Script 'mail_helper' called by obssrc
Hello community,

here is the log from the commit of package openshot-qt for openSUSE:Factory 
checked in at 2026-04-14 17:49:01
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Comparing /work/SRC/openSUSE:Factory/openshot-qt (Old)
 and      /work/SRC/openSUSE:Factory/.openshot-qt.new.21863 (New)
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Package is "openshot-qt"

Tue Apr 14 17:49:01 2026 rev:9 rq:1346411 version:3.5.1

Changes:
--------
--- /work/SRC/openSUSE:Factory/openshot-qt/openshot-qt.changes  2025-01-27 
20:57:18.135698733 +0100
+++ /work/SRC/openSUSE:Factory/.openshot-qt.new.21863/openshot-qt.changes       
2026-04-14 17:49:33.038592921 +0200
@@ -1,0 +2,616 @@
+Wed Apr 08 06:16:33 UTC 2026 - Paolo Stivanin <[email protected]>
+
+- Update to version 3.5.1:
+  * Fixing some Codacy nitpicks in our unit tests mostly.
+  * Translating "User Interface Scale" in UI and documentation
+  * Adding User Interface Scale to preferences, to allow users to easily 
change the scale of OpenShot. When Enabling the Legacy Timeline it disables 
this option, and sets it back to 100% scale (our legacy web timeline backends 
have ugly artifacts when using some fractional scales)
+  * Fixing SHIFT+Scroll Wheel regression, where our new qwidget timeline 
backend did not scroll horizontally when SHIFT was pressed. Added unit tests as 
well.
+  * Align razor icon with mouse position hotspot in QCursor (qwidget timeline 
backend). It was offset incorrectly, so the dotted line was not aligned with 
the actual cursor position, thus cutting in the wrong place.
+  * Invalidate all qwidget timeline clip thumbnail caching when "Edit Title" 
is used, so thumbnails of titles stay in sync after edits (dragging thumbnail, 
initial timeline thumbnail, and dropped thumbnail)
+  * Updating file menu image used in docs, due to missing "Optimize" option in 
the menu. Updating README.md to add missing English PDF for website.
+  * Updating supporters.json file with credits from donations.
+  * Fixing a regression which added Preview->Optimize to the main windows top 
menu - oops.
+  * Fixing many Codacy nitpicks and complaints. Mostly minor changes - I 
validated all the code still works and tests pass.
+  * Adding final 3.5.1 translation compiled files (qm files)
+  * Adding missing translatable strings for Optimize feature
+  * Bumping version to 3.5.1, minimum libopenshot version to 0.7.0
+  * Fixing sphynx syntax error on korean user-guide translation
+  * Fixing broken language headers in user guide po files
+  * Updating user-guide translations in all po website po files.
+  * Improving test_translations.py to also check user-guide translation 
syntax, for missing placeholders, accidental unescaped % signs, and gettext 
correctness. Updating README.md with better instructions.
+  * Updating user-guide language README instructions, for calling `make 
gettext` and merging those template changes into all *.po files. Also, updating 
all po files with new pot tepmlate changes.
+  * Updating user-guide docs translation templates
+  * Updating core OpenShot translation templates for new release
+  * Fixing regression on auto-play preview dialog (was slow to open, resize 
event was fighting auto-play signal), now it's faster to start playback, and 
still resizes correctly when window changes sizes.
+  * Updating a few more user-guide images to have white backgrounds instead of 
black (for alpha)
+  * Updating a few more user-guide images to have white backgrounds instead of 
black (for alpha)
+  * - Updating all preferences images in user-guide to use white background, 
and updating preview tab image with new optimized video settings. - Refactoring 
asset moving code, to be sure we remove the assets as we move them (titles, 
optimized videos, protobuf files, thumbnails, etc...) - New Unit tests
+  * Fixing some Optimize issues found during testing: - Optimize on an 
unlinked file, just links it again - Delete & Unlink now works again (broke due 
to file path issues) - Fixed a backup_assets/ folder path issue, when a 
backup.osp is used (it just just resolve to the .openshot_qt/ folder. - Added a 
main window, Edit->Clear->Optimized Videos option (if optimized files exist)
+  * Lots of improvements to optimize feature, renaming menu, updating docs, 
improving file name matching for "Link to Existing" to more closely match 
behavior of other common suffixes and established/expected patterns.
+  * Fixing a regression on the preview dialog - so SVG images stay sharp and 
auto playback still works
+  * Adjusting preview dialog, split file, and region dialogs to use 
proxy_reader (optimize preview) when available, for performance.
+  * Fixing build failure related to executor (and possibly older verison of 
python) - our Linux builder was failing.
+  * Integrating new, faster thumbnailing, with a 3X thumbnail size max decode 
(so our Thumbnail function can still scale it down with higher accuracy - 
clearer picture)
+  * Add to Timeline dialog now selects clips after adding them to the timeline 
- similar to dragging files directly onto the timeline (identical behavior now)
+  * Improving zooming on timeline (CTRL+Scroll Wheel, and CTRL+Press Scroll 
Wheel) to leave playhead at existing position, and center zoom around that 
position. Also fixing a small snapping regression.
+  * Fixing thumbnail cache issues on qwidget timeline backend, causing some 
thumbnails to flicker when smooth zooming
+  * - Add new Proxy Service to manage a new "Optimize Preview" feature - 
Optimize Preview intercepts JSON to libopenshot, and substitutes lower 
resolution preview media, and manages this process (generating new 
low-resolution files, linking pre-existing low-resolution files, removing low 
resolution proxy files. - Moving thumbnails into folders per file id - New unit 
tests - New settings to manage "Optimize" process (in Preview tab on 
preferences)
+  * Applying new thread preference clamps (2 to 3X cpu count), logging on 
thread init, and a more descriptive "Copy Version Info" context menu - with way 
more system info (and cache settings).
+  * Adding link to our step by step guide on installing ComfyUI: Advanced AI 
server for OpenShot.
+  * - Upgraded Change Video Style... to WAN VACE with reference image + 
depth/Canny control. - Added Extract -> Depth and Extract -> Lines for images 
and videos. - Added reference image to generate dialog: Reference tab. - 
Renamed img2video-svd to img2video-wan and updated docs/icons.
