies kept whole in the editor and in runs, a run button beside each statement, and a choice of where the date goes in saved SQL file names.
- A schema can be restored from a script. Restore from script… on a schema or catalog takes a backup, pg_dump --inserts, mysqldump, mariadb-dump, or an SSMS or SQL Developer script, plain or gzipped, UTF-8 or UTF-16. It checks the script through once and then runs it on a session of its own opened beside the connection's, with progress, cancel and Continue past failed statements.
- A backup restores into the schema it is opened over. A backup names every object in full, so restoring it from another schema's menu created the objects in the schema it was taken of. The restore now reads which scope a script names and, when that is not the selected one, offers to move it: qualified names, USE and CREATE SCHEMA, and the names passed to nextval, setval, OBJECT_ID, DBCC CHECKIDENT, sp_addextendedproperty and COMPILE_SCHEMA are rewritten into the selected scope, inside routine bodies and MySQL executable comments too, while other strings are left as written. Backups now start with a comment naming the scope they were taken of.
- Scripts run the way each product's client reads them. Run all, and Run selection when the selection holds a client line, read a script as psql, mysql, sqlcmd, SQL*Plus, the Db2 CLP and sqlite3 do: DELIMITER, GO batches, PL/SQL units ended by /, --#SET TERMINATOR, CREATE TRIGGER … END;, MySQL executable comments, and the client commands that are skipped, end the script or are refused before anything runs.
- Statement bodies stay whole in the editor and in runs. The statement shaded under the caret, and the one Run statement sends, was the piece between two semicolons, so a CREATE PROCEDURE … BEGIN …; END, a PL/SQL block or package, a SQLite trigger or a PostgreSQL BEGIN ATOMIC body was cut at its first inner semicolon. Run all cut MySQL routines typed without DELIMITER, BEGIN ATOMIC bodies, SQLite triggers holding a CASE and Db2 routines the same way, and ran an Oracle unit typed without its slash together with everything after it. Both now count a body, and the editor takes its statements from the reader Run all uses, so the shading follows DELIMITER, GO and the SQL*Plus slash and Run statement sends the unit as Run all would.
- A run button beside each statement. The SQL editor's gutter shows a ▶ to the right of the line number on the first line of each statement; pressing it runs that statement without moving the caret.
- Saved SQL file names put the date where you choose. Settings → Database → SQL file name date puts today's date after the tab's name, as before, before it, or leaves it out of the name the save dialog suggests for a query tab that has never been saved.
- A schema backup's suggested file name is dated. The save dialog suggests <schema>_backup_<yyyyMMdd>.sql (.sql.gz with gzip), so a second day's backup does not offer to replace the first's.
- The comment toggle keeps the selection. The same lines stay selected, a backwards selection stays backwards, and the caret stays beside the character it was next to, instead of collapsing to a caret at the old offset.
- Shared libraries are updated to rugpui 0.1.17 and rudbkit 0.1.7, and russh to 0.65 for RUSTSEC-2026-0492 and RUSTSEC-2026-0495. UI dependencies, framework patches and the database toolkit are pinned to their published release commits.

Requirements
- Linux: glibc 2.35 or newer: Ubuntu 22.04, Debian 12 and later, or a distribution built on them, MX Linux 23 among them.

What's Changed
- chore: release 0.3.9 by @xcomart in #63
0.3.8 12 november 2026 — 64-bitars EXE (Inno Setup)
Versionsinformation
rudbman 0.3.8 — schema backups that carry sequences, views, routines and triggers and restore to the same schema on every product, a keyboard shortcut sheet in the app, a statement timeout that query tabs obey, and the reasons a transfer skipped rows.
- A schema backup carries the whole schema. Under Also include, the backup dialog now writes sequences, views, procedures and functions (and Oracle packages) and triggers, all on by default, in the order a replay needs: sequences, tables, routines, views, rows, foreign keys, counters, materialized views, triggers. Tables use the server's own CREATE on H2, Oracle, MySQL, MariaDB and SQLite; elsewhere the reconstruction reads PostgreSQL's and SQL Server's CHECK constraints and computed columns, keeps serial defaults and ALWAYS identities, and writes SQL Server's (max) types and seeds. Routine bodies carry each client's delimiter (DELIMITER, /, GO, --#SET TERMINATOR), identity values and sequence positions carry over, PostgreSQL extensions, enums and domains come first, unicode text is N'' on SQL Server, and DROPs take foreign keys off first. Objects whose DDL cannot be read are left out and listed on the finished card.
- A backup restores to the same schema on every product. Backups replayed through each product's own client and compared catalog by catalog and row by row no longer fail or restore something else. MySQL keeps DEFAULT CURRENT_TIMESTAMP columns in every row, backslashes, multi-bit BIT and YEAR values and INVOKER views, with rows under NO_AUTO_VALUE_ON_ZERO and routines and triggers under their own sql_mode. PostgreSQL keeps expression and partial indexes, bit varying lengths, view options and CHECK OPTION, materialized view indexes, deferrable keys and identity counters. SQL Server keeps uniqueidentifier, money and datetime as the server accepts them, alias types, collations, filtered and INCLUDE indexes, clustering, UNIQUE constraints, default names, indexed views, extended properties and identity and sequence counters. Oracle writes timestamps that do not depend on NLS settings, keeps materialized view indexes, leaves views valid and does not reuse deleted identity keys. Db2 keeps computed columns, identity attributes, CHECK and UNIQUE constraints, expression and INCLUDE indexes in their own schema, VARGRAPHIC lengths, binary values, line breaks in strings, TIMESTAMP(12), overloaded routine drops and counters. H2 keeps domains, ALWAYS identities, arrays and JSON, and SQLite keeps values not of the declared type and the AUTOINCREMENT mark.
- The app lists its keyboard shortcuts. Keyboard shortcuts in the ☰ menu and the macOS menu opens a sheet of the chords in use, grouped by window, tabs and panes, explorer, SQL editor, data grid and ERD, written from the bindings and the menus' own words. It is a dialog like the others: it holds the window's chords back while it is up, Escape puts it away and the keyboard goes back to where it was. User guide, beside it and on the sheet, opens the user guide in the browser.
- The statement timeout applies to query tab statements. Settings → Statement timeout was saved but never sent to the driver. A query tab now runs each statement, and each page of a paged result, under it, read at the start of every run so open tabs pick up a change.
- A transfer shows why it skipped rows. With Skip the row and record why, the finished card now lists the first five reasons and how many more rows were skipped, in all eight languages.
- CSV and template extracts carry no session settings. An extraction wrote MySQL's SET FOREIGN_KEY_CHECKS and SQL Server's SET QUOTED_IDENTIFIER lines into every file. They now go only into a SQL script: one with DDL or INSERT rows.
Requirements
- Linux: glibc 2.35 or newer: Ubuntu 22.04, Debian 12 and later, or a distribution built on them, MX Linux 23 among them.

