Compare commits

..
22 changed files with 419 additions and 444 deletions
-5
View File
@@ -1,10 +1,6 @@
name: Release
on:
# A nightly dry run on main: every release build except macOS, nothing signed or released
# (craft-common-actions release.yml `dry-run`). Finds release-only breakage before release day.
schedule:
- cron: "16 1 * * *"
push:
branches: [release]
workflow_dispatch:
@@ -34,7 +30,6 @@ jobs:
# Intentionally tracked at @main (not SHA-pinned) so org-wide CI improvements, security fixes, and build optimizations propagate without PR churn across all craftapp repos.
uses: storytold/craft-common-actions/.github/workflows/release.yml@main
with:
dry-run: ${{ github.event_name == 'schedule' }}
app-name: deckcraft
app-title: DeckCraft
# Keep identical to ci.yml and freebsd.yml. The shared release workflow sets
Generated
+24 -24
View File
@@ -1038,7 +1038,7 @@ checksum = "0c87e182de0887fd5361989c677c4e8f5000cd9491d6d563161a8f3a5519fc7f"
[[package]]
name = "deckcraft"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"arboard",
"chrono",
@@ -1060,7 +1060,7 @@ dependencies = [
[[package]]
name = "deckcraft-anim"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"deckcraft-geom",
"deckcraft-model",
@@ -1069,7 +1069,7 @@ dependencies = [
[[package]]
name = "deckcraft-av1"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"deckcraft-bitstream",
"rayon",
@@ -1078,14 +1078,14 @@ dependencies = [
[[package]]
name = "deckcraft-bitstream"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"craft-media",
]
[[package]]
name = "deckcraft-cli"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"deckcraft-engine",
"deckcraft-mcp",
@@ -1095,14 +1095,14 @@ dependencies = [
[[package]]
name = "deckcraft-color"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"serde",
]
[[package]]
name = "deckcraft-engine"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"chrono",
"deckcraft-anim",
@@ -1128,7 +1128,7 @@ dependencies = [
[[package]]
name = "deckcraft-fonts"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"epaint_default_fonts",
"harfrust 0.13.3",
@@ -1140,7 +1140,7 @@ dependencies = [
[[package]]
name = "deckcraft-format"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"deckcraft-model",
"serde",
@@ -1151,7 +1151,7 @@ dependencies = [
[[package]]
name = "deckcraft-geom"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"kurbo",
"serde",
@@ -1159,7 +1159,7 @@ dependencies = [
[[package]]
name = "deckcraft-h264"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"deckcraft-bitstream",
"rayon",
@@ -1168,7 +1168,7 @@ dependencies = [
[[package]]
name = "deckcraft-hevc"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"deckcraft-bitstream",
"rayon",
@@ -1177,7 +1177,7 @@ dependencies = [
[[package]]
name = "deckcraft-mcp"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"deckcraft-engine",
"deckcraft-render",
@@ -1187,7 +1187,7 @@ dependencies = [
[[package]]
name = "deckcraft-media"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"craft-audio",
"craft-media",
@@ -1203,7 +1203,7 @@ dependencies = [
[[package]]
name = "deckcraft-metafile"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"image",
"kurbo",
@@ -1212,7 +1212,7 @@ dependencies = [
[[package]]
name = "deckcraft-model"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"deckcraft-color",
"deckcraft-geom",
@@ -1224,7 +1224,7 @@ dependencies = [
[[package]]
name = "deckcraft-pdf"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"craft-pdf",
"deckcraft-fonts",
@@ -1242,7 +1242,7 @@ dependencies = [
[[package]]
name = "deckcraft-pptx"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"deckcraft-color",
"deckcraft-engine",
@@ -1258,7 +1258,7 @@ dependencies = [
[[package]]
name = "deckcraft-render"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"deckcraft-anim",
"deckcraft-color",
@@ -1277,7 +1277,7 @@ dependencies = [
[[package]]
name = "deckcraft-text"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"chrono",
"deckcraft-color",
@@ -1291,7 +1291,7 @@ dependencies = [
[[package]]
name = "deckcraft-ui-egui"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"craft-ui",
"craft-update",
@@ -1317,7 +1317,7 @@ dependencies = [
[[package]]
name = "deckcraft-vp9"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"deckcraft-bitstream",
"rayon",
@@ -1326,7 +1326,7 @@ dependencies = [
[[package]]
name = "deckcraft-web"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"craft-web-api",
"craft-web-host",
@@ -6311,7 +6311,7 @@ checksum = "9440ea3e5aeabb0ac63af70daf835274065238cdd0cec83418f417eae38bacee"
[[package]]
name = "xtask"
version = "0.18.1"
version = "0.18.0"
dependencies = [
"serde_json",
]
+1 -1
View File
@@ -4,7 +4,7 @@ members = ["crates/*", "apps/*", "xtask"]
default-members = ["crates/*", "apps/deckcraft", "apps/deckcraft-cli"]
[workspace.package]
version = "0.18.1"
version = "0.18.0"
edition = "2024"
license = "MIT OR Apache-2.0"
rust-version = "1.95"
+3
View File
@@ -148,6 +148,9 @@ fn warn_if_damaged(path: &str, opened: &serde_json::Value) {
if slides > 0 {
eprintln!("deckcraft-cli: warning: {path}: {slides} slide{} couldn't be read and left out", if slides == 1 { "" } else { "s" });
}
for w in opened.get("importWarnings").and_then(|v| v.as_array()).into_iter().flatten().filter_map(|w| w.as_str()) {
eprintln!("deckcraft-cli: warning: {path}: {w}");
}
if media > 0 {
eprintln!("deckcraft-cli: warning: {path}: {media} picture or media file{} couldn't be read", if media == 1 { "" } else { "s" });
}
+5 -55
View File
@@ -118,8 +118,6 @@ enum Output {
grain: FilmGrainParams,
apply_film_grain: bool,
pts: i64,
/// What copying the picture out costs, in the reference's own format.