+  * Fixing transitions / clip paint artifacts, during updates (which sometimes 
caused a tiny glitch in the UI when resizing them)
+  * Updating docs with multi-selection trimming / re-timing notes.
+  * Preview the correct clip/transition when multiple clips are selected and 
being trimmed/re-timed. Which ever item initiated the trim, will be the one 
previewed.
+  * Improving transition resizing and retiming, to always show the static 
keyframes for brightness, and ensure they are still visible after resizing. 
Also, some cleanup work around this same feature.
+  * Fixing multi-item trims/retiming to use the same exact position for all 
items, to prevent drift as updates are applied and mouse is still moving.
+  * Fixing first and last thumbnail slots (on Entire mode) to use center of 
slot as the frame #, instead of anchoring to the start/end frames - which was 
causing all sorts of issues. This is simpler, and consistent with how the other 
slots report their center frames.
+  * Only adding overlay play icon to video file types (in Project Files), and 
leaving images and audio just thumbnails - so each has a distinct look, and we 
no longer use the film strip edges for videos.
+  * Adding retries to Project Files thumbnailing, better support for device 
pixel scaling and ui-scale support, and support for a new thumbnailing override 
(scale_mode), so our thumbnails no longer have black bars based on aspect ratio 
(they zoomin enough to remove black bars) making thumbnails appear clearer.
+  * Improving the cursor waiting / spinner behavior to be more reliable, and 
less likely to spin forever.
+  * Improving thumbnailing with some experimental overlay changes, and dynamic 
icons in Project Files
+  * - Added Ctrl + middle-drag smooth zoom to the qwidget timeline (smooth 
scroll) - Multi-selection now only resizes/retimes on shared outer edges; 
interior edges drag instead. - Grouped trim/retime now applies correctly to 
aligned selected clips. - Preview during grouped resize/retime follows the 
topmost clip. - Fixed grouped resize mouse-release flashing. - Reduced zoom 
thumbnail churn, while keeping retimed clip thumbnails stable. - Added focused 
helper tests for zoom, grouped edge targeting, trim, and retime.
+  * Sentry bug fix: Protect save_recovery() - now returns immediately if 
file_path does not exist yet
+  * Sentry bug fix: Protecting AddTransition from None item_ids list (similar 
bug to AddClip was fixed)
+  * Sentry bug fix: Fixing advanced title editor launch error handling, to not 
over verbose report this to Sentry
+  * Sentry bug fix: Missing blender exe error logging was too verbose, and 
moving from log.error to log.info
+  * Sentry bug fix: Protect int formatting on properties tableview, to avoid a 
formatting error: %d format: a number is required, not str.
+  * Sentry bug fix: protecting item_ids attribute, so we don't get a 
'NoneType' object has no attribute 'clear' exception.
+  * - UpdateManager sets a simple in-progress flag while undo/redo is running 
- any re-entrant undo() / redo() call returns immediately instead of touching 
history - history entries are removed by stable index with pop(), not by object 
lookup with remove()
+  * Fixing sentry bug: Removing old reference to actionPause on Tutorial (no 
longer exists), the actionPlay just toggles the icon now.
+  * Fixing Sentry bug: TimelineView.dragLeaveEvent() now delegates to 
TimelineWidget.dragLeaveEvent() on the qwidget backend, instead of   running 
the legacy cleanup path that could leave item_ids in the wrong state.
+  * Fixing multi-selection drags of transitions to not apply any 
auto-direction updates (just move them). Also, apply clip moves first, and then 
apply transition moves, to keep things in batches and process them in a better 
order. This fixes a bug where by moving a bunch of clips + transitions, it 
could reverse the transitions due to invalid ordering.
+  * Fixing a bug on launch, where scale is compared for equality, but it's a 
float, so 14.99999999 != 15.0, and it causes the initial empty project to be 
"dirty" and show a save project dialog. Using tolerance check now.
+  * Additional logging to identify what updates mark the project as dirty, and 
log using debug level. Trying to identify why Windows thinks project has 
changed on launch.
+  * Fixing Qt compatibility helpers for Ubuntu 18.04 (our oldest supported 
launchpad PPA build), which was missing 2 functions from later versions of Qt, 
and crash when painting a clip on the timeline.
+  * Do not mark project dirty, if default sample rate changes on launch (on 
some environments this is more common)
+  * Fixing keyframe detection logic to better detect keyframes (even empty 
ones with no points), for example when you delete all keyframes for a clip 
using the keyframe panel. Now, all keyframes available still appear when right 
clicking on the keyframe panel. Also, some log adjustments have been made 
around deleting keyframes.
+  * Animate>Reset menu now clears rotation, shear_x/y, origin_x/y, as well as 
location_x/y, and scale_x/y.
+  * Updating translations for all core translation files (website languages)
+  * Updating translation template
+  * - Larger FPS fraction limits in File Properties, Export, and Edit Profile. 
- Live calculated FPS (2 decimals) labels added to all three dialogs, so you 
can understand easily that 30000 / 1001 = 29.97. - File Properties now shows 
live Duration above Frame Rate. - FPS editing is image-sequence-only
+  * Updating gitlab artifact expiration from 4 weeks to 2 weeks, running out 
of space.
+  * Fixing recent projects unit test to be deterministic, and not fail at 
certain times of the day. Take 2.
+  * Fixing recent projects unit test to be deterministic, and not fail at 
certain times of the day (oops)
+  * Fixing sphinx configuration to be backwards compatible for older Ubuntu 
versions: config-inited / builder-inited
+  * Fixing sphinx configuration to be backwards compatible for older Ubuntu 
versions: add_css_file and add_js_file methods.
+  * More work on making unit tests more compatible with build servers, which 
are headless (no display). Take 4.
+  * Ensuring webengine is not required for python unit tests (take 3)
+  * I added a test bootstrap in src/tests/__init__.py that runs before 
unittest discover imports the Qt-heavy test modules. It now:
+  * - Added a shared test helper at src/qt_test_app.py so tests can retrofit 
an existing Qt app instance with the OpenShot-like methods/state they need. - 
Changed src/language/test_translations.py to create its QCoreApplication lazily 
instead of at import time, which removes the discovery-order poison pill.