What's Changed
- chore: release 0.3.8 by @xcomart in #62
0.3.7 8 november 2026 — 64-bitars EXE (Inno Setup)
Versionsinformation
rudbman 0.3.7 — several explorer rows and detail-panel columns picked at once and copied as names or as SELECT and INSERT statements, dated save names that always end in .sql, a selection run one statement at a time, and Shift+Backspace that deletes with the completion list open.
- The explorer takes several rows at once. Ctrl/Cmd-click adds a row to the selection or takes it out, and Shift-click and Shift+Up/Down select a run of rows; a plain click or arrow narrows it back to one. Ctrl/Cmd+C, or Copy names on the right-click menu, puts the names of the picked tables, views, routines, schemas and catalogs on the clipboard, one per line; folders are passed over. A right-click on one of several picked rows keeps the others, so the menu acts on all of them.
- A table's menu copies its SELECT and INSERT statements. Copy SELECT statement (tables and views) and Copy INSERT statement (tables) read the columns of every picked table and copy one statement per table, a blank line between them, laid out in the Format SQL river with one column to a line. The INSERT gives each value a :column placeholder on the same line as its column, and names that need quoting are quoted for the connection's dialect. A toast says when the copy is done, and nothing is copied if a table's columns cannot be read.
- The detail panel's column list picks rows too. Click, Ctrl/Cmd-click and Shift-click pick columns, Ctrl/Cmd+A picks them all, and Ctrl/Cmd+C copies their names; the right-click menu also copies them comma-separated, ready for a select list, or as whole tab-separated rows.
- Saving a query tab suggests a dated name and always writes .sql. The first save offers the tab's name with today's date after it (Query 3_20261108.sql), and a name typed without an extension is saved with .sql after it; when a file of that longer name already exists, the save dialog asks again over it.
- Run selection runs each statement on its own. The selected text is split the way Run all splits the script, each statement keeping its place in the editor so a failure selects it, so CREATE USER; GRANT …; FLUSH PRIVILEGES runs on MySQL and MariaDB without allowMultiQueries and every statement's result is shown. A procedural block or a MySQL executable comment standing as its own statement is still sent whole.
- Shift+Backspace deletes while the completion list is open. With Shift still held from a capital, the first Backspace over an open completion popup did nothing on Windows and Linux; it now deletes like Backspace.
- Shared libraries are updated to rugpui 0.1.16. UI dependencies and framework patches are pinned to its published release commit, bringing the multi-row tree selection and the Shift+Backspace binding in the editor and text inputs.
Requirements
- Linux: glibc 2.35 or newer: Ubuntu 22.04, Debian 12 and later, or a distribution built on them, MX Linux 23 among them.

What's Changed
- chore: release 0.3.7 by @xcomart in #61
0.3.6 5 november 2026 — 64-bitars EXE (Inno Setup)
0.3.5 1 november 2026 — 64-bitars EXE (Inno Setup)
0.3.4 29 oktober 2026 — 64-bitars EXE (Inno Setup)
0.3.3 24 oktober 2026 — 64-bitars EXE (Inno Setup)
0.2.9 10 oktober 2026 — 64-bitars EXE (Inno Setup)
0.2.3 26 augusti 2026 — 64-bitars EXE (Inno Setup)

Populärast i Databaser

Visa alla
  1. MariaDBDatabaser · MariaDB Corporation Ab
  2. HeidiSQLDatabaser · Ansgar Becker
  3. SQLiteDatabaser · SQLite
  4. DoltDatabaser · DoltHub
  5. MongoDBDatabaser · MongoDB Inc
  6. Redis on WindowsDatabaser · MSOpenTech
  7. DB Browser for SQLiteDatabaser · DB Browser for SQLite Team
  8. Beekeeper StudioDatabaser · Beekeeper Studio Team
  9. LiquibaseDatabaser · Liquibase
  10. Another Redis Desktop ManagerDatabaser · Another

Mer från Xcomart

Visa alla
  • rulogmanTerminaler och CLI-verktyg · Xcomart

Taggar

Dela