GRI-2156: acceptance test for in-slide object animations (forge #197) #291

Merged
vivy merged 3 commits from test/GRI-2156-pptx-object-animations into dev 2026-11-16 20:08:04 +00:00
Member

Summary

Acceptance test, before implementation, for forge #197 (GRI-2156): the "done
when" criterion that a deck with an entrance, an emphasis and a motion path
must play those build steps on the screen, not only the slide's final state.

  • New fixture crates/server/tests/fixtures/animated-deck.pptx: a one-slide
    deck built with python-pptx, with a real p:timing tree injected into
    the slide XML afterwards (entrance "Appear", emphasis colour change,
    motion path), since python-pptx itself cannot author animations.
  • New test pptx_video::tests::a_deck_with_object_animations_shows_its_build_steps_over_time
    (#[ignore]d, real tools, same convention as the existing
    the_sample_deck_renders_with_the_real_soffice_and_pdftoppm): renders the
    fixture with the real render_fallback_pages + join_slides_into_mp4,
    then pulls two frames out of the joined MP4 at different points in the
    one slide's dwell time and asserts they differ.

This fails today: render_fallback_pages renders exactly one PNG per slide
(the slide's final design state, animations or not — see this module's own
doc comment, "What this cannot do"), and for a single slide
build_filter_complex is [v0]copy[outv], so the two sampled frames come
out byte-identical regardless of the deck's animations.

Ran on this build host with a locally-extracted pdftoppm (poppler-utils is
not installed on this particular sandbox, though install.sh puts it on
every real deployment target):

cargo test -p server --lib "pptx_video::tests::a_deck_with_object_animations_shows_its_build_steps_over_time" -- --ignored --exact
test pptx_video::tests::a_deck_with_object_animations_shows_its_build_steps_over_time ... FAILED
thread '...' panicked at crates/server/src/pptx_video.rs:1592:9:
assertion `left != right` failed: a deck with an entrance, an emphasis and
a motion path must look different early and late in its one slide's dwell
time; today's render is the slide's final state for the whole dwell, so
these two frames come out byte-identical
test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 214 filtered out

./scripts/local-gates.sh --rust-only passes clean on this branch (fmt,
clippy, the full cargo test including the real-DB acceptance suites and
pptx_video_acceptance_integration.rs, cargo test --doc, cargo deny).
The new test does not run under that gate — it is #[ignore]d like its
sibling, needing real soffice/pdftoppm/ffmpeg on PATH.

Criterion mapping

  • "a deck with an entrance, an emphasis and a motion path plays them on a
    screen" → a_deck_with_object_animations_shows_its_build_steps_over_time.

Not covered by this PR — a contract gap for GRI-2133/Nyssa

The ticket's second criterion — "the #189 upload note is not shown for a
fully supported deck, and is still shown for a deck with unsupported
effects" — has no existing interface to test against. There is no "note"
or "warning" field anywhere in the server's media API or the frontend
upload flow today (#189's own scope never added one; it was proposed only
for the SVG-slideshow option that was not chosen). Writing a test for it
means picking where that note surfaces (a new media-item field? a response
on upload? a job-result field?), which is an interface decision, not a test
case. Flagging this on GRI-2156 and GRI-2133 rather than guessing a shape.