+  * Updating ai.po translations for user-guide doc translations
+  * Fixing some typos on ai.rst (ComfyUI docs), primarily where to find the 
setting for ComfyUI URL.
+  * Fixing missing razor icon on qwidget timeline backend, and also enhancing 
the razor to also cut transitions.
+  * Bumping version to 3.5.0-dev
+  * Fixing an issue in our deploy script, if a GitHub release asset already 
exists.
+  * Packaging libsndfile in AppImage, to prevent some audio issues on systems 
with incompatible system libsndfile libraries
+  * Fixing small regression failure in our new unit tests (on Linux gitlab 
runner)
+  * Making qwidget keyframe dragging silky smooth, and fixing issue with 
disappearing partial thumbnails during trimming.
+  * Fixing a bunch of qwidget clip and thumbnailing issues: - trimming of 
re-timed clips - thumbnailing system big refactor, use time keyframes, and 
volume keyframes - lots of new unit tests around thumbnailing and clip sizes - 
new squash/stretch thumbnail behavior when re-timing - new trim thumbnail 
behavior on start, and start-end
+  * Translating docs in all website languages
+  * Bumping version to 3.5.0, minimum libopenshot version to 0.6.0. Major 
upstream release of OpenShot!
+  * Updating donors and supporters
+  * Fixing more Codacy nitpicks
+  * Fixing Codacy nitpicks about /tmp/ file paths being hardcoded.
+  * Fix gitlab-ci to only sign windows builds when all builds succeed
+  * Fixing test failures on github and gitlab ci's
+  * Added new unittest coverage for json_data, project_data, main_window, and 
timeline helpers. Actions and GitLab Linux CI to run full test discovery 
command.
+  * Removing legacy "resource" attribute on transitions, which break on 
project save for AppImages (due to /tmp/.mount/ folder locations changing 
between runs)
+  * - Recent Projects now normalizes paths, removing Windows slash-mismatch 
duplicates. - Missing-file paths now display cleanly, wrap better, and Browse 
picks a folder instead of a file.
+  * I patched src/classes/sentry.py so sentry.init_tracing() fails closed: if 
Sentry transport setup throws, OpenShot logs a warning and continues without 
telemetry.
+  * Stop playback when trimming starts, and reset play button correctly
+  * Improving dragging multiple selections on qwidget timeline backend, to 
ensure they are moved as an atomic unit, with no drift during fast dragging and 
dropping.
+  * Fixing transition snapping direction to correctly handle overlapping clips
+  * When dropdowns are used to keyframe a value, use constant interpolation, 
since these values always represent binary choices (like types, enums, modes), 
and never ranges (linear, bezier).
+  * Add fade_audio_hint to transitions created automatically, which informs 
the Timeline to fade audio as well as the image. If the transition is on the 
left side of a clip, it will fade in audio. If the transition is on the right 
side of a clip, it will fade out audio. And if the transition is overlapping 2 
clips, it will fade out one, and fade in the other, using an equal-power curve.
+  * Improved ComfyUI async request checking on launch to avoid lagging when 
checking for server availability in the UI.
+  * Fixing snap regression which caused snapping on the qwidget timeline to 
stop working after zooming into the timeline (CTRL+Scroll Wheel)
+  * Adding missing webp dependencies for windows freeze.py, to move them into 
the root so windows can find them when loading the webp imageformat.
+  * Renaming Comfy UI to ComfyUI in translations as well
+  * Renaming Comfy UI to ComfyUI.
+  * Updating all translations for the next release of OpenShot, including the 
full set of website languages 100% translated.
+  * Updating some Qt packaging changes for imageformat dependencies, to move 
them into the root folder so the imageformat plugins do not fail to load at 
runtime.
+  * Updating language template POT file with new English strings from Openshot
+  * Fixing a few tiny media path differences between JS and qwidget timeline 
backends (attempting to fix a JPEG issue on Windows)
+  * Improving the Check Hardware decoding button to use a new libopenshot 
method for verifying hardware decoding actually worked
+  * Fixing a bug in the autosave system, if not modifications happened since 
last time
+  * Removed Experimental label from Blender GPU acceleration
+  * Removing "Transform" menu from OpenShot, since it's not the default 
(always-on) behavior when selecting a clip (or multiple clips). Also updating 
the documentation.
+  * Adding live previewing to transition trimming, for both static and 
animated masks, to give more feedback to what portion of the mask will be used 
after trimming.
+  * Incorporating Raffi's suggestions into documentation. Updating a few 
annotated screenshots with arrows, moving some sections around, and other small 
refactors.
+  * - Mask handling now uses has_single_image to split static wipes from 
animated masks. - Animated masks keep real start/end and remove wipe brightness 
keyframes. - start/end now show for effects that use a video mask source. - 
Split-file mask sources now apply their saved start/end.
+  * Huge update of documentation / user guide, updating every single 
screenshot, improving tables / descriptions, and links. All screenshots now use 
Cosmic Dusk theme, and also improved ai.rst a bunch, adding new images and new 
content.
+  * Large refactor of preferences and categories, to simplify and clarify some 
settings. Also fixing some cosmic dusk theme issues related to preferences, and 
making the comfyui URL check button use a green checkmark, instead of a dialog 
popup. Lastly, defaulting to new timeline (qwidget - no longer called 
experimental), and adding a legacy web timeline option - if a user needs to use 
the old webengine or webkit timeline. Also, defaulting thumbnail style to 
"Entire" as the default value.  Lastly, removing "Auto Transform" from 
preferences, and making that enabled perminetely (since OpenShot is very buggy 
feeling with it disabled).
+  * Small bug fix for placeholder on profile edit dialog, where formatting was 
being escaped
+  * Clear export dialog settings cache when the project profile is changed, to 
prevent confusion when switching profiles - we don't want to use old, cached 
export of our old profile.
+  * Adjusting mp4/webm and many common export presets to use crf quality 
settings, and setting options for software codecs to use improved values for 
"gop", "allow_b_frames" and "max_b_frames". The end result is smaller file 
sizes across the board, improved quality for med/low quality settings, and 
similar export times. Some are faster, some are slower by a few %, but file 
sizes are reduced across the board.
