Native PostgreSQL schema engineering

*Changes since:* ***v1.2.3*** **Summary:** pgModeler 2.0.0 is a major release and the direct successor to the 1.2.x series. The headline addition is multi-scene support, letting a single model split its canvas across multiple scenes so very large databases stay manageable to browse and edit. This cycle also brings a complete visual redesign of the application, a full translation of the interface into dozens of languages, initial PostgreSQL 18 support including temporal keys, and a substantial round of performance work targeting large models. Alongside the application itself, pgModeler now ships in two editions — Community and Plus — detailed below, with pgModeler Plus gaining a mature Git-based versioning workflow and an improved SSH tunnel plugin. **pgModeler Community and pgModeler Plus:** Starting with this release, pgModeler is distributed in two editions. pgModeler Community remains free and open-source, and keeps the core modeling experience: creating and editing database models, generating SQL/XML, the model fix and metadata handling tools, and everything described in the rest of these notes unless stated otherwise. pgModeler Plus is a paid tier that, in addition to Community, includes the database import, model diff, SQL execution tool, data grid/data handling, and database explorer tools — all of which shipped with every build up to 1.2.3 and are now Plus-exclusive — plus the versioning and SSH tunnel plugins covered further below. **Multi-scene support:** A model's canvas can now be segmented into multiple scenes instead of a single, ever-growing one, either automatically by schema or by a configurable object-count threshold. This keeps the editor responsive on databases with thousands of tables, where a single canvas used to become sluggish to pan, zoom and select on. A new panel lets you switch between scenes and change the segmentation mode, with previous/next navigation buttons for quick browsing. A table that relates to a table in a different scene shows a small marker instead of a relationship line crossing scenes; double-clicking it opens a dialog listing those relationships, with direct navigation to the related table or relationship. An option to also hide relationships between tables of the same scene is available, representing them the same way — useful for models where the drawn relationships add more visual noise than value. Both the segmentation mode and the hide-relationships option can be set up front when importing a database, in the GUI and via new `pgmodeler-cli --import-db` options (`--seg-mode`/`--seg-threshold` and `--hide-rels`); opening an existing model whose objects are concentrated in very few scenes now also offers to switch segmentation mode. The command-line model fix tool (`--fix-model`/`--fix-sdbm`) preserves a model's segmentation data instead of discarding it. **Orthogonal relationship lines:** Relationships can now be drawn with an orthogonal line style, made only of horizontal and vertical segments routed around tables and the crow's foot areas of other relationships. Where lines cross, one hops over the other through a small arc whose shading tells which line is on top. Holding SHIFT and dragging a segment moves it, and the end segments can be shifted through a jog next to the table; these adjustments are saved with the model and in the metadata files (ALT+SHIFT+click discards them). The style (straight, curved or orthogonal) is now chosen in the relationships settings, with orthogonal as the new default, and replaces the former "Use curved lines" option. **Relationship colors and line widths:** Relationships gained a "Line color" submenu in their context menu: with one or more relationships selected (and nothing else), you can assign a random custom color to them, which makes it easier to tell apart the lines of densely connected models, or disable the custom color again. Both actions can be undone. The appearance settings also got two new options, "Objects border width" and "Relationships line width", that adjust the thickness of the object borders and of the relationship lines as a percentage of the default width (from 50% to 200%) with a live preview in the settings page; the built-in themes now use 150% for the borders and 110% for the relationship lines. **PostgreSQL 18 support:** Added initial support for PostgreSQL 18, including the new temporal keys: primary and unique constraints can be defined with `WITHOUT OVERLAPS`, and foreign keys can reference them using `PERIOD`, letting you model time-bound data such as validity periods or scheduling ranges following the SQL standard, instead of relying on ad-hoc check constraints. **Integrated workflow tools:** The database import, model diff and model export tools are no longer separate dialogs — they are now integrated into the main window as dedicated views ("Import view", "Diff view", "Export view"), consistent with the main design view. The model fix and metadata handling tools were likewise merged into a single, integrated tool. **Visual redesign:** The application went through a full visual overhaul. A new theming engine renders widgets contextually across all built-in themes (light, dark, inksaver and a custom-theme option), including tooltips, scrollbars, tab bars, progress bars, menus, checkboxes, radio buttons and more, with theme-adaptive accent colors and smart theme detection. A new default canvas font and a refreshed color scheme make tables, column types (PK, FK, NN, UQ) and constraints easier to tell apart at a glance; editing forms were modernized into cleaner tabbed and vertical layouts with consistent tab-key navigation; and the application's logo, icons and splash screen were redesigned across every platform. **Internationalization:** The user interface can now be