bytes: u64,
},
}
@@ -134,7 +132,7 @@ impl Output {
fn take(self) -> Result<Picture> {
match self {
Output::Decoded(s) => s.wait().map(|p| Arc::try_unwrap(p).unwrap_or_else(|p| (*p).clone())),
Output::Existing { slot, seq, width, height, grain, apply_film_grain, pts, .. } => {
Output::Existing { slot, seq, width, height, grain, apply_film_grain, pts } => {
let rf = slot.wait()?;
Ok(output_picture(&seq, &rf, width, height, &grain, apply_film_grain, pts))
}
@@ -381,8 +379,8 @@ impl Decoder {
// Output policy (7.18.1 note): with scalability, show one frame per temporal unit, the
// highest spatial layer present.
let scalable = self.seq.as_ref().is_some_and(|s| s.op_idc != 0);
let kept: Vec<Output> = if !scalable {
out.into_iter().map(|(o, _, _)| o).collect()
if !scalable {
self.out.extend(out.into_iter().map(|(o, _, _)| o));
} else {
let mut last: Option<(usize, u32)> = None;
let mut sel: Vec<Output> = Vec::new();
@@ -402,27 +400,8 @@ impl Decoder {
}
}
}
sel
};
// Every repeat is copied out whole, so the repeats this call keeps (after the layer
// selection) must not queue more than half the decoder budget; the first one over it and
// everything after is dropped.
let mut r = r;
let mut repeat_bytes = 0u64;
let mut ok = Vec::with_capacity(kept.len());
for o in kept {
if let Output::Existing { bytes, .. } = &o {
repeat_bytes = repeat_bytes.saturating_add(*bytes);
if repeat_bytes > self.memory_budget / 2 {
if r.is_ok() {
r = Err(Error::Invalid("repeated frame outputs exceed the decoder memory budget"));
}
break;
}
}
ok.push(o);
self.out.extend(sel);
}
self.out.extend(ok);
r
}
@@ -524,15 +503,6 @@ impl Decoder {
let idx = fh.frame_to_show_map_idx;
let slot = self.ref_frames[idx].clone().ok_or(Error::Invalid("show_existing_frame of an empty slot"))?;
let info = self.refs[idx].clone();
// Sized in the reference's own format: a new sequence header may differ from it.
let shape = deckcraft_bitstream::budget::PictureShape {
width: u64::from(info.upscaled_width),
height: u64::from(info.frame_height),
shift_x: u32::from(info.subsampling_x),
shift_y: u32::from(info.subsampling_y),
chroma_planes: if info.mono_chrome { 0 } else { 2 },
bytes_per_sample: 2,
};
let o = Output::Existing {
slot: slot.clone(),
seq: seq.clone(),
@@ -541,7 +511,6 @@ impl Decoder {
grain: fh.film_grain.clone(),
apply_film_grain: self.apply_film_grain,
pts: self.pts,
bytes: shape.bytes().unwrap_or(u64::MAX),
};
if fh.frame_type == KEY_FRAME as u8 {
// reference frame loading process (7.21) then refresh every slot (7.20)
@@ -660,9 +629,6 @@ impl Decoder {
ri.lf_deltas = fh.lf.deltas;
ri.seg_features = fh.seg.features;
ri.grain = fh.film_grain.clone();
ri.subsampling_x = job.seq.color.subsampling_x;
ri.subsampling_y = job.seq.color.subsampling_y;
ri.mono_chrome = job.seq.color.mono_chrome;
self.ref_frames[i] = Some(job.slot.clone());
}
}
@@ -811,14 +777,6 @@ fn decode_frame(job: &FrameJob, pool: &Pool, stats: &mut DecodeStats) -> Result<
Ok((rf, pic))
}
/// Bytes all parallel tile states may hold of whole-frame CDEF and restoration metadata; past it
/// the tiles decode one after another on the frame's own metadata.