+  * Fixing duration calculation on new libopenshot handling of mp3 files (with 
has_single_image = true)
+  * Support high DPI thumbnails on qwidget experimental timeline
+  * Adding new docs *.rst file for timeline usage, including keyframe panel.
+  * Adding modifiers to the keyframe panel "Insert Keyframe" buttons:  ALT = 
bezier, CTRL = constant, SHIFT = linear.  Also fixed a few bugs with keyframe 
right click menu inside the keyframe panel, for single and multiple selections.
+  * Fixing regression in SHIFT click on qwidget timeline backend, to correctly 
select a range of clips/transitions.
+  * Fix regression which clamped playhead to max frame on timeline.
+  * Fix small regression to the split file / preview / cutting dialog, so it 
restores normal functionality
+  * Adding QToolTip support for qwidget experimental timeline backend, for 
clips, effects, transitions, and track icons.
+  * Small fix to our GetMaxFrame calculation, to clamp seeking (next/prev 
buttons) to the start of the last frame, not after the last frame.  This 
prevents a strange bug which can allow playback to go past the last frame.
+  * Removing noisy log thumbnailing missing on qwidget timeline backend
+  * Fixing selection when single clicking on multiple selections, to unselect 
the other items.
+  * Display file reader "file_name" attribute, which can differ from the 
actual original file name. For example, emoji's have their emoji name (for 
readability), and Split File dialog can have user-defined file names, and file 
names with timestamps (if no name is given). These custom names now appear in 
the properties dock dropdown for clarity.
+  * Fixing missing loop icon on Windows builds (hopefully)
+  * Fixing bug in "Add to Timeline" for JS timeline backend, probably a 
regression from all the qwidget changes)
+  * Fixing qwidget experimental timeline to drag all selected clips, even when 
they are outside of the viewable area. Also, improving our mp3/audio file 
detection, so thumbnails are faster loading when thumbnail style set to 
"entire" mode.
+  * Improving transform in video preview widget to handle multiple clips 
better, even if they don't all intersect the playhead - still create keyframes 
at the same relative points in all selected clips.
+  * Improving timeline auto extend duration when using "Add to Timeline" 
dialog or dragging large or multiple files past the right edge of the 
project/timeline. Also, dragging a clip past the right edge now smoothly 
extends the project automatically.
+  * Removing noisy logs in video transforms
+  * - Snapping is now 12px. - Trim snapping is directional and stable 
(left/right both work). - Out-of-range keyframes (including frame 1 / 0.0) now 
snap correctly. - Preview drag updates are smoother (less seek throttling).
+  * Fixing regressions on JS timeline backend seeking/playhead movement, to 
correctly seek and clear cached frames.
+  * Adding a new trace UI recorder, which can assist in automated UI testing 
and validation.
+  * Upgrading text to video to wan2.2 family, 5B model
+  * Fixing some minor regressions on ComfyUI models that mention a prefix path 
(removing the prefix), and preventing an invalid SRT file path from being set 
as the file's caption - if Whisper fails.
+  * Fixing bug which caused the property selection dropdown to not show clip's 
effects, when a clip was selected in the Experiemntal qwidget timeline.
+  * Auto generate waveforms for audio-only files when adding them to the 
timeline. Also, on the qwidget experimental timeline, correctly display the 
audio thumbnail when in thumbnail mode.
+  * Updating documentation with ComfyUI / Advanced AI, Animated effect mask 
support, and more doc updates.
+  * Fixing CTRL+Click, SHIFT+Click, ALT+Click handling on experimental qwidget 
timeline backend.
+  * Removing older legacy pipeline code and support for ComfyUI, since we 
split them into JSON files. Also removing usage of turbo model.
+  * Fixing regression in Seek Next/Prev buttons due to caching fixes and 
stronger seek protections (for smooth caching and playhead movement)
+  * Fixed tracked-object selection consistency so Object Detection property 
edits/transform handles target the selected object (supports both legacy index 
keys and new <effect-uuid>-<index>   IDs), preventing misaligned or delayed box 
updates.   Also fixed region preview seek-signal compatibility by allowing 
QueueSeek to accept both int and int,bool, and kept ESC handling simple/local 
in SelectRegion (no global focus/event hacks).
+  * Fixing a couple seek / refresh frame regressions, which were not 
previewing our property changes for clips/effects
+  * - Timeline/seek call sites were updated to use SeekSignal(frame, 
start_preroll) consistently, so commit seeks explicitly pass True. - Playhead 
dragging logic was simplified and made deterministic: emits preview/commit 
seeks directly, avoids duplicate emits, and forces a final commit on release. - 
Preview thread seek handling was switched to a thread-safe latest-wins queue, 
so rapid scrubs don’t backlog stale seeks. - A small timeline mouse-move path 
fix ensures playhead drag events are handled immediately. - Temporary debug 
noise in this repo was cleaned up (notably playback-mode and cache-thread 
toggle logs).
+  * Fixing qwidget timeline backend to allow for clips to drop beyond the 
duration of the project (i.e. past the right edge of the timeline) and 
extending the timeline automatically.
+  * Fixing "Add to Timeline" dialog bug which did not correctly extend the 
timeline when files extend past the duration of the timeline.
+  * Improving missing file dialog to be more clear, adding instructions, and 
making long paths use relative paths, with a bold file name.
+  * Modifying the launch.py script to append lib/PyQt5 path to the environment 
variable on Windows. This is gated for windows only, to fix an issue with our 
packaging of imageformats not loading due to the dependencies not being found 
on the PATH. This impacted JPEG and other image formats.
+  * Fixing mixed slashes in windows paths for Transition extra JPG files, due 
to the /extra/ folder. These paths need to be normalized. This only was 
breaking in the Experimental timeline though, but this fixes it.
+  * Unselect a project file when dropping it on the timeline, and focus the 
timeline widget. This allows the user to drop a file on the timeline and press 
DELETE immediately, and it only deletes the new clip (not the previously 
selected Project File too)
+  * Populating project files on "Open Project" is now batched, and populates 
the model much faster (30%+ on some larger projects).
+  * Fixing duplicate file importing bug, where a user can import the same file 
after opening a project - but then it prevents further importing until the next 
time you open the project. Weird.