fully translated, and this release ships with translations for the languages that best represent pgModeler's current audience, based on usage statistics, including Brazilian Portuguese (pt_BR) — all of them generated with AI assistance, with human review of their accuracy still pending (collaborations are welcome!). More languages will be added as that audience grows. Every executable, not only the main GUI, now follows the configured UI language. **Performance improvements:** Several bottlenecks that mattered most on large, real-world models were addressed. The internal template engine behind every bit of generated SQL and XML code was substantially reworked for speed, on top of no longer reloading and reprocessing those internal template files on every use — together, SQL and XML generation on real-world models is 50-70% faster, which also speeds up opening and saving large models. Also addressed: slow object lookups by name, and a multi-second stall when applying changes through the database editing form on large models. Canvas rendering was also optimized to stay smooth while panning, zooming or dragging, and when working with large or densely connected models. **Stability and reliability:** This cycle closes a substantial number of crashes and memory-related issues, among them: a crash during model validation when a column was restored via undo; a crash/freeze condition in the model close and reload flow; data corruption when closing or reloading models that used certain user-defined types; crashes in the model export and SQL tool; crashes when moving segments of orthogonal lines; an intermittent crash when generating SQL/XML code from more than one thread at the same time (export, validation, database import and model diff all do); and a `--fix-sdbm` bug that was discarding almost every object of a split model saved with ID comments. A batch of memory leaks was also closed across relationship handling, the undo/redo history, copy-and-paste, and model loading, which should help keep memory usage in check during long editing sessions. **Credentials protection:** The passwords of database connections and SSH tunnels are now saved encrypted in `connections.conf` and `sshtunnel.conf`, and these files become readable only by their owner. This keeps the passwords from being read by simply opening the files, e.g., by another user of the same machine or in a copy of the file; the protection is not meant to withstand someone who already controls your account. Existing passwords are converted the next time the settings are saved. Older pgModeler versions can't read the encrypted passwords, so after a downgrade the passwords must be typed again, and the configuration folder copied from the previous version (`pgmodeler-1.2`) still holds the old passwords in plain text. In the pgModeler Plus versioning plugin, the credentials (user and password, or access token) embedded in HTTPS remote URLs are masked in the plugin's log, in its error and warning messages and in the remote URL it shows; note that Git itself still stores them in plain text in the repository's `.git/config`. **pgModeler Plus — Safer database dropping:** Dropping a database in the SQL tool (from the database explorer's toolbar or the database list) now has a second confirmation. After accepting the usual warning, a "Danger zone" dialog asks for the name of the database to be typed, and the drop only proceeds when it matches exactly, protecting against accidental drops. **Compatibility notes:** The "Use curved lines" option no longer exists and the default relationship line style is now orthogonal; pick the style you prefer in the relationships settings after upgrading. The default foreign key column naming pattern changed to a table-first convention; if your models rely on the previous defaults, review your `relationships.conf` after upgrading. Developers building plugins should consult `CHANGELOG.md` for the complete list of code-level changes, since several widgets and classes were renamed or moved as part of the Community/Plus split and the visual redesign. Passwords saved by previous versions are re-saved encrypted; see "Credentials protection" above for what that means when downgrading. **pgModeler Plus — Versioning plugin:** This cycle brings a Git client embedded directly in pgModeler as part of Plus. You can clone remote repositories with live progress, HTTPS credentials or SSH keys; manage remotes and their push URLs from Repository Settings; stage or unstage individual files with filters and a side-by-side diff view before committing (split-model files included); push, pull and sync with a configurable reconcile strategy and upstream tracking; resolve merge/pull conflicts in a dedicated mode that inspects base/local/remote versions of each file; merge any branch into the current one; cherry-pick commits; amend the last commit; reset to any historical commit (soft, mixed or hard); revert a commit without rewriting history; manage tags (create, remove, sync, annotate); and save, browse, apply, pop or drop stash entries. SSH agent keys are picked up automatically, and the Git credential manager can be configured directly from Repository Settings. **pgModeler Plus — SSH tunnel plugin:** The SSH tunnel plugin gained more granular control over your tunnels: automatic startup on launch is now configured per tunnel instead of a single all-or-nothing switch, the toolbar button shows whether any tunnels are currently running with a tooltip summarizing active/stopped counts, and the tunnels menu lists every configured tunnel individually so each can be started or stopped on its own. **Looking ahead:** With 2.0.0 out, the focus now shifts to collecting feedback from the community and addressing whatever issues come up with it, while development of 2.1.0 has not yet begun.