Release notes

File Manager

• Two-panel File Manager: A two-panel file manager built into FastEditor, accessible from Tools > File Manager... or via command line with FastEditor.exe --mode=fm.
• Local file operations: Copy, move, rename, and delete (Recycle Bin or permanent) with progress tracking, cancellation, and conflict handling. Incomplete destination files from an interrupted transfer are removed on the next start, sources are left alone, and a notice lists what was cleaned up.
• Search and filters: In-panel quick search by name and mask; search (Alt+F7) by name, size, date, and file contents (supporting PCRE2 regular expressions and large files).
• Editor and Database Builder integration: Press F3 to view local files read-only in the editor, F4 to open them for editing, or press Enter/F4 on SQLite files to open Database Builder.
• FTP and FTPS client: Connection profiles with encrypted passwords (via Windows DPAPI); UTF-8 and CP1251 filename encodings; copy, move, rename, and delete between local folders and servers with progress, cancel, and transfer resume.
• FTPS certificate trust and verification: Interactive certificate trust dialog on TLS verification failure offering public-key pinning (SPKI sha256 with change detection) or per-profile hostname check opt-out (preserving CA validation for shared hosting). Note: remote server files cannot be viewed or edited directly with F3/F4; they must be copied to a local panel first. Direct site-to-site (FXP) transfer is not supported.
• Folder synchronization (Pro): Compare a local folder with a server folder, review differences, then synchronize.
• Included components: libcurl, LibreSSL, and related components (licenses in the licenses folder).

Editor Core & CLI

• Read-only view switch: Added --view launch parameter to open large files read-only without triggering the Pro trial prompt.
• CLI threshold consistency: Headless --diff command now enforces the 50 MB Core threshold, matching --convert, --sql, --join, and --script.
• Documentation: 10-section offline help manual accessible via F1 or Help > FastEditor Help, covering core concepts, formats, search engine, automation, File Manager, Database Builder, and CLI.
• Shortcut adjustments: Resolved shortcut conflicts between Follow Mode (Live Tail) and international AltGr key combinations; Ctrl+0 resets font zoom.

Bug Fixes & Limitations

• Fixed: The editor rejected --mode=… command-line switches, which also broke "Open in Editor" from DB Builder.
• Fixed: Opening a large file that has no line breaks no longer freezes the editor; lines are shown truncated at 1 MB.
• Fixed: DB Builder forms showed an empty attachment/photo field until the record was navigated away and back.
• Fixed: Stored DB Builder form layouts now preserve foreign key dropdown combobox bindings instead of reverting to generic text inputs.
• Known limitation: The File Manager runs with the privileges of the user who started it; operations in protected locations fail, and there is no elevation prompt.
• Known limitation: SQLite stores dates as text; DB Builder displays DATE columns as ISO 8601 text strings.

Fixed

• DB Builder: form and list layouts no longer lose their field bindings when you open and leave the layout designer.
• Printing: the text editor's print output was unreadably small on every previous version; it now prints at the correct size, honours your configured tab width, respects printer margins, and adds a filename and page-number footer.
• High-DPI: text in the layout designer, report designer, relationship graph, label designer, grid and list views now scales correctly on scaled displays.

Report output is now a Pro feature

Printing a report, exporting a report to PDF, and printing a label sheet now require FastEditor Pro. Everything else in DB Builder stays free: tables and schema editing, forms, portals, the grid, search, import, CSV export, the report designer, the label designer, and on-screen report preview. The bundled demo database follows the same boundary.