Test plan

  • cargo test -p server --lib "pptx_video::tests::a_deck_with_object_animations_shows_its_build_steps_over_time" -- --ignored --exact fails against unimplemented code (output above)
  • ./scripts/local-gates.sh --rust-only passes on this branch
  • cargo fmt -p server -- --check clean
## Summary Acceptance test, before implementation, for forge #197 (GRI-2156): the "done when" criterion that a deck with an entrance, an emphasis and a motion path must play those build steps on the screen, not only the slide's final state. - New fixture `crates/server/tests/fixtures/animated-deck.pptx`: a one-slide deck built with `python-pptx`, with a real `p:timing` tree injected into the slide XML afterwards (entrance "Appear", emphasis colour change, motion path), since `python-pptx` itself cannot author animations. - New test `pptx_video::tests::a_deck_with_object_animations_shows_its_build_steps_over_time` (`#[ignore]`d, real tools, same convention as the existing `the_sample_deck_renders_with_the_real_soffice_and_pdftoppm`): renders the fixture with the real `render_fallback_pages` + `join_slides_into_mp4`, then pulls two frames out of the joined MP4 at different points in the one slide's dwell time and asserts they differ. This fails today: `render_fallback_pages` renders exactly one PNG per slide (the slide's final design state, animations or not — see this module's own doc comment, "What this cannot do"), and for a single slide `build_filter_complex` is `[v0]copy[outv]`, so the two sampled frames come out byte-identical regardless of the deck's animations. Ran on this build host with a locally-extracted `pdftoppm` (poppler-utils is not installed on this particular sandbox, though `install.sh` puts it on every real deployment target): ``` cargo test -p server --lib "pptx_video::tests::a_deck_with_object_animations_shows_its_build_steps_over_time" -- --ignored --exact ``` ``` test pptx_video::tests::a_deck_with_object_animations_shows_its_build_steps_over_time ... FAILED thread '...' panicked at crates/server/src/pptx_video.rs:1592:9: assertion `left != right` failed: a deck with an entrance, an emphasis and a motion path must look different early and late in its one slide's dwell time; today's render is the slide's final state for the whole dwell, so these two frames come out byte-identical test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 214 filtered out ``` `./scripts/local-gates.sh --rust-only` passes clean on this branch (fmt, clippy, the full `cargo test` including the real-DB acceptance suites and `pptx_video_acceptance_integration.rs`, `cargo test --doc`, `cargo deny`). The new test does not run under that gate — it is `#[ignore]`d like its sibling, needing real `soffice`/`pdftoppm`/`ffmpeg` on `PATH`. ## Criterion mapping - "a deck with an entrance, an emphasis and a motion path plays them on a screen" → `a_deck_with_object_animations_shows_its_build_steps_over_time`. ## Not covered by this PR — a contract gap for GRI-2133/Nyssa The ticket's second criterion — "the #189 upload note is not shown for a fully supported deck, and is still shown for a deck with unsupported effects" — has no existing interface to test against. There is no "note" or "warning" field anywhere in the server's media API or the frontend upload flow today (`#189`'s own scope never added one; it was proposed only for the SVG-slideshow option that was not chosen). Writing a test for it means picking where that note surfaces (a new media-item field? a response on upload? a job-result field?), which is an interface decision, not a test case. Flagging this on GRI-2156 and GRI-2133 rather than guessing a shape. ## Test plan - [x] `cargo test -p server --lib "pptx_video::tests::a_deck_with_object_animations_shows_its_build_steps_over_time" -- --ignored --exact` fails against unimplemented code (output above) - [x] `./scripts/local-gates.sh --rust-only` passes on this branch - [x] `cargo fmt -p server -- --check` clean
GRI-2156: acceptance test for in-slide build steps (forge #197)
All checks were successful
CI / the pre-push hook (pull_request) Successful in 3s
CI / cargo fmt (pull_request) Successful in 3s
CI / install.sh client upgrade (pull_request) Successful in 13s
CI / cargo deny (pull_request) Successful in 14s
CI / cargo deny advisories (pull_request) Successful in 17s
CI / repository boundary (pull_request) Successful in 5s
CI / cargo clippy (pull_request) Successful in 1m7s
CI / frontend-web (pull_request) Successful in 59s
CI / cargo test (pull_request) Successful in 3m22s
CI / Notify gate failed (pull_request) Has been skipped
CI / CI (pull_request) Successful in 1s
7c8f6d8a59
Adds a_deck_with_object_animations_shows_its_build_steps_over_time,
an ignored real-tool test (needs soffice, pdftoppm, ffmpeg): a one-slide
fixture deck with an entrance, an emphasis and a motion path must show
different content early and late in its dwell time once build steps are
rendered as extra frames/segments. Today it renders one static PNG per
slide regardless of animations, so the two sampled frames come out
byte-identical and the test fails -- recorded on GRI-2156.
Merge origin/dev into test/GRI-2156-pptx-object-animations
All checks were successful
CI / the pre-push hook (pull_request) Successful in 4s
CI / cargo fmt (pull_request) Successful in 4s
CI / install.sh client upgrade (pull_request) Successful in 12s
CI / cargo deny (pull_request) Successful in 17s
CI / cargo deny advisories (pull_request) Successful in 18s
CI / repository boundary (pull_request) Successful in 5s
CI / cargo clippy (pull_request) Successful in 1m6s
CI / frontend-web (pull_request) Successful in 59s
CI / cargo test (pull_request) Successful in 3m16s
CI / Notify gate failed (pull_request) Has been skipped
CI / CI (pull_request) Successful in 1s
bd0447b0d1
vivy merged commit fc90997a3c into dev 2026-11-16 20:08:04 +00:00
vivy deleted branch test/GRI-2156-pptx-object-animations 2026-11-16 20:08:05 +00:00
Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
Grimoirix/grimoire-of-digital-signage!291
No description provided.