mltframework/shotcut
Pull Request Overview
- Opened on September 30, 2026
- Status Merged
- Commit count 2 with first commit September 30, 2026
Total Delta
Open Days
Test Delta
How long has this pull request spent in each phase of its lifecycle?
| Fraction of total time | Business days | Phase |
|---|---|---|
|
|
0.0 days | Authoring 1 commit before pull request opened for review |
|
|
0.0 days | Awaiting first review |
|
|
0.2 days | Revising work with 1 commit in response to 3 reviews that left 5 comments |
| Set up deploy tracking to begin tracking full PR start-to-deploy times | ||
Total time for pull request
0.2 business days
from first commit authored to merge
Deploy tracking has not been set up
Fix paste filter to track regression
See bug report https://forum.shotcut.org/t/copying-a-filter-from-a-clip-to-a-track-creates-an-empty-fake/51989
This pull request improves the handling of filter ordering and indexing in the filter management logic, particularly to ensure correct placement of user-visible and hidden filters (such as the track audio-level meter) and to fix related bugs in filter paste and model row mapping. The changes enhance both internal correctness and user interface consistency.
Filter Ordering and Placement Improvements:
- Added a new
firstPostUserFilterIndexfunction inmltcontroller.cppto accurately locate the first hidden post-user filter, ensuring that track volume filters and pasted filters are positioned correctly relative to hidden filters like the audio-level meter. Updated themoveTrackVolumeFilterToEndfunction to use this logic for more precise placement. - Updated the filter pasting logic in
copyFiltersto ensure that pasted filters are placed ahead of hidden post filters and that GPU filters are inserted at the correct destination index, preventing UI binding issues and maintaining correct filter order. - Removed redundant adjustment of filters in
pasteFilters, as this is now handled more accurately withincopyFiltersandmoveTrackVolumeFilterToEnd.
Model Row and Filter Index Mapping:
- Replaced the
userFilterCountfunction with a newuserFilterMltIndexfunction inattachedfiltersmodel.cppto correctly map model rows to MLT filter indices, skipping hidden filters and ensuring user actions target the correct filters. UpdatedmltFilterIndexto use this new logic.
Comments Threads Pending Resolution
## Copilot review overview
### 🔵 Needs a closer look
A moderate filter-paste ordering issue remains when GPU and non-GPU filters are interleaved.
**Review effort:** Lite
**Findings:** None
Resolved since last review (1)
- 
[Use correct user-filter index when inserting service filters](#discussion_r4140155930)
## Copilot review overview
### 🟡 Changes recommended
Moderate issues remain in paste refresh behavior, GPU insertion ordering, and filter insertion mapping.
**Review effort:** Lite
**Findings:** 1
Open (1)
- [Use correct user-filter index when inserting service filters](#discussion_r4140155930) · New
What changed in this PR
Fixes filter ordering and visible-row-to-MLT index mapping when pasting filters around hidden track filters.
**Changes:**
- Adds hidden-filter-aware user filter indexing.
- Improves pasted filter and track-volume placement.
- Removes redundant paste adjustment.
| File | Description |
|---|---|
| `src/models/attachedfiltersmodel.cpp` | Maps visible filter rows to actual MLT indices. |
| `src/mltcontroller.cpp` | Handles pasted filter placement and hidden-filter ordering. |
---
💡 Add a `code-review` agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
Resolved Comment Threads
No resolved comments have been left on this PR.
Adding the track volume and level hidden filters certainly added some complexity. This change looks good to me. I did not test this. But the logic is sound and I will do some testing after it merges.