+  * Improve missing file prompt, to have a "Skip File", "Skip All", and 
"Browse..." button. Also, check some additional spots in the project file where 
file paths are used, and better duplicate prevention when a file is used on 
multiple things (transition effect path and a clip path... for example)
+  * Patched a likely root cause for this Windows-only crash: synchronous 
QWidget.repaint() calls during active paint/update cycles.   - Replaced 
repaint() with update() in playback/zoom-heavy paths:
+  * Fixing playhead snapping inaccuracy on the Experimental timeline, which 
snapped +1 frame after the playhead instead of on top of the playhead frame.
+  * Fixing Retro theme and removing theme specific code from our region dialog.
+  * Improvements to theming for AI context menus, generate dialog, and region 
dialog (for tracking)
+  * Make meta batches for tracking objects with SAM2 use dynamic batch size 
(so we minimize chance for OOM on higher resolution videos)
+  * Fixing SAM2 tracking on images to mirror the functionality on videos:   - 
Track -> Blur   - Track -> Highlight   - Track -> Mask
+  * Fixing selection handling after a generated file was added and auto 
selected
+  * - Added Music... ACE 1.5 template + music icon. - Renamed Audio... to 
Sound.... - Fixed music runs: random seeds, no stale output reuse, and 
prompt-to-tags/lyrics handling.
+  * - Prevented cross-job state contamination with a per-run cache key for 
windowed SAM2 state. - Improved chunk-boundary stability using richer carries 
(point+bbox) plus 4-frame boundary replay. - Added detailed DINO/SAM2 debug 
logging (currently always on).
+  * Added a loop icon on video preview dialog, and SPACE now controls 
play/pause, and the play button resets correctly when end of video is reached.
+  * - Fixed Track Object output routing: Blur/Highlight now only import final 
video, Mask only imports mask; blocked fallback paths that were re-adding extra 
outputs. - Fixed filename prefix handling so openshot_mask/openshot_gen 
suffixes are preserved (with payload appended), enabling reliable output 
filtering. - Expanded Generate/Highlight UX: simplified Prompt + Tracking tabs, 
added Qt color pickers, plus new Mask Brightness and Background Brightness 
controls wired end-to-end.
+  * - Region UI now supports multi-tool annotations (pos/neg points + rects), 
frame markers, and clear-all, plus new toolbar icons. - Generate + SAM2 
workflows now pass full tracking JSON (seed frame + per-frame 
points/rects/auto) instead of seed-only inputs. - Tracker preprocess flow was 
fixed (dialog parenting + guards/imports) to avoid region selection 
crashes/regressions.
+  * - Comfy scene splitting now returns range metadata and OpenShot creates 
split-style file entries from it, reusing the original source path with 
start/end trims (no duplicate segment MP4     imports). - Scene split file 
names now match Split File dialog format exactly (base (start to end)), with 
only start and end attributes applied.
+  * Replaces the WAN-based image workflow with a standard SDXL graph 
(CheckpointLoaderSimple, CLIPTextEncode, KSampler, VAEDecode,     SaveImage). - 
The image model/settings are now quality-oriented SDXL Base: ckpt_name = 
sd_xl_base_1.0.safetensors, steps = 28, and cfg = 6.5.
+  * - Replaced built-in txt2img-basic and txt2video-svd workflows with 
WAN-based graphs (UNET/CLIP/VAE loaders, ModelSamplingSD3, WAN model paths), 
and set video defaults to 832x480, 16 fps,     length: 64 (~4s). - Switched 
these template IDs (and img2video-svd) off legacy Python workflow generation so 
OpenShot now uses JSON templates directly at runtime. - Added WAN node types to 
template classification, renamed the image-to-video menu label to Image to 
Video (WAN 2.2 TI2V), and added/used a new img2video-svd WAN 2.2 TI2V template 
in the     Enhance flow.
+  * - Fix Comfy progress tracking for meta-batch follow-up prompts and cleaner 
WS event selection. - Add queue progress metadata (progress_detail, 
sub_progress) and improve fallback behavior when /progress is unavailable. - 
Switch blur-anything SAM2 workflow to OpenShot windowed/chunked nodes with 
masked blur and 96-frame batching.
+  * - Harden Comfy progress updates after sleep/wake: detect stale WS, 
reconnect with backoff (60s -> 300s), and fallback to /progress when WS is 
quiet. - Fix template input rewriting for multi-loader workflows by matching 
both file/video (and audio/file) keys. - Tune Blur Anything defaults: switch to 
sam2.1_hiera_tiny, keep model loaded, reduce batch size to 32.
+  * Check more known paths used in effects and clip effects, and properly call 
find_missing_files() flow - to prompt the user, or remove the effect if asset 
can not be found. Also it logs when assets/effects are removed.
+  * Fix recursive paint on Experimental timeline by adding repaint_after_paint
+  * - Added new SAM2 Blur Anything templates:       - 
src/comfyui/image-blur-anything-sam2.json       - 
src/comfyui/video-blur-anything-sam2.json - Added Comfy availability gating + 
URL validation:       - Hide AI menus when Comfy is unreachable       - 
Preferences now has Check for comfy-ui-url - Added Comfy error 
truncation/sanitizing so failures don’t dump huge payloads in UI dialogs. - 
Enabled proper maximize/min-max behavior for Region and Split dialogs. - Moved 
Queued badge to the bottom of file thumbnails (list + tree views).
+  * Large refactor to support user-defined ComfyUI JSON files in 
.openshot_qt/comfyui/ folder, and write a debug.json file (what we send to 
ComfyUI for debugging purposes). Also split out all our built-in workflows as 
simple JSON files, with a few overrides to help menus/categories. Also now use 
the category icon when generating new items (audio icon, image icon, etc...). 
Menu is now dynamic, and will load user-defined JSON workflows as well.
+  * Add OS-dropped file on Drag-Enter on the Experimental timeline, so the 
clip is draggable like a normal clip.
+  * Auto show properties when a clip or transition is double-clicked (all 
timeline backends supported)
+  * Integrating recent fix from comfy-ui branch, to resolve accidentally 
unselecting Project Files (multiple selections) when right clicking.
