mirror of
https://github.com/discourse/discourse.git
synced 2026-08-04 10:39:43 +08:00
Previously, important image-processing behavior lacked precise regression coverage, making image-library upgrades difficult to assess safely. This change adds coverage without changing production behavior: - Exercises public upload and image-model outcomes for SVG sanitization, format detection and conversion, transparency, animation metadata, EXIF orientation normalization, malformed input rejection, crop positioning, tiny-image resizing, dominant-color caching, and image-quality selection. - Keeps normal LetterAvatar RSpec coverage portable by generating one representative PNG and checking its format and requested dimensions. - Adds a separate Linux-only `script/letter_avatar_pixel_diff` workflow that compares A–Z renders with committed baselines and generates an HTML expected/actual/highlighted-diff report. Use `--update-baselines` when intentionally accepting rendering changes. - Uses deterministic fixtures with regeneration guidance beside the relevant specs.
318 B
Vendored
3x5px
318 B
Vendored
3x5px