const TILE_METADATA_BUDGET: usize = 256 << 20;
fn tile_metadata_fits(tiles: usize, per_tile: usize) -> bool {
tiles.checked_mul(per_tile).is_some_and(|n| n <= TILE_METADATA_BUDGET)
}
/// Decode the tiles of a frame (tile number, data), in parallel when several; returns the final
/// CDFs of tile context_update_tile_id.
fn decode_tiles(pool: &Pool, fs: &mut FrameState, cdfs: &Cdfs, tiles: &[(usize, Vec<u8>)]) -> Result<Option<Box<Cdfs>>> {
@@ -832,7 +790,7 @@ fn decode_tiles(pool: &Pool, fs: &mut FrameState, cdfs: &Cdfs, tiles: &[(usize,
};
// Intra block copy reads the current frame through the tile's own buffer: keep whole-frame
// buffers (tiles one after another) for it.
if tiles.len() > 1 && pool.is_parallel() && !fs.fh.allow_intrabc && tile_metadata_fits(tiles.len(), fs.tile_metadata_bytes()) {
if tiles.len() > 1 && pool.is_parallel() && !fs.fh.allow_intrabc {
let mut states: Vec<TileState> = tiles
.iter()
.map(|(num, _)| {
@@ -1117,14 +1075,6 @@ mod guard_tests {
#[cfg(test)]
mod budget_tests {
use super::*;
#[test]
fn tile_metadata_copies_are_budgeted() {
assert!(tile_metadata_fits(64, 1 << 20));
assert!(!tile_metadata_fits(4096, 1 << 20));
assert!(!tile_metadata_fits(usize::MAX, 2));
}
use deckcraft_bitstream::budget::DECODE_MEMORY_BUDGET;
fn color(mono: bool) -> crate::header::ColorConfig {
-4
View File
@@ -360,10 +360,6 @@ pub struct RefInfo {
pub lf_deltas: LoopFilterDeltas,
pub seg_features: SegmentationFeatures,
pub grain: FilmGrainParams,
/// The colour format the frame was decoded in (a later sequence header may change it).
pub subsampling_x: u8,
pub subsampling_y: u8,
pub mono_chrome: bool,
}
/// A parsed uncompressed frame header with its derived variables.
-7
View File
@@ -253,13 +253,6 @@ impl FrameState {
self.lr = t.lr;
}
/// Bytes of CDEF and loop-restoration metadata each parallel tile state copies whole.
pub fn tile_metadata_bytes(&self) -> usize {
let lr = &self.lr;
// 1 (type) + 6 (Wiener) + 1 (SGR set) + 2 (SGR weights) bytes per unit.
self.cdef.len() + lr.lr_type.iter().map(|v| v.len() * 10).sum::<usize>()
}
/// Private buffers for the tile covering mode info rows [r0, r1) x columns [c0, c1); tiles
/// on the right / bottom frame edge also cover the allocation margins.
pub fn tile_state(&self, r0: usize, r1: usize, c0: usize, c1: usize) -> TileState {
+9
View File
@@ -433,6 +433,15 @@ pub(crate) fn shape_of(s: &Session, id: ShapeId) -> Result<deckcraft_model::Shap
pub fn xfrm_of(doc: &deckcraft_model::Presentation, sel: &crate::Selection, shape: &deckcraft_model::Shape) -> Xfrm {
let mut x = shape.xfrm.unwrap_or_else(|| ctx_xfrm(doc, sel, shape));
if let deckcraft_model::ShapeKind::Table(t) = &shape.kind {
// A loaded table may not be normalized yet: measure the bounded grid an edit will
// leave, not every cell the file stored.
let normalized;
let t = if t.is_normalized() {
t
} else {
normalized = t.normalized();
&normalized
};
let drawn: f64 = table_drawn_heights(doc, sel, t).iter().sum();
if drawn.is_finite() && drawn > x.h {
x.h = drawn;
+17 -9
View File
@@ -139,11 +139,13 @@ pub struct Damage {
pub media: Vec<String>,
/// Slides a `.pptx` lists but that couldn't be read, so they are not in the presentation.
pub slides: Vec<String>,
/// Content a `.pptx` had that was cut to fit a limit; the file still opened.