+  * Fixing automatic selections when adding files from OS (file explorer), Add 
to Timeline, and Dropping emojis and files on timeline (Experimental qwdiget 
timeline backend)
+  * Fixing file multi-selections in this branch (so right click doesn't not 
clear them)
+  * Fixing qwidget timeline clip selections, so that multiple clips can be 
right clicked on without unselecting most of them.  However, the clip menu 
button does unselect other clips. Also, wrapping a transaction around "Remove 
Clip" menu so a single UNDO will bring all removed clips back.
+  * Fixed some issues with file tree/list selection, and clicking off the 
selects (margin or empty spot) now deselects items correctly. Also, fixed 
preview window to correctly close on ESC/reject async (fast shutdown, and stops 
playback)
+  * Large refactor of ComfyUI context menus, trying to simplify and improve 
integration with many of the AI tasks. Also, new icons.
+  * Lowering quality and fps of text to video (for now)
+  * Add TransNet scene segmentation template and robust segment import handling
+  *  - Add Comfy templates for img2video-svd and RIFE 2x frame interpolation, 
with full workflow wiring.  - Use remote Comfy node/model discovery (RIFE + 
flexible SVD checkpoint selection).  - Show immediate generation progress in 
Project Files (including queued state) without mutating file names.
+  * Adding support for Whisper/SRT subtitles to the Generate dialog. Also, 
auto Caption effect when a file has caption data already.
+  *  - Normalize transition drag payload paths and restore robust transition 
reader loading on drop (deferred preview still uses lightweight path-only 
reader metadata).   - Harden deferred transition drop finalization to 
detect/log failed inserts and only refresh timeline state after successful 
commits.
+  * Autoplay when preview window is opened. Seems like we are just saving 
users a click, lol.
+  * Fix auto transition when adding new emoji or file to experimental timeline
+  * Fixing region dialog used by Tracker effect to select a region (this was 
broken in a recent commit due to a new signal that was not implemented)
+  * qwidget timeline: defer drag inserts until drop, keep real preview
+  * Do not delete new items (clips and transitions) when mouse goes over track 
name (annoying), for example when trying to drop a new item at position 0.0 
(next to the track name).  Also, select items (and unselect others) when the 
menu icon is clicked (similar to our JS timeline backend)
+  * Adding some additional protection and logging around transition readers 
and drag data handling for transitions (trying to fix a windows specific bug 
where transitions fail to drag)
+  * Fix playback state checks and refresh behavior
+  * Improve split-file UX and fix export correctness for split clips
+  * Refactoring auto transitions so they work the same in our JS and qwidget 
experimental timeline backends, by overlapping 2 clips.
+  * Add Comfy audio/video generation pipelines and generalized output import
+  * Refactoring Comfy UI service code to it's own file, improving code to work 
with remote servers (i.e. LoadFile -> Upload files). Updating file headers with 
correct header and copyright. Updating some base/default models for image 
generation and upscaling. Much more stable now at reconnecting, losing 
connection, and not giving up.
+  * Initial ComfyUI pipeline and client implementations. - Generate with no 
input file - Generate with 1 input image - Progress displayed in "Project 
Files" - Cancel Job menu - Full circle generations and editing proven
+  * timeline(qwidget): polish transition visuals and live resize feedback
+  * Fix transition auto-direction and keyframe handling on drag/trim
+  * Timeline(qwidget backend): fix keyframe panel sync/selection across drag, 
snap, tracks, and multi-clip edits
+  * • Improve QWidget timeline theme parity and keyframe panel styling across 
Humanity Dark/Retro
+  * - Adjusting lock colors for JS timeline backends (dimmer version of normal 
track style and border style). - Moving thumbnails downward on Experimental 
timeline (Humanity Dark and Retro themes) - Adding lock functionality to 
Experimental Timeline (undraggable, undeletable)
+  * - Fix many dragging / snapping related issues on the Experimental timeline 
keyframe panel. - Remove / Reorder track buttons on Experimental timeline 
(Lock, Keyframe Panel) - Improve selection box on keyframe panel, so the tiny 
gaps between the property rows do not unselect the clip.
+  * Adding the name of the emoji as the Project Files name, which is much 
easier to read (instead of the real file name of the emoji SVG file)
+  * Fixing a regression which broke the preview and split file screen from 
appearing. Not sure when this broke, but this feels like a merge regression.
+  * Removing debug code causing feature branches to upload artifacts to GitHub 
daily builds
+  * Enable "Use Blender GPU rendering for Animated Titles" setting by default 
(seems very stable now)
+  * Temporary test the develop branch only upload and mktorrent logic for our 
new code sign job
+  * Hiding a gitlab ci "re-init-builders" job, and gating it behind a release 
branch filter.
+  * Renaming windows gitlab ci job names, and splitting the code signing job 
by architecture (x64 vs x86)
+  * Fixing python executable name on new code signing commands in gitlab ci
+  * Adding msys bin to path for code signing gitlab ci job
+  * Fixing yml syntax error on gitlab ci section for code signing
+  * Updating Windows code signing flow, to separate the gitlab runner jobs, so 
another server can sign and upload.
+  * Updating developer certification for Mac code
+  * Fixing many Codacy nitpicks, and protecting against a missing timing mode 
"reader" attribute in a clip (which happened to break while testing these 
changes)
+  * Adding transaction around adding multiple clips to the timeline, so a 
single CTRL+Z will undo them all.
+  * Adding transaction to emoji adding (adds a file and a clip), so a single 
CTRL+Z should undo this
+  * Removing noisy debug logs for property model updates
+  * Removing incorrect log during project save, which claimed to copy title 
assets when they were already copied.
+  * Fixing project title update during an export (i.e. auto save firing during 
an export caused an error)
+  * Improving thumbnailing on qwidget timeline backend (more stable thumbnail 
slots, less flicker, better handling of tiny clip widths). This applies 
improvements to all thumbnail styles (start, start/end, entire clip).
+  * Removing delay on thumbnail trimming refresh, so preview is updated 
immediately after clip is trimmed.
+  * General improvements to caching and playback: - New caching defaults (0.5 
play ahead cache, 768 MB limit, 900 frames max) - Seek playhead to trim 
position after trimming is done, and refresh preview only one time (more stable 
preview after trimming)
+  * Fixing crash on deleting clips with effects, and fixing regression that 
causes effects to break the UNDO/REDO system - not allowing the user to undo an 
effect added to a clip.