Added

• Embedded DB Builder: FastEditor now includes DB Builder, an embedded visual SQLite database workspace. Open it via Tools → DB Builder..., the Command Palette (Ctrl+Shift+P → DB Builder...), or by opening a database file directly.
• Schema & Table Designer: Create and modify SQLite tables and fields, configure data types (Text, Integer, Real, Date, Time, and Attachments), define primary keys, and establish foreign key relationships with relational integrity.
• Grid & Form Views: Browse and edit records in a tabular grid or in Form View with type-specific input widgets, field labels, and relational portals displaying records from related tables.
• Visual Script Builder & Step Debugger: Build automated workflows using a visual action tree with branching logic, loops, record navigation, and field calculations. An interactive debugger powered by QuickJS allows stepping through actions, setting breakpoints, and inspecting variables.
• Vector Reports & Multi-Column Label Printing: Design banded reports (header, detail, summary, footer) or multi-column label sheets using Avery presets (Avery 5160, Avery L7163, Avery L7651) or custom dimensions. Preview paginated layouts, print to physical printers, or export to vector PDF documents.
• Attachments & Media Playback: Attach images, PDF documents, and media to records using hybrid storage (inline BLOB or companion disk files). Audio attachments support in-app playback with duration badges, and video opens in the system player.
• SQLite Metadata Detection & Bundled Demo: Opening an existing SQLite database containing DB Builder metadata prompts to switch to DB Builder. A pre-configured sample database (demo.sqlite) is bundled with the installer and accessible from the DB Builder welcome menu.

Changed / Improved

• DB Builder Output Licensing: PDF export, report printing, and label-sheet printing require FastEditor Pro, while schema, forms, scripts, the report and label designers, and on-screen preview remain completely free in FastEditor Core.
• Executable & Installer Size: FastEditor.exe and the setup installer have increased in size by approximately 3 MB due to the inclusion of the DB Builder UI, SQLite database engine, QuickJS scripting runtime, and vector PDF rendering pipeline.

Added

• JavaScript Scripting Host Modal Dialogs: Added app.showMessage(message, title?) and app.confirm(message, title?) methods to the JavaScript scripting environment, allowing user scripts to present informational alerts and interactive confirmation prompts.

Changed / Improved

• Flat Seamless Scrollbar Rendering: Render flat, unbordered scrollbar tracks and thumbs matching the editor background theme, removing legacy 3D beveled borders and thumb grips while preserving decoupled mouse hit zones.
• Find & Replace Panel Controls Layout: Positioned Find and Replace bar buttons, checkboxes, and scope selectors using sequential relative offsets to prevent control overlaps during window and panel resizing.

Fixed

• Startup Window Dark Theme Initialization: Eliminated white client flash during application startup by creating the main window invisibly, assigning a dark theme background brush, and configuring Windows DWM immersive dark mode before revealing the window.
• Startup Window Maximize Transition: Eliminated the visible size transition on startup by maximizing the window while invisible before revealing it on screen.
• High-DPI In-Cell Editor and Coordinate Scaling: Corrected coordinate scaling on displays scaled above 100% for CSV in-cell editor placement, column header resize boundary hitboxes, live status bar resizing, and Markdown preview split-pane divider tracking.
• High-DPI Popup Hit-Testing and Overlay Dismissal: Normalized click hit-testing coordinates on high-DPI displays for status bar popup trigger zones (encoding, line endings) and outside-click dismissal of the command palette.

Changed / Improved

• Integrated Custom Scrollbars: Replaced the default Windows scrollbars and separate 6px overview strip with custom, theme-drawn vertical and horizontal scrollbars. Search matches and diff markers now render directly on the vertical scrollbar track, and matches stay visible through the thumb while dragging. Reclaims 6px of horizontal editor width.
• Harmonized Scrollbar Geometry: Standardized both horizontal and vertical scrollbars to an 8px track with a 10px thumb, high-contrast borders, and proportional grip indicators for consistent visual hierarchy across all dark and light themes.

Fixed

• Display Glitch When Editing in Scrolled Files: Fixed an issue where editing or pressing Enter near the bottom of a scrolled document could leave a line missing from the screen, or show lines at the wrong vertical position, until the view was scrolled or redrawn. File contents were never affected.
• High-DPI Split-View Rendering (250% Scaling): Resolved coordinate scaling and ceiling row calculation defects in split viewports under 250% display scaling, eliminating viewport bottom truncation and misaligned horizontal scrollbar click hit-testing.

Added

• GitHub Flavored Markdown (GFM) Preview: Upgraded the Markdown preview rendering engine to cmark-gfm (0.29.0.gfm.13) with full native support for GFM pipe tables (| col | col |), column alignments (:---, :-:, ---:), strikethrough (~~text~~), autolinks, security tag filtering, and task lists (- [ ], - [x]).
• Find Input Clear ("×") Button: Added a responsive, theme-aware overlay clear button inside the right edge of the Find input field to instantly reset search queries with a single click while preserving keyboard focus.