pub warnings: Vec<String>,
}
impl Damage {
pub fn is_empty(&self) -> bool {
self.media.is_empty() && self.slides.is_empty()
self.media.is_empty() && self.slides.is_empty() && self.warnings.is_empty()
}
}
@@ -151,17 +153,17 @@ impl Damage {
pub fn open_checked(name: &str, bytes: &[u8]) -> Result<(Presentation, Damage)> {
if deckcraft_format::sniff(bytes) {
let (p, media) = deckcraft_format::load_checked(bytes).map_err(|e| EngineError::Other(e.to_string()))?;
return Ok((p, Damage { media, slides: vec![] }));
return Ok((p, Damage { media, ..Default::default() }));
}
open_other(name, bytes).map(|(p, slides)| (p, Damage { media: vec![], slides }))
open_other(name, bytes).map(|(p, slides, warnings)| (p, Damage { slides, warnings, ..Default::default() }))
}
fn open_other(name: &str, bytes: &[u8]) -> Result<(Presentation, Vec<String>)> {
fn open_other(name: &str, bytes: &[u8]) -> Result<(Presentation, Vec<String>, Vec<String>)> {
if deckcraft_pptx::sniff(bytes) {
let (mut p, lost_slides) = deckcraft_pptx::import_checked(bytes).map_err(|e| EngineError::Other(e.to_string()))?;
let (mut p, lost_slides, warnings) = deckcraft_pptx::import_report(bytes).map_err(|e| EngineError::Other(e.to_string()))?;
deckcraft_format::repair(&mut p);
probe_clips(&mut p);
return Ok((p, lost_slides));
return Ok((p, lost_slides, warnings));
}
if deckcraft_pptx::is_encrypted(bytes) {
return Err(EngineError::Other(format!("{name} is password-protected. DeckCraft can't open password-protected presentations yet.")));
@@ -180,7 +182,7 @@ fn open_other(name: &str, bytes: &[u8]) -> Result<(Presentation, Vec<String>)> {
if p.slides.is_empty() {
p = defaults::new_presentation(None);
}
return Ok((p, vec![]));
return Ok((p, vec![], vec![]));
}
Err(EngineError::Other(format!("{name}: not a presentation DeckCraft can read")))
}
@@ -276,7 +278,7 @@ fn adopt_loaded(s: &mut Session, f: LoadedFile) -> Result<Value> {
let st = s.doc()?;
Ok(json!({
"document": i, "slides": st.doc.slides.len(), "missingFonts": st.missing_fonts,
"damagedMedia": damaged.media, "lostSlides": damaged.slides,
"damagedMedia": damaged.media, "lostSlides": damaged.slides, "importWarnings": damaged.warnings,
}))
}
@@ -311,7 +313,9 @@ fn open_bytes(s: &mut Session, p: &Value) -> Result<Value> {
let (doc, damaged) = open_checked(name, &data)?;
let i = add_opened(s, name, None, doc);
report_damage(s, &damaged)?;
Ok(json!({"document": i, "missingFonts": s.doc()?.missing_fonts, "damagedMedia": damaged.media, "lostSlides": damaged.slides}))
Ok(
json!({"document": i, "missingFonts": s.doc()?.missing_fonts, "damagedMedia": damaged.media, "lostSlides": damaged.slides, "importWarnings": damaged.warnings}),
)
}
/// Tell the user which pictures, media and slides a damaged file lost, and mark the copy repaired.
@@ -339,6 +343,10 @@ fn report_damage(s: &mut Session, damaged: &Damage) -> Result<()> {
if n == 1 { "was" } else { "were" },
));
}
for w in &damaged.warnings {
text.push_str(w);
text.push_str(". ");
}
text.push_str("It was opened as a repaired copy; save it under a new name to keep the original file.");
s.ui_requests.push(crate::UiRequest::Message { text });
Ok(())
+17 -4
View File
@@ -286,6 +286,12 @@ fn with_table(s: &mut Session, p: &Value, cmd: &str, f: impl Fn(&mut Table) -> R
};
let r = edit_shapes(s, &json!({"id": id}), cmd, |sh| match &mut sh.kind {
ShapeKind::Table(t) => {
// Edits walk the grid: bound it first (a loaded table need not be normalized yet).
// An over-cap table is refused, not trimmed: a trim would delete rows while reporting success.
