What's New in v1.0.8

Converter

• OpenDocument support – ODT, ODS, and ODP files can now be converted to DOCX, XLSX, PPTX, and PDF, using Office COM → LibreOffice, with native zip/XML fallbacks for the document conversions. OpenDocument extensions are recognized by the file picker, context menu, notifications, and achievements.
• OpenDocument previews – ODT, ODS, and ODP files can be previewed directly in the preview dialog.
• Markdown to HTML – Markdown files can be converted to standalone UTF-8 HTML documents with embedded CSS, from the app or the Windows context menu, by @Othman-Benbrahim in #31
• Raster to SVG – images can now be vectorized to SVG with VTracer.
• Image format detection – HEIC/HEIF/AVIF files are detected from their content even when the extension is wrong, and are transparently decoded through a temporary PNG for PyMuPDF, Pillow, and OCR.
• OCR for image to PDF – a per-conversion OCR option uses the scanlayer engine to create searchable PDFs. It supports none, automatic, or a custom angle in degrees, falls back to a plain image embed when no text is found, and parallelizes OCR when merging multiple images into one PDF.
• OCR in watch folders and scheduled tasks – automation configs accept ocr = true, so image → PDF rules and image merges produce searchable PDFs in the background too, with the same orientation setting and plain-embed fallback.
• OCR output safety – Main and automation now share non-UI OCR helpers. OCR PDFs are staged and validated before replacing existing files; invalid OCR output falls back to a plain image embed, failed conversions preserve the destination, and merged-OCR achievements are recorded only after validation and commit.
• Unified, validated conversion pipeline – PDF ↔ Word, PDF/DOCX merge, DOCX → PDF, and image-to-PDF OCR paths now use shared staged validation and atomic replacement, so failed or cancelled conversions leave the previous destination intact and clean up temporary files.
• Better PDF-to-Word – LibreOffice PDF → DOCX now selects writer_pdf_import; text-only mode reports PDFs with no extractable text instead of silently producing an empty DOCX, and Quick/Watcher/Scheduler use the richer fitz fallback.
• Safer DOCX-to-PDF – DOCX inputs are validated before engines run. Windows tries direct Word COM first, keeps docx2pdf only as a .docx fallback after a COM error, and then falls back to LibreOffice or ReportLab. Existing destinations survive failed conversions.
• Deterministic, safer merges – Main, Quick Convert, Watcher, and Scheduler share merge helpers with staged validation, stable ordering, pypdf PDF bookmarks, shared DOCX page-break behavior, and collision-safe timestamp names.
• Backup before conversion – an optional setting creates a .bak copy of each source file before conversion starts.
• Automatic temp file cleanup – tracked temporary files are removed when the app closes and at the start of each batch conversion.
• Quick Launch – the conversion chooser is now a searchable dialog with keyboard navigation and theme-aware styling.
• Quick Convert safety – Quick Convert uses staged output and cooperative cancellation, preserves existing destinations on failure or cancellation, and cleans temporary staging files.
• Batch conversion hardening – batch routing now reports per-file failures, handles supported PDF/DOCX/PNG targets consistently, and reserves unique output paths.
• TIFF routing – TIFF inputs and outputs now dispatch consistently through the shared image conversion registry.
• Custom image-to-ICO dimensions – Settings now supports custom WxH icon sizes in addition to the preset sizes.

Templates

• Frequent-configuration templates – when enabled in Settings, the app tracks how often each conversion configuration is used. A configuration used 3 times is promoted to an [Auto] template (up to 3 per conversion type). Duplicates of existing templates are skipped, and deleted auto templates are not recreated.
• Auto-save frequent configurations – the tracking that feeds automatic templates can be toggled in Settings.
• Semantic template configuration – new templates store language-independent keys such as text_only, alpha_az, high, and by_pages instead of French display labels. Legacy database rows and exported templates are normalized automatically, so existing templates keep working.

Themes

• QSS-themed settings and dialogs – Settings, templates, dashboard, and dialogs are now fully styled by settings.qss and theme files instead of inline styles.
• Custom themes in Advanced Conversions – the dialog re-applies its styles instantly on theme changes, without restarting.
• Themeable splash screen – each theme can provide an optional splash.qss; themes without one fall back to their dark/light base.
• Themeable Quick Convert window – the context-menu window card uses an optional window_card.qss per theme, including readable combo-box selection colors in every theme.
• Themeable automation buttons – the automation folder and reload buttons in Settings now follow the active theme.
• Theme gallery shortcut – the Theme settings tab now has a translated Get themes link to the community FCP-themes repository.
• Theme-aware dashboard – stat cards, charts, tables, filters, and database controls now derive their palette from the active built-in or custom theme and refresh instantly on theme changes. Custom themes can also define an optional [dashboard] palette in metadata.ini for precise, readable colors.
• Theme-aware achievements – the achievements dialog, cards, XP ring, rank badge, stat bars and unlock/rank popups now use an optional achievements.qss. Built-in light and dark styles are included, and custom themes without the file fall back to the matching built-in style. The stylesheet also accepts an AchievementsPalette block for the custom-painted widgets, including ring/progress colors, card sheen/shadow, fonts, tier/category colors and per-rank colors.
• Built-in theme alignment – dark and light themes now match Obsidian's layout dimensions.
• Theme switching fixes – the logo label updates immediately, and tab close / preview button emojis are visible with custom themes.
• Custom theme names in Settings – the active custom theme now displays the name from metadata.ini instead of its folder name.
• Readable achievement category highlights – category hover and selection backgrounds are subtle tints, and category text colors are checked for contrast across all bundled themes.