Changed / Improved

• Markdown Table & Element Styling: Added clean, modern CSS styling for Markdown Preview tables including border collapse, header backgrounds, alternating row zebra striping, strikethrough typography, and tasklist checkbox alignment.
• Theme Contrast & Adaptive Glyph Rendering: The Find clear button dynamically derives its foreground and hover colors from active theme tokens, ensuring high contrast across both dark and light themes (including GitHub Light).
• Pro Feature Informative Dialogs in Core Mode: Pro feature calls (Export Filtered Rows, Join CSV Files, CSV Diff & Review) now show an informative dialog in Core mode instead of silently doing nothing. The dialog explains which feature requires Pro and how to activate it.

Fixed

• Dark Theme Toolbar Rendering & Contrast: Resolved an issue where toolbar buttons in dark themes (Default Dark, Dracula, Monokai) could render with white background tiles. Implemented seamless theme-matched rendering for normal, hover, pressed, and disabled button states across all color themes.
• Enhanced Button Hover & Pressed Highlights: Improved visual feedback in dark themes with clear, distinct background contrast on mouse hover and active toggle states.
• Disabled Action Visibility: Clarified the visual appearance of inactive toolbar buttons (such as Undo/Redo on clean documents) with balanced optical dimming.
• Windows Visual Styles & High-DPI Manifest: Configured release and CI build scripts (release.ps1, p0_ci_gate.ps1) to compile with /XP and /DPIAWARE flags, ensuring the official Microsoft Common-Controls 6.0.0.0 manifest is embedded for modern Visual Styles, crisp High-DPI rendering, and theme-consistent toolbar button states.
• High-DPI Paint Extent on Main Editor Canvas: Switched logical dimension queries to physical drawing surface extents (OutputWidth() / OutputHeight()) across LayoutMainUi and FlushDirtyRows, eliminating unpainted canvas right/bottom margins and correctly centering the welcome screen and click targets under display scaling.
• High-DPI Canvas Rendering in Popups & Overlays: Resolved unpainted canvas margins under High-DPI display scaling across the Command Palette list, Scroll Overview vertical track, SQL Query Popup history list, and Preferences color preview swatches.
• Headless Startup Error Handling: Guarded main window creation failure paths with non-blocking console error reporting and graceful termination instead of unparented modal dialogs hanging headless/CI environments.
• Find Clear Button Visual Persistence: Resolved a Win32 sibling window clipping and Z-order issue where clicking outside the Find field caused the edit control to overpaint the clear button.
• Status Bar Client Placement & Menu Bar Alignment: Restored native Win32 client rectangle measurement in core layout calculations, eliminating a 20px menu bar offset that previously displaced the status bar below the visible window edge.
• Status Bar Temporary Status Lifetime: Fixed status bar temporary message persistence across plugin reloads and asynchronous search actions.
• Filtered CSV Export Header Duplication: Resolved header line duplication and empty row handling in filtered CSV exports.
• Dialog & Popup Window Visual Contrast: Restored native OS dialog and control rendering across all popup dialogs (Preferences, Highlighting Rules, Recipes Manager, Data Profiler, Wizards, and Requesters), eliminating dark theme color propagation and ensuring crisp, high-contrast text readability.

Added

• Natural Language to SQL Generation (NL→SQL via Local LLM): Generate analytical SQL queries from plain English descriptions tailored to the schema and sample rows of the active CSV, Parquet, or NDJSON file. Access it directly from the SQL Query Bar (Ctrl+Q) via the new [✨ AI] button.
• Interactive SQL AI Preview: Inspect generated SQL queries before execution with syntax highlighting, column schema validation, model latency metrics, and one-click actions to run the query immediately (Run Query), insert it into the SQL bar (Insert to SQL Bar), or refine the prompt.
• AI & Local LLM Preferences: Configure OpenAI-compatible local LLM endpoints (LM Studio, Ollama, vLLM, default: http://localhost:1234/v1) under Preferences → AI / Local LLM (Ctrl+,) with model selection, temperature tuning, max tokens, custom system instructions, and a live connection test button (Test Connection).
• DuckDB Scalar Functions Routing: Extended analytical query routing to recognize standard SQL scalar functions (LOWER, UPPER, TRIM, LTRIM, RTRIM, LENGTH, SUBSTR, SUBSTRING, COALESCE, NULLIF, IFNULL, NVL, CAST, CASE/WHEN, REGEXP_MATCHES, CONCAT, REPLACE, ABS, ROUND, STRPTIME, DATE_TRUNC, EXTRACT), automatically evaluating them in DuckDB rather than falling back to literal column name matching.