if t.over_cap() {
return Err(bad(cmd, format!("the table is larger than {} cells; edit refused", deckcraft_model::table::MAX_CELLS)));
}
t.normalize();
f(t)?;
t.normalize();
let (w, h) = (t.width(), t.height());
@@ -319,10 +325,17 @@ fn rows_cols(s: &mut Session, p: &Value, op: &str, name: &'static str) -> Result
let op = op.to_string();
with_table(s, p, name, move |t| {
match op.as_str() {
"rowAbove" => t.insert_row(r),
"rowBelow" => t.insert_row(r.saturating_add(1)),
"colLeft" => t.insert_col(c),
"colRight" => t.insert_col(c.saturating_add(1)),
"rowAbove" | "rowBelow" | "colLeft" | "colRight" => {
let ok = match op.as_str() {
"rowAbove" => t.insert_row(r),
"rowBelow" => t.insert_row(r.saturating_add(1)),
"colLeft" => t.insert_col(c),
_ => t.insert_col(c.saturating_add(1)),
};
if !ok {
return Err(bad(name, format!("the table is at its {}-cell limit", deckcraft_model::table::MAX_CELLS)));
}
}
"delRow" => {
if !t.delete_row(r) {
return Err(bad(name, "can't delete the last row"));
+39
View File
@@ -6921,6 +6921,45 @@ fn master_body_placeholder_format_keeps_per_level_formatting() {
assert_eq!(shown(&s), vec![Some(28.0), Some(24.0)]);
}
#[test]
fn table_edits_past_the_cell_cap_are_refused_not_trimmed() {
use deckcraft_model::table::{MAX_CELLS, Row};
let mut s = session();
let id = table_at(&mut s, &[100.0, 100.0], json!([["a", "b"], ["c", "d"]]));
let cols = 200;
let rows = MAX_CELLS / cols;
let big = deckcraft_model::Table {
cols: vec![1.0; cols],
rows: vec![Row { height: 1.0, cells: vec![Default::default(); cols] }; rows],
..Default::default()
};
for sl in std::sync::Arc::make_mut(&mut s.doc_mut().unwrap().doc).slides.iter_mut() {
for sh in std::sync::Arc::make_mut(sl).shapes.iter_mut() {
if sh.id == id {
sh.kind = deckcraft_model::ShapeKind::Table(big.clone());
}
}
}
for cmd in ["table.insertRowAbove", "table.insertRowBelow"] {
let e = s.execute(cmd, &json!({"id": id.0, "row": 0})).unwrap_err().to_string();
assert!(e.contains("cell limit"), "{cmd}: {e}");
}
let e = s.execute("table.insertColumnLeft", &json!({"id": id.0, "col": 0})).unwrap_err().to_string();
assert!(e.contains("cell limit"), "{e}");
assert_eq!(table_of(&s, id), big, "a refused insert leaves every row in place");
// A table already over the cap is refused too, not trimmed behind a success.
let huge = deckcraft_model::Table { rows: vec![Row { height: 1.0, cells: vec![Default::default(); cols] }; rows + 1], ..big.clone() };
for sl in std::sync::Arc::make_mut(&mut s.doc_mut().unwrap().doc).slides.iter_mut() {
for sh in std::sync::Arc::make_mut(sl).shapes.iter_mut() {
if sh.id == id {
sh.kind = deckcraft_model::ShapeKind::Table(huge.clone());
}
}
}
assert!(s.execute("table.deleteRow", &json!({"id": id.0, "row": 0})).is_err());
assert_eq!(table_of(&s, id).n_rows(), rows + 1);
}
/// A shape named many times in one edit is edited once and not copied per mention.
#[test]
fn repeated_target_ids_edit_the_shape_once() {
+2 -28
View File
@@ -404,36 +404,10 @@ impl Decoder {
chroma_planes: if chroma == 0 { 0 } else { 2 },
bytes_per_sample: 2,
};
let frame_bytes = |f: &crate::picture::Frame| {
(f.width as u64)
.saturating_mul(f.height as u64)
.saturating_add((f.cwidth as u64).saturating_mul(f.cheight as u64).saturating_mul(2))
.saturating_mul(2)
};
// Output frames waiting to be returned stay allocated after they leave the DPB, at their
// own size: a resolution change in this call can leave pictures of the previous SPS.
let queued: u64 = self
.out_queue
.iter()
.filter(|o| !self.dpb.entries.iter().any(|e| Arc::ptr_eq(&e.frame, &o.frame)) && !self.in_flight.iter().any(|f| Arc::ptr_eq(f, &o.frame)))
.map(|o| frame_bytes(&o.frame))
.fold(0, u64::saturating_add);
// Frames the DPB holds now count at their own size too (an SPS change flushes them into the
// queue before the new pictures are made): the DPB part is the larger of its new slots and that.
let held: u64 = self.dpb.entries.iter().map(|e| frame_bytes(&e.frame)).fold(0, u64::saturating_add);
// A queued frame still being decoded is one of the in-flight pictures already reserved, so the
// queue and the in-flight reservation overlap: charge the larger, not both.
let pictures = u64::from(sps.max_dec_pic_buffering) + 1 + self.max_in_flight as u64;
let slots = shape.store_bytes(u64::from(sps.max_dec_pic_buffering));
let flight = shape.store_bytes(self.max_in_flight as u64);
let total = shape
.store_bytes(1)
.zip(slots)
.zip(flight)
.map(|((one, slots), flight)| slots.max(held).saturating_add(one).saturating_add(flight).saturating_add(queued));
ensure!(
total.is_some_and(|b| b <= self.memory_budget),
"{}x{} pictures with {pictures} in the decoder and {queued} bytes queued exceed the memory budget",
shape.fits(pictures, self.memory_budget),
"{}x{} pictures with {pictures} in the decoder exceed the memory budget",
sps.width,
sps.height
);
+186 -16
View File
@@ -83,6 +83,12 @@ impl Default for Table {
}
}
/// Cells a table may hold after [`Table::normalize`] (PowerPoint's own limit is 75 x 75).
