migrations/egui
1
0
Fork 0
mirror of https://github.com/emilk/egui.git synced 2026-09-02 23:00:04 -04:00
Code Activity

Rename failed_pixel_count_threshold to max_failed_pixels (#8383)

Browse Source 
It was confusing that both tolerances had "threshold" in the name

---------

Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Emil Ernerfeldt
2026-08-03 09:02:41 -07:00
committed by GitHub
parent 49d4befe6b
commit dae9adf307
3 changed files with 163 additions and 51 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
crates/egui_kittest/README.md
View File

@@ -44,25 +44,32 @@ All possible settings and their defaults:
# path to the snapshot directory
output_path = "tests/snapshots"
# default threshold for image comparison tests
# maximum weighted squared YIQ color distance between two corresponding pixels
# (a per-pixel color tolerance, applied to each pixel pair on its own)
threshold = 0.6
# default failed_pixel_count_threshold
failed_pixel_count_threshold = 0
# how many pixels may exceed the `threshold` before the test fails
# (an absolute pixel count, not a fraction of the image)
max_failed_pixels = 0
[windows]
threshold = 0.6
failed_pixel_count_threshold = 0
max_failed_pixels = 0
[macos]
threshold = 0.6
failed_pixel_count_threshold = 0
max_failed_pixels = 0
[linux]
threshold = 0.6
failed_pixel_count_threshold = 0
max_failed_pixels = 0
```
Raise `max_failed_pixels` only very carefully: a high value (more than ~10) is enough to hide a
real change, such as a moved separator, a shifted one-pixel border, or a small icon rendering
incorrectly. Prefer the smallest value that makes the test pass, and re-check it whenever you
update the snapshot.
## Snapshot testing
There is a snapshot testing feature. To create snapshot tests, enable the `snapshot` and `wgpu` features.
Once enabled, you can call `Harness::snapshot` to render the ui and save the image to the `tests/snapshots` directory.
@@ -105,7 +112,7 @@ However, especially when you're using custom rendering, you may observe images d
First check whether the difference is due to a change in enabled rendering features, potentially due to difference in hardware (/software renderer) capabilities.
Generally you should carefully enforcing the same set of features for all test runs, but this may happen nonetheless.
Once you validated that the differences are miniscule and hard to avoid, you can try to _carefully_ adjust the comparison tolerance setting (`SnapshotOptions::threshold`, TODO([#5683](https://github.com/emilk/egui/issues/5683)): as well as number of pixels allowed to differ) for the specific test.
Once you validated that the differences are miniscule and hard to avoid, you can try to _carefully_ adjust the comparison tolerances (`SnapshotOptions::threshold` and, as a last resort, `SnapshotOptions::max_failed_pixels`) for the specific test. See also TODO([#5683](https://github.com/emilk/egui/issues/5683)).
⚠️ **WARNING** ⚠️
Picking too high tolerances may mean that you are missing actual test failures.