Changed / Improved

• Platform HTTP Client Enhancements: Added support for custom URL ports (such as http://localhost:1234), robust multi-byte UTF-8 response decoding from local LLM servers, and configurable connection timeouts.
• SQL Literal Boundary Protection: String literals enclosed in quotes (e.g. 'Case Western', 'LOWER(x)') are now properly masked during query parsing to prevent false-positive analytical routing on text values.

Fixed

• Preferences & SQL Bar Control Interaction: Fixed an issue where the Run and Close buttons in the SQL Query Bar became unresponsive when the AI Preferences tab was opened.
• Session Table Resolution for SELECT *: Fixed schema discovery when referencing active tab sessions via SELECT * FROM current.
• AI Popup Lifecycle & Visual Stability: Resolved popup flickering and focus loss during background LLM streaming and window repaints.

Added

• Multi-Click Word and Line Selection: Double-clicking in text mode instantly selects the word under the caret, while triple-clicking selects the entire line. Dragging after a multi-click expands the selection smoothly in word or line increments while preserving CSV grid boundaries.

Changed / Improved

• Search Export Status Feedback: Exporting search results to clipboard now triggers a timed temporary notification in the status bar with automatic clearance.

Fixed

• Archive Search Result Paths: Resolved an issue in Find in Files where ZIP archive matches displayed truncated file names instead of full absolute archive paths, preventing ambiguity across duplicate archive names.
• Archive Preview Notifications: Routed read-only mode alerts and archive extraction limits through non-blocking status bar feedback.

Added

• Go to Line (Ctrl+G): Jump directly to any line number in the active document (Edit → Go to Line... or Ctrl+G) with instant input validation.
• Change Case Transformations: Transform selected text instantly between UPPERCASE, lowercase, and Title Case via Edit → Change Case.
• Trim Trailing Whitespace: Strip trailing spaces and tabs across the document on demand (Edit → Trim Trailing Whitespace) or automatically upon saving via Preferences.
• Remove Duplicate Lines: Clean up logs and text files by stripping duplicate lines while preserving original line order (Edit → Remove Duplicate Lines).
• Document Bookmarks: Toggle visual bookmarks at specific lines to navigate large files quickly (Ctrl+F2 to toggle, Alt+F2 / Alt+Shift+F2 to jump between previous/next bookmarks, and clear all bookmarks from the menu).
• Visual Tab Expansion: Tab characters (\t) are now expanded visually to the configured tab width (TabSize) across all text rendering, caret positioning, mouse clicking, and rectangular/stream selections.
• Dedent with Shift+Tab: Pressing Shift+Tab now dedents the current line or selection by one tab stop instead of inserting a tab character.

Changed / Improved

• Enhanced Indent Guides: Indent guide lines now render with crisp alignment and distinct visibility, including full positioning support when Word Wrap is active.

Fixed

• Preferences Persistence: Resolved an issue where toggling certain settings (such as Trim Trailing Whitespace on Save and Show Indent Guides) failed to persist reliably across restarts.
• Tab Key Focus Handling: Fixed intermittent duplicate tab insertions occurring under specific focus change events.