pub const MAX_CELLS: usize = 100_000;
/// Cell visits [`Table::merge`] may spend growing its rectangle over existing merges: a few
/// passes over the largest grid, where a real table needs one or two.
const MAX_MERGE_WORK: usize = 16 * MAX_CELLS;
impl Table {
pub fn new(rows: usize, cols: usize, width: f64, row_height: f64) -> Self {
let rows = rows.clamp(1, 1000);
@@ -113,15 +119,24 @@ impl Table {
self.rows.iter().map(|r| r.height).sum()
}
/// Insert an empty row before `at`. A merge the new row falls inside grows to cover it.
pub fn insert_row(&mut self, at: usize) {
/// Returns false, leaving the table as it was, when the row would take it past [`MAX_CELLS`].
pub fn insert_row(&mut self, at: usize) -> bool {
if self.cols.len().saturating_mul(self.rows.len().saturating_add(1)) > MAX_CELLS {
return false;
}
let at = at.min(self.rows.len());
let merges = self.merges();
let h = self.rows.get(at.saturating_sub(1)).map(|r| r.height).unwrap_or(30.0);
self.rows.insert(at, Row { height: h, cells: (0..self.cols.len()).map(|_| Cell::new("")).collect() });
self.set_merges(merges.into_iter().map(|m| m.insert_row(at)));
true
}
/// Insert an empty column before `at`. A merge the new column falls inside grows to cover it.
pub fn insert_col(&mut self, at: usize) {
/// Returns false, leaving the table as it was, when the column would take it past [`MAX_CELLS`].
pub fn insert_col(&mut self, at: usize) -> bool {
if self.cols.len().saturating_add(1).saturating_mul(self.rows.len()) > MAX_CELLS {
return false;
}
let at = at.min(self.cols.len());
let merges = self.merges();
let w = self.cols.get(at.saturating_sub(1)).copied().unwrap_or(80.0);
@@ -131,6 +146,7 @@ impl Table {
r.cells.insert(i, Cell::new(""));
}
self.set_merges(merges.into_iter().map(|m| m.insert_col(at)));
true
}
/// Delete row `at`. Merges through it shrink; a merge anchored on it moves its anchor (and
/// text) to the next row it covers.
@@ -170,8 +186,12 @@ impl Table {
true
}
/// The merged areas: every uncovered cell spanning more than one cell, clipped to the grid.
/// Merges of a well-formed table do not overlap, so together they cover at most the grid; a
/// file's overlapping spans past that are dropped, so rebuilding them cannot revisit cells
/// without bound.
fn merges(&self) -> Vec<Merge> {
let (nr, nc) = (self.rows.len(), self.cols.len());
let mut area = nr.saturating_mul(nc);
let mut out = vec![];
for (r, row) in self.rows.iter().enumerate() {
for (c, cell) in row.cells.iter().enumerate().take(nc) {
@@ -180,7 +200,10 @@ impl Table {
}
let rs = (cell.row_span.max(1) as usize).min(nr - r);
let gs = (cell.grid_span.max(1) as usize).min(nc - c);
if rs > 1 || gs > 1 {
if (rs > 1 || gs > 1)
&& let Some(left) = area.checked_sub(rs.saturating_mul(gs))
{
area = left;
out.push(Merge { r, c, rs, gs });
}
}
@@ -233,9 +256,17 @@ impl Table {
if r1 >= self.rows.len() || c1 >= self.cols.len() || (r0 == r1 && c0 == c1) {
return false;
}
// Grow the rectangle until no existing merge sticks out of it.
// Grow the rectangle until no existing merge sticks out of it. Each pass reads the whole
// grid and may grow it by one row, so the passes are charged against
// [`MAX_MERGE_WORK`]: past it the merge is refused, before anything changed.
let (mut r0, mut c0, mut r1, mut c1) = (r0, c0, r1, c1);
let pass = self.rows.len().saturating_mul(self.cols.len()).max(1);
let mut work = 0usize;
loop {
work = work.saturating_add(pass);
if work > MAX_MERGE_WORK {
return false;
}
let before = (r0, c0, r1, c1);
for (r, row) in self.rows.iter().enumerate() {
for (c, cell) in row.cells.iter().enumerate().take(self.cols.len()) {
@@ -289,21 +320,50 @@ impl Table {
return false;
}
// Spans come from the file and may reach past the table; only walk the cells that exist.