+  * Adding CTRL and SHIFT modifiers to snap rotation in the video preview 
transform handles. We snap by 15 degree increments only when pressing the 
modifier key.
+  * Prevent a crash on shutdown when using the qwidget timeline backend (by 
ensuring the thumbnail thread is properly shut down first)
+  * Sort missing files before prompting the user to find the missing files.
+  * Improving missing file prompts to be less annoying, and adding a 'cancel' 
button which stops prompting the user, and no longer prompts per "clip", only 
per "file".  Less prompts, easy bail out.
+  * Fixing title editor to correctly edit titles or duplicate titles from our 
primary Project Files context menu (this is a regression from our tabstop 
branch)
+  * Silence more theme misses (when applying theme colors, icons, and styles)
+  * Reduce grid size of Emoji list view (they have too much horizontal 
padding/margin)
+  * Quite theme miss logs (too noisy at the moment)
+  * Calling apt-get update on GitHub translations action, since it is failing 
with a 404 trying to install dependencies for qttranslations5-l10n python3-pyqt5
+  * Fixing a few focus-related colors on Cosmic Dusk, so checkboxes, browse 
buttons, and tabs (horizontal and vertical) look more distinctive, for easier 
TAB/keyboard navigation.
+  * Fixing a bunch of Codacy nitpicks on out new tabstop code
+  * Fixing a new regression in our test_translations.py script, due to 
changing our "Line %s" substitutions with QLocale
+  * Making ENTER on Preferences window close the dialog - not press the 
"Restore Defaults" button
+  * Improvements to the top menu styling for Cosmic Dusk, adding a focus style 
(to ALT allows you to see which menu is focused using keyboard and arrow keys). 
 Also improving the top menu's PINK debug focus border 
(OPENSHOT_DEBUG_FOCUS=1). Also fixing many Property related keyboard editing 
issues (SPACE/ENTER now being editing the value OR show the dropdown menu). 
Pressing a ,.0-9 numeric key enters edit mode, and types that key value.
+  * Improvements to the top menu styling for Cosmic Dusk, adding a focus style 
(to ALT allows you to see which menu is focused using keyboard and arrow keys). 
 Also improving the top menu's PINK debug focus border (OPENSHOT_DEBUG_FOCUS=1)
+  * Updating translations for our 25 fully translated website languages 
(adding missing emojis and missing translations)
+  * Improving numeric translation in OpenShot, for patterns like "Track " + 
number, "Transition " + number to use QLocale().toString() instead, so we can 
natively support different number systems.
+  * Standardizing the template text in our SVG titles (Replacing UPPER version 
with Title casing), and adding them to translation template (so "Title", 
"Sub-Title" and similar text values will be included in the translation system.
+  * Fixing a regression on set_single_column mode for screen readers (like 
Orca). Now our Thumbnail and Detail views work great with screen readers (and 
no longer duplicate the icon text next to the Thumbnail) since both views have 
their own Proxy now.
+  * Improving focus styles for cosmic dusk on main window, to work well with 
hover and click styles, and toggle buttons (like snapping, razor, timing)
+  * Adding focus styles into main window toolbars, so it's more clear what has 
focus when pressing TAB.
+  * Improvements to Export dialog TAB order, and pink debug outline for 
Advanced Tabs added.
+  * Improvements to Export dialog TAB order, and pink debug outline for 
Advanced Tabs added.
+  * Fixing some pink debug TAB styles, consolidating some treeview padding 
styles in base.py theme, fixing checkboxes and a few other pink outline issues.
+  * Improving TAB order on main window - better logic to ensure order goes 
from TOP LEFT to BOTTOM RIGHT and adjusts dynamically to changes in dock 
position.
+  * Fixing some Codacy nitpicks, code complexity, some unused imports, etc...
+  * Fixing some missing focus debug pink borders (for spinners and sliders) 
for OPENSHOT_DEBUG_FOCUS=1, and fixing some TAB order issues in Animated Titles 
screen.
+  * Fix creeping timeline height on launch - timeline dock should restore to 
the same height as the previous size when OpenShot is closed
+  * Improvements to list views (transitions, emojis, files, effects) to output 
the correct name for screen readers (in thumbnail views), and the dock icons 
now have a name.
+  * Allow for pressing buttons with SPACE key when buttons are focused, so 
timeline doesn't steel control of the space bar. This allows for dock buttons 
to be TAB'ed to and SPACE closes a dock.  Also, Allow for any # numeric value 
to edit a property value in our property dock (0-9,.)
+  * Improvements to tabstop on main window, and added an env variable: 
OPENSHOT_DEBUG_FOCUS=1 to clearly show a pink bright focus border for debuggin.
+  * Stop blender from rendering when cancel is pressed. Also, prevent Animated 
Title screen from jumping focus back to listview after thumbnail preview is 
generated
+  * Adding Title and Animated title to our tabstop improvements - both screens 
now work great with TAB key.
+  * Adding initial tabstop implementation, to improve the tab order (i.e. 
using TAB key to jump through the various screens). This commit improves the 
Preferences, Export, Title, and Animated Title screens.
+  * Removing auto update sphinx docs - due to translations and git pulls on 
production website
+  * Completing more translations for other domains (transitions, blender, 
emojis) for our 25 website languages
+  * Completing 100% translations on the 25 languages translated on 
openshot.org, to match our website fully
+  * Adding GA4 tracking to sphinx updates on website (gitlab-ci)
+  * Adding GA4 tracking instructions for user-guide documentation for 
openshot.org website
+  * Adding instructions for generating all user-guide translations for the 
openshot website, and adding a toggle to make the html docs use relative parent 
paths for (_static, _images, _sources) for easier serving of the user-guide 
translations.