Added

• Extensibility & Plugin Architecture: Extend FastEditor with lightweight, sandboxed JavaScript plugins. Create custom visual line highlighting patterns, security threat alerts, and custom streaming data exporters (e.g. XML records, custom schemas) that appear seamlessly in GUI export dialogs, the QuickJS scripting API, and CLI pipelines. Browse and install curated plugins directly from the new Plugin Directory (Tools → Plugins → Browse Plugin Catalog...).
• Pipeline Recipes (.recipe.js): Encapsulate multi-step data pipelines (open, filter, SQL aggregate, join, export) into single, human-readable JavaScript recipe files designed for Git version control. Manage, edit, and run project-level, global, or bundled recipes from the interactive Recipes Manager (Tools → Recipes → Recipes Manager...) or execute them headlessly in CI/CD pipelines via FastEditor.exe --recipe <file>.
• Live Tail & Streaming Follow Mode: Monitor active, rapidly growing log files in real time with constant low memory usage. FastEditor smoothly tracks log streams up to millions of lines, automatically pauses scrolling when inspecting historical events, resumes follow mode when scrolling back to the bottom, and continuously reclaims memory in the background (Ctrl+Shift+L).
• Highlighting & Alert Rules Dialog: Configure persistent visual styling rules, gutter badges, and alerting patterns with full PCRE2 regular expression support (Ctrl+Shift+H). When tracking live log streams, matching high-priority events (such as fatal errors or security indicators) trigger customizable sound alerts and native Windows Action Center notifications.
• Multi-Log Timestamp Correlation: Merge and correlate disparate log streams from multiple open tabs or disk files into a single unified chronological view (Log → Correlate Multiple Logs... or --correlate-logs). Multi-line log entries (such as stack traces and JSON dumps) remain cleanly attached to their parent timestamps without interleaving.
• Automated Log Report & Spike Analytics: Generate instant analytical summaries for multi-gigabyte log files in seconds (Log → Generate Log Report... or --log-report). Features message frequency spike detection, timestamp gap analysis, error severity breakdowns, and one-click export to standalone HTML5 dashboards, Markdown tables, or plain text.
• Interactive Data Profiler & Sparklines: Inspect column health and data distributions across massive tabular datasets with instant summary statistics (CSV → Interactive Data Profiler...). Displays null ratios, distinct value counts, min/max values, 10-bin inline Unicode sparkline histograms, and automatic detection of mixed-type schema conflicts.
• Tabular Diff & Keyed Comparison: Compare two CSV or TSV datasets side by side using single or composite matching key columns (CSV → Diff CSV Files... or --diff). Highlights added, removed, and modified records with full column schema diff reporting.
• Smart Cell Inspector Tooltip: Hover over tabular cells to instantly inspect decoded values for Unix epoch timestamps (in local time and UTC), Base64 strings, Hex byte dumps, formatted JSON payloads, and decoded JWT authentication tokens.
• Data Sanitizer on Export: Anonymize and protect sensitive data during export (Ctrl+Alt+E or --export-sanitize). Supports deterministic salted SHA-256 hashing (preserving relationships for downstream JOIN queries), email/card masking, and realistic synthetic data generation across all 7 export formats.
• Split & Combine Datasets: Split massive multi-gigabyte files into manageable chunks by line count, file size, or unique column values (File → Split / Combine Files... or --split / --merge) with full cancellation support and zero disk residue on abort.

Changed / Improved

• Ultra-Fast Batch Replace & Scripted Mutations: Re-engineered the in-memory text replacement engine to eliminate line-index recalculation overhead during repeated edits. Batch text mutations — including Find & Replace All, privacy masking scripts, and tight loops calling document.replaceRange() or csv.setCell() — now run up to 25,000x–50,000x faster on 100+ MB files while preserving single-step Undo/Redo.
• Huge-File In-Place Editing (50+ GB Files): Direct edits and cell modifications in multi-gigabyte mapped files now save in milliseconds with guaranteed crash safety.
• Streaming Regex Replace on Massive Files: Perform length-altering search and replace operations across arbitrarily large files exceeding system RAM via FastEditor.exe --stream-replace at a constant memory footprint under 25 MB.
• Optimized Large File Line Indexing: Enhanced memory pre-allocation when indexing large text and log files, reducing peak memory usage by up to 76 MB on 10 GB datasets.