for rr in r..r.saturating_add(rs).min(self.rows.len()) {
for cc in c..c.saturating_add(gs).min(self.cols.len()) {
if let Some(x) = self.cell_mut(rr, cc) {
x.h_merge = false;
x.v_merge = false;
x.grid_span = 1;
x.row_span = 1;
}
let end = r.saturating_add(rs).min(self.rows.len());
let gs = gs.min(self.cols.len().saturating_sub(c));
for row in self.rows.get_mut(r..end).unwrap_or_default() {
for x in row.cells.iter_mut().skip(c).take(gs) {
x.h_merge = false;
x.v_merge = false;
x.grid_span = 1;
x.row_span = 1;
}
}
true
}
/// Make every row have exactly `cols.len()` cells (repair after loading).
pub fn normalize(&mut self) {
/// Whether the grid is as [`Table::normalize`] leaves it: within [`MAX_CELLS`], every row
/// exactly `cols.len()` cells wide.
pub fn is_normalized(&self) -> bool {
let n = self.cols.len();
n <= MAX_CELLS && self.rows.len() <= (MAX_CELLS / n.max(1)).max(1) && self.rows.iter().all(|r| r.cells.len() == n)
}
/// Whether [`Table::normalize`] would drop rows or columns to fit [`MAX_CELLS`].
pub fn over_cap(&self) -> bool {
let n = self.cols.len();
n > MAX_CELLS || self.rows.len() > (MAX_CELLS / n.max(1)).max(1)
}
/// What [`Table::normalize`] would leave of this table, cloning only the rows and cells that
/// survive (an oversized table is not copied whole first).
pub fn normalized(&self) -> Table {
let n = self.cols.len().min(MAX_CELLS);
let keep = (MAX_CELLS / n.max(1)).max(1);
let rows = self.rows.iter().take(keep).map(|r| Row { height: r.height, cells: r.cells.iter().take(n).cloned().collect() }).collect();
let mut t = Table { cols: self.cols[..n].to_vec(), rows, background: self.background.clone(), style: self.style.clone(), ..*self };
t.normalize();
if self.over_cap() {
t.clip_spans();
}
t
}
/// Make every row have exactly `cols.len()` cells (repair after loading). A grid over
/// [`MAX_CELLS`] loses its last rows (and columns) first, so a file naming many rows and many
/// columns but few cells cannot make this allocate their product.
pub fn normalize(&mut self) {
let truncated = self.over_cap();
self.cols.truncate(MAX_CELLS);
let n = self.cols.len();
self.rows.truncate((MAX_CELLS / n.max(1)).max(1));
for r in &mut self.rows {
r.cells.resize_with(n, || Cell::new(""));
for c in &mut r.cells {
@@ -311,6 +371,20 @@ impl Table {
c.row_span = c.row_span.max(1);
}
}
if truncated {
self.clip_spans();
}
}
/// Cut every merge span to the rows and columns that remain, so no merge reaches past a
/// table that lost rows or columns to the cell cap.
fn clip_spans(&mut self) {
let (nr, nc) = (self.rows.len(), self.cols.len());
for (ri, r) in self.rows.iter_mut().enumerate() {
for (ci, c) in r.cells.iter_mut().enumerate() {
c.row_span = c.row_span.min((nr - ri) as u32).max(1);
c.grid_span = c.grid_span.min((nc - ci) as u32).max(1);
}
}
}
}
@@ -519,6 +593,102 @@ mod tests {
}
}
/// Each cell spans two rows, uncovered: merging the last two rows would grow the rectangle a
/// row per pass over the whole grid, 10,000 passes. The merge is refused at its work bound
/// instead, leaving the table as it was; a small merge still works.
#[test]
fn merge_expansion_stops_at_its_work_bound() {
let mut t = Table::new(1_000, 10, 400.0, 30.0);
for row in &mut t.rows {
for c in &mut row.cells {
c.row_span = 2;
}
}
let before = t.clone();
assert!(!t.merge(998, 0, 999, 9));
assert_eq!(t, before);
let mut small = Table::new(4, 4, 400.0, 30.0);
assert!(small.merge(0, 0, 1, 1));
assert!(small.is_normalized() && !Table { cols: vec![1.0; 3], rows: vec![Row::default(); 2], ..Default::default() }.is_normalized());
}
#[test]
fn normalize_caps_a_sparse_grid_at_max_cells() {
// 1,000 columns and 1,000 empty rows would be a million cells.
let mut t = Table { cols: vec![10.0; 1000], rows: vec![Row::default(); 1000], ..Default::default() };
t.normalize();
assert_eq!(t.cols.len(), 1000);
assert_eq!(t.rows.len(), MAX_CELLS / 1000);
assert!(t.rows.iter().all(|r| r.cells.len() == 1000));
// Columns alone past the cap keep one row of MAX_CELLS cells.