i18n

• Turkish localization – full Turkish UI translation, installer strings, and legal pages. Issue #32 fixed.
• Achievements translated to English – achievement names and descriptions are now available in English.
• Semantic translation keys – UI text now uses stable semantic keys instead of French literals, making translations consistent across the app.
• Legacy language-file compatibility – legacy_aliases.json maps old French keys to semantic keys, so existing community .lang files and saved values continue to resolve.
• Bundled language coverage – new and migrated strings are present in English, French, Italian, Russian, Turkish, Chinese, English-revisited, and the blank language template.
• Hardcoded UI cleanup – remaining hardcoded French fallbacks were removed, duplicate labels were consolidated, and file-dialog filters, encryption warnings, and template summaries are now localized.
• New UI strings – all new OCR, OpenDocument, template, and theme strings are translated across the existing languages.
• OCR tooltip re-translation – the OCR tooltip now updates when switching languages without restarting.

Licensing

• GPLv3 only – the separate commercial license was removed and the README and terms of use now reference GPLv3 exclusively.

Build & Packaging

• src/ layout – application code, assets, styles, and languages moved under src/, with all build and installer files consolidated under packaging/.
• Stable local paths – configuration, encryption keys, and SQLite databases are anchored to the source directory so they stay alongside the app.
• QSS recognized as CSS – added .gitattributes entries for *.qss files.
• New Themes available – added two new custom themes by default in the bundle also available at FCP-Themes
• New dependencies – scanlayer (OCR), VTracer (SVG), and Markdown.
• Dependency cleanup – removed the unused ffmpeg-python package.
• Asset organization – application fonts, SFX, and icons are grouped under src/Assets/ for packaging.
• Build automation – the main build auto-detects UPX from PATH with fallback locations, and build_all.ps1 creates the portable ZIP plus SHA-256 checksums in Output/.
• Version 1.0.8 – application, packaging, and metadata version bumped for the release.

Testing

• 1,066 tests passed, 8 skipped – expanded from 193 in v1.0.7, with new suites for PDF → Word fallbacks and cancellation, DOCX → PDF engine ordering and validation, PDF/DOCX merge safety, OCR consolidation and achievement accounting, OpenDocument conversions, ODF previews, image format detection, image-to-SVG, OCR, parallel OCR merge workers, automation OCR, Quick Launch, custom OCR orientation, scheduler ordering, asset layout, Terms & Privacy enforcement, console suppression, auto-save, backup, temporary-file cleanup, dialog sizing, theme fixes, dashboard theming, achievements theming, semantic template configuration, and translation-key consistency.
• Real integration coverage – real Word COM and Tesseract tests run when available; the 8 skipped tests require unavailable LibreOffice/Word integration tools in the current environment.
• Shared fixtures – the QApp session fixture, offscreen platform, and isolated app fixture are consolidated in conftest.py.

Fixes

• Config and key files are now stored in the source directory instead of the working directory.
• SQLite database files are anchored to the source directory.
• Refreshed tzdata for Morocco UTC+0 and removed the stale example job.
• Fixed ModuleNotFoundError when running achievements_manager directly.
• Removed the unimplemented "complete text" PDF-to-Word mode.
• Conversions from the Windows context menu now require Terms & Privacy acceptance, matching the main app.
• The background daemon now refuses to start watch folders and scheduled tasks until Terms & Privacy are accepted.
• No more console window flash at launch: playsound3's import-time backend probe (ffplay -version) now runs hidden, and tesseract/ffplay share one console-suppression helper.
• Fixed a crash when opening Settings caused by the console-suppression patch breaking asyncio imports.
• Per-file preview buttons are restored after a batch rename.
• Template tests now use an isolated temporary database, so test runs no longer create Test Template rows in the real app database.
• Template detail rows use normal bullets instead of leftover ?? placeholders.
• History search now matches the localized Type column shown in the app, not only the internal operation key.
• History conversion types now come from the active language file for every supported language, with readable format-arrow fallbacks for untranslated advanced operations.
• The Word to PDF options dialog now opens at its layout size hint, avoiding the Windows geometry warning.
• Merged PDF/DOCX timestamp collisions now append a unique suffix instead of overwriting an earlier output.
• Scheduled tasks now process input files in deterministic case-insensitive order instead of relying on filesystem enumeration order.
• PowerPoint COM conversions use isolated temporary workspaces so concurrent conversions cannot remove each other's files.
• Invalid or missing DOCX inputs fail with a clear message before opening Word.
• Failed or cancelled PDF/Word, merge, and OCR conversions preserve existing destinations and remove temporary staging files.
• PDF merges preserve bookmarks/outlines through pypdf.
• LibreOffice PDF → DOCX conversions use the Writer PDF import filter.
• Text-only PDF → Word reports image-only/scanned PDFs instead of returning an empty DOCX as success.
• Merged-image OCR achievements count only pages from a successfully validated and committed PDF.
• Dialog buttons now follow the active theme's QPushButton rules, including custom themes.
• Cleaned up the Advanced Conversions title and close button, removed emoji prefixes from Word-to-PDF labels and achievement names, and kept Markdown HTML output light-themed with embedded CSS.

Full Changelog: v1.0.7...v1.0.8