Fixed

• CSV Join Semantic Modes & Column Naming: Extended the CSV Join wizard and CLI engine to fully support RIGHT and FULL outer joins alongside INNER and LEFT joins, with automatic column collision suffixing (_left / _right) for non-key duplicate columns.
• Clipboard Memory Optimization: Safeguarded large clipboard copy operations to eliminate excessive memory allocation and avoid unexpected prompts during headless automation.
• Automated Headless Test Stability: Cleaned up background test window lifecycles and modal guards for uninterrupted continuous integration execution.

Added

• Native Apache Parquet Support (.parquet): Open and inspect massive columnar Parquet files directly in the CSV Grid with zero setup. Metadata, row groups, and schema data types are extracted in milliseconds, and columns display data types (col [INT64], [VARCHAR]) with virtual scrolling at a constant ~28 MB memory footprint.
• DuckDB Analytical SQL & Multi-Tab JOINs: Run rich analytical queries directly from the SQL Query Bar (Ctrl+Q) using GROUP BY, HAVING, and aggregate calculations (COUNT, SUM, AVG, COUNT DISTINCT) with strict type protection against grouping collisions. Perform SQL JOINs across multiple open tabs (FROM tab1 JOIN tab2 ON ...) without needing a wizard.
• Multi-Format Structured Export (7 Formats): Export active documents, filtered rows, or SQL projections to Parquet, Microsoft Excel (.xlsx), JSON Lines (.jsonl), TSV, Markdown tables, batch SQL INSERT statements, or CSV via Ctrl+Shift+E, the QuickJS scripting API (app.exportProjection), or the CLI.
• JSON Lines / NDJSON Tabular Grid: Open streaming .jsonl and .ndjson files directly into a structured table. Schema keys and nested properties (a.b.c) are automatically inferred and projected into column headers.
• Transparent Archive Viewing (.gz, .zst): Open and inspect compressed Gzip (.gz) and Zstandard (.zst) log files directly without manual extraction to disk.
• Enhanced SQL Bar UX: Interactive SQL syntax highlighting with dark and light theme adaptation, persistent query history (up to 100 queries), named query bookmarks, column/table autocomplete (Ctrl+Space), and precise inline error range highlighting.
• Headless Automation CLI v1.0 & GUI-to-CLI Bridges: Use the FastEditor binary directly in shell scripts and CI/CD pipelines with subcommands for --convert, --sql, --join, --script, --filter, --split, --merge, and --log-report. One-click "Copy as CLI Command" buttons in GUI panels allow seamless transitions from interactive exploration to reproducible pipelines.

Changed

• SQL Query Input Component: Replaced single-line string input with an editor control supporting real-time lexical coloring, smooth horizontal scrolling, and custom selection handling.
• Streaming Export Engine: Export operations now use an asynchronous 64 KB chunked streaming pipeline with real-time progress and cancellation support.

Fixed

• Added a 5-iteration retry loop for atomic file replacements during saves and license updates to eliminate transient file lock collisions with antivirus and search indexers.
• Fixed regex boundary parsing in pii_masking.js to prevent false negatives on isolated credentials and added negative lookahead to eliminate false positives on IPv4 addresses.
• Restored visual error badge overlays on plain text files by routing encoding diagnostics through all exit paths in the syntax styling pipeline.
• Added defensive null-session guards across mouse and keyboard event handlers in the viewport input router.

Added

• JavaScript Scripting Engine (QuickJS-NG): Automate complex multi-step editing, log transformations, and custom data processing with modern ES2023 JavaScript. Includes script hotkey binding, interactive execution (Scripts → Run Script...), and one-click export of recorded macros directly into editable JavaScript files.
• PCRE2-16 Regular Expressions: Full support for Perl-Compatible Regular Expressions in Find, Replace, and Filtering. Use lookahead, lookbehind, non-greedy quantifiers, named capture groups, and Unicode character properties to search intricate patterns across massive datasets.
• Direct Archive Search (.gz, .zip, .zst): Search inside compressed server logs and data archives directly from Find in Files (Ctrl+Shift+F). FastEditor decompresses streams on the fly without extracting archives to disk or cluttering temporary folders.
• Streaming Multi-File Replace & Industrial Results Panel: Safely replace patterns across thousands of files without opening tabs or exhausting system memory. The new dockable Results Panel organizes millions of search matches in a high-performance tree with line context, match counts, and instant double-click navigation.
• 28+ Legacy Encodings, Error Detection & Batch Convert: Open and repair legacy documents across Windows (CP1250–1258, CP866), ISO-8859, Cyrillic (KOI8-R/U), and East Asian (Shift-JIS, GB18030, Big5, EUC-KR) codepages alongside UTF-8 and UTF-16. FastEditor highlights invalid byte sequences and mojibake with a status bar badge, and the Batch Convert wizard (File → Batch Convert...) allows bulk transcoding and newline normalization.