+  * Fixed latex user-guide pdf output for most languages, including bn and ar 
- still RTL issues though, and a few languages still have issues: "ar" "hi" 
"ja" "ko"
+  * Adding fully translated languages for user-guide: - ro - ru - tr - uk - vi 
- zh_hans - zh_hant - pt
+  * Adding fully translated "nl", "nb", "pl" languages for user-guide
+  * Adding fully translated "it", "ja", "ko" languages for user-guide
+  * Adding fully translated "hr", "id", "is" languages for user-guide
+  * Adding fully translated "hi", "fi" languages for user-guide
+  * Adding fully translated "de", "es", and "fa" languages for user-guide
+  * Adding translations for ar and bn lang codes
+  * Adding initial po files for all website supported languages - to our 
user-guide will initially match up with these 25 languages
+  * Updating README.md instructions
+  * Reverting invalid icon names in French translation
+  * Removing doc folder from make html command in README.md
+  * Initial POT english templates for user-guide for v3.4 of OpenShot, and 
initial French translation - for testing.
+  * Experimental ssl and crypto libraries included in AppImage freeze to 
prevent ssl errors on newer distros
+  * Replace old universal analytics tracking code with updated GA4 tracking - 
for basic anonymous usage tracking and exception reporting
+  * Added a safe __deepcopy__ override to JsonDataStore 
(src/classes/json_data.py) that skips deep-copying compiled regex objects 
(which Python 3.6 can’t clone)
+  * Importing typing for Tuple for python 3.6 compatibility
++++ 319 more lines (skipped)
++++ between /work/SRC/openSUSE:Factory/openshot-qt/openshot-qt.changes
++++ and /work/SRC/openSUSE:Factory/.openshot-qt.new.21863/openshot-qt.changes

Old:
----
  openshot-qt-3.3.0.obscpio

New:
----
  openshot-qt-3.5.1.tar.xz

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Other differences:
------------------
++++++ openshot-qt.spec ++++++
--- /var/tmp/diff_new_pack.yJPyno/_old  2026-04-14 17:49:35.666701551 +0200
+++ /var/tmp/diff_new_pack.yJPyno/_new  2026-04-14 17:49:35.666701551 +0200
@@ -1,7 +1,7 @@
 #
 # spec file for package openshot-qt
 #
-# Copyright (c) 2025 SUSE LLC
+# Copyright (c) 2026 SUSE LLC and contributors
 #
 # All modifications and additions to the file contributed by third parties
 # remain the property of their copyright owners, unless otherwise agreed
@@ -18,7 +18,7 @@
 
 %define appname org.openshot.OpenShot
 Name:           openshot-qt
-Version:        3.3.0
+Version:        3.5.1
 Release:        0
 Summary:        Non-linear video editor with broad format support
 License:        GPL-3.0-or-later
@@ -76,9 +76,6 @@
 %suse_update_desktop_file -r %{appname} AudioVideo AudioVideoEditing
 %fdupes -s %{buildroot}%{python3_sitelib}
 
-%check
-python3 -B ./src/tests/query_tests.py -platform minimal
-
 %files
 %doc README.md
 %license COPYING

++++++ _service ++++++
--- /var/tmp/diff_new_pack.yJPyno/_old  2026-04-14 17:49:35.706703205 +0200
+++ /var/tmp/diff_new_pack.yJPyno/_new  2026-04-14 17:49:35.710703370 +0200
@@ -2,17 +2,11 @@
   <service name="obs_scm" mode="manual">
     <param name="url">https://github.com/OpenShot/openshot-qt.git</param>
     <param name="scm">git</param>
-    <param name="revision">v3.3.0</param>
+    <param name="revision">v3.5.1</param>
     <param name="versionformat">@PARENT_TAG@</param>
     <param name="versionrewrite-pattern">v(.*)</param>
     <param name="changesgenerate">enable</param>
   </service>
   <service name="set_version" mode="manual"/>
-
-  <service name="tar" mode="buildtime" />
-  <service name="recompress" mode="buildtime">
-    <param name="file">*.tar</param>
-    <param name="compression">xz</param>
-  </service>
 </services>
 

++++++ _servicedata ++++++
--- /var/tmp/diff_new_pack.yJPyno/_old  2026-04-14 17:49:35.738704528 +0200
+++ /var/tmp/diff_new_pack.yJPyno/_new  2026-04-14 17:49:35.746704858 +0200
@@ -1,6 +1,6 @@
 <servicedata>
 <service name="tar_scm">
                 <param 
name="url">https://github.com/OpenShot/openshot-qt.git</param>
-              <param 
name="changesrevision">8672b30defd0530d3aea8269d4400dab3d6b4391</param></service></servicedata>
+              <param 
name="changesrevision">ad3cb296d9066cd60e910cc816b07afe710cd72d</param></service></servicedata>
 (No newline at EOF)
 

++++++ openshot-qt-disable-sending-metrics.patch ++++++
--- /var/tmp/diff_new_pack.yJPyno/_old  2026-04-14 17:49:35.766705685 +0200
+++ /var/tmp/diff_new_pack.yJPyno/_new  2026-04-14 17:49:35.770705850 +0200
@@ -1,14 +1,14 @@
-Index: openshot-qt-3.2.0/src/settings/_default.settings
+Index: openshot-qt-3.5.1/src/settings/_default.settings
 ===================================================================
---- openshot-qt-3.2.0.orig/src/settings/_default.settings
-+++ openshot-qt-3.2.0/src/settings/_default.settings
-@@ -379,7 +379,7 @@
-     "restart": true
+--- openshot-qt-3.5.1.orig/src/settings/_default.settings
++++ openshot-qt-3.5.1/src/settings/_default.settings
+@@ -515,7 +515,7 @@
+     "restart": false
    },
    {
 -    "value": true,
 +    "value": false,
      "title": "Send Anonymous Metrics and Errors",
      "type": "bool",
-     "category": "Debug",
+     "category": "Advanced",
 

++++++ openshot-qt.obsinfo ++++++
--- /var/tmp/diff_new_pack.yJPyno/_old  2026-04-14 17:49:35.802707173 +0200
+++ /var/tmp/diff_new_pack.yJPyno/_new  2026-04-14 17:49:35.806707339 +0200
@@ -1,5 +1,5 @@
 name: openshot-qt
-version: 3.3.0
-mtime: 1734835066
-commit: 8672b30defd0530d3aea8269d4400dab3d6b4391
+version: 3.5.1
+mtime: 1775428957
+commit: ad3cb296d9066cd60e910cc816b07afe710cd72d
 

Reply via email to