Changed

• Core Mode Capabilities: Fast log error navigation (F8 / Shift+F8 / F9), interactive CSV cell editing (F2), column sorting, and the SQL Query Bar (Ctrl+Q) for slicing and filtering tabular data are now permanently available in the free FastEditor Core edition.

Fixed

• Fixed a loop condition in SQL query evaluation when limiting row projections with no WHERE clause.
• Improved memory stability and per-session resource reclamation during heavy multi-gigabyte find and replace operations.

Changed

• Start/Stop Macro Recording is now Ctrl+Shift+R, and Toggle Word Wrap is now Ctrl+Alt+W. The old Ctrl+Alt+M and Alt+Z did nothing on many machines: GPU overlays and screen recorders claim them as system-wide hotkeys, and Windows hands such a key to its owner before any application can see it. Play Last Macro keeps its own shortcut, Ctrl+Shift+M.

Added

• FastEditor now tells you, once, if another program holds one of its shortcuts as a system-wide hotkey. The notice names the shortcut and the command, so a key that quietly does nothing is no longer a mystery.

Fixed

• The right-click menu on a CSV column header is now in English. It previously showed unreadable characters in place of the column commands.

Added

• JSON files can be read as a table. When the file is an array of objects, Ctrl+Shift+J opens it as a read-only grid, and the status bar offers an Open in Grid badge for files that qualify. Press Ctrl+Shift+J again to go back to the text.
• Columns are the keys of the objects, in the order they first appear in the file. Keys missing from a row show as an empty cell, and nested objects and arrays show as compact JSON so a row still fits on one line.

Fixed

• Pasting several lines of text now scrolls the view to follow the caret, the way typing and Enter already did. The pasted text used to land below the bottom of the window, leaving you to scroll down by hand before you could carry on typing.
• The view no longer snaps back to its previous position a moment after an edit scrolled it. The scrollbar could lag behind the text and then pull the view back to where it had been.

Fixed

• Ctrl+V, Ctrl+C, Ctrl+X and Ctrl+A now act on the input field you are typing in — the SQL query bar, Find and Replace, the command palette and the CSV cell editor. Previously they went to the document behind the field, so pasting a query pasted it into your file and moved the caret out of the field.
• The SQL query bar now uses the editor theme instead of showing a gray strip above the text, and it keeps matching after you switch themes.
• Joining two CSV files that use the same name for their key column no longer produces two identically named columns in the result. Shared names are labelled _left and _right, and the Join wizard now reports the collision instead of reporting none.
• Changing a key column in the Join wizard by clicking a step tab is now respected. The join previously ran with whichever columns were selected the last time you pressed Next, which looked like the two files had been swapped.

Added

• Ctrl+Right and Ctrl+Left move the caret to the start of the next and previous word. Punctuation counts as its own word, so the caret stops at every meaningful boundary.
• Hold Shift — Ctrl+Shift+Right / Ctrl+Shift+Left — to select text word by word.

Fixed

• Using the mouse wheel in a split view no longer crashes the editor. Scrolling now works over both panes, and the wheel scrolls the pane the pointer is over.

Fixed

• The view no longer jumps to a different part of the file when you delete or type text after scrolling. Editing keeps the lines you were looking at in place, and only scrolls when the caret would otherwise leave the screen.
• Scroll position is now tracked in on-screen rows everywhere, so the view lands on the right line with word wrap, filters and CSV views enabled.

Fixed

• Word wrap no longer splits words in the middle of a line. Lines now break at word boundaries, and words too long to fit still wrap as before.
• Home and End, and clicking past the end of a wrapped line, now stay on the visual line you are on.

First public release.