let mut t = Table { cols: vec![1.0; MAX_CELLS + 5], rows: vec![Row::default(); 3], ..Default::default() };
t.normalize();
assert_eq!((t.cols.len(), t.rows.len()), (MAX_CELLS, 1));
// `normalized` leaves what `normalize` does.
let wide = Table { cols: vec![10.0; 1000], rows: vec![Row { height: 1.0, cells: vec![Cell::new("x"); 1500] }; 150], ..Default::default() };
let mut n = wide.clone();
n.normalize();
assert_eq!(wide.normalized(), n);
// A normal table is untouched.
let mut t = Table::new(4, 4, 400.0, 30.0);
t.normalize();
assert_eq!((t.n_rows(), t.n_cols()), (4, 4));
}
#[test]
fn overlapping_file_spans_rebuild_in_bounded_work() {
// Every cell an uncovered anchor spanning to the far corner: their areas add up to about
// n^4 / 4, but only what fits the grid is kept, so an edit's rebuild stays near n^2.
let n = 200;
let mut t = Table::new(n, n, 400.0, 30.0);
for row in &mut t.rows {
for c in &mut row.cells {
c.grid_span = u32::MAX;
c.row_span = u32::MAX;
}
}
assert_eq!(t.merges().len(), 1, "the first anchor already covers the whole grid");
assert!(t.insert_row(n));
assert_eq!(t.n_rows(), n + 1);
assert!(t.split(0, 0));
}
/// A table at the cell cap refuses another row or column instead of losing rows to
/// normalization, and the cap sits above PowerPoint's own 75 x 75.
#[test]
fn insert_past_the_cap_is_refused() {
const { assert!(MAX_CELLS >= 75 * 75) };
let mut t =
Table { cols: vec![1.0; 200], rows: vec![Row { height: 1.0, cells: vec![Cell::new(""); 200] }; MAX_CELLS / 200], ..Default::default() };
assert!(!t.over_cap());
let before = t.clone();
assert!(!t.insert_row(0));
assert!(!t.insert_col(0));
assert_eq!(t, before);
assert!(Table::new(75, 75, 100.0, 10.0).insert_row(0));
}
/// A merge that crosses the truncation boundary is clipped to the rows and columns that remain.
#[test]
fn merges_are_clipped_when_the_cap_drops_rows() {
let cols = 1000;
let rows = MAX_CELLS / cols + 1;
let mut t = Table { cols: vec![1.0; cols], rows: vec![Row { height: 1.0, cells: vec![Cell::new(""); cols] }; rows], ..Default::default() };
t.rows[0].cells[0].row_span = rows as u32;
t.rows[0].cells[cols - 1].grid_span = 5;
let n = t.normalized();
assert_eq!(n.rows.len(), rows - 1);
assert_eq!(n.rows[0].cells[0].row_span as usize, rows - 1);
assert_eq!(n.rows[0].cells[cols - 1].grid_span, 1);
t.normalize();
assert_eq!(t.rows[0].cells[0].row_span as usize, rows - 1);
// Within the cap, spans are left as the file wrote them.
let mut small = Table::new(3, 3, 30.0, 10.0);
small.rows[2].cells[2].grid_span = 4;
small.normalize();
assert_eq!(small.rows[2].cells[2].grid_span, 4);
}
fn merged() -> Table {
// 4x4 with a 2x2 merge at (1,1) holding "M".
let mut t = Table::new(4, 4, 400.0, 30.0);
@@ -552,10 +722,10 @@ mod tests {
fn row_and_column_edits_keep_merges_consistent() {
for at in 0..=4 {
let mut t = merged();
t.insert_row(at);
assert!(t.insert_row(at));
assert_topology(&t);
let mut t = merged();
t.insert_col(at);
assert!(t.insert_col(at));
assert_topology(&t);
}
for at in 0..4 {
+6
View File
@@ -70,6 +70,12 @@ pub fn import_checked(bytes: &[u8]) -> Result<(Presentation, Vec<String>), PptxE
read::import_checked(bytes)
}
/// [`import_checked`], plus warnings for content cut to fit a limit (a table past the cell cap
/// keeps its leading rows); the file still opens.
pub fn import_report(bytes: &[u8]) -> Result<(Presentation, Vec<String>, Vec<String>), PptxError> {
read::import_report(bytes)
}
/// [`import`] with a lower presentation-wide chart point budget, so tests reach the over-budget
/// path without a multi-million-point chart.
#[doc(hidden)]
+53 -196
View File
@@ -10,7 +10,6 @@ mod timing;
pub(crate) use chart::{chart_from as chart_from_el, read_chart_xml};
pub(crate) use chartex::{chartex_from, read_chartex_xml};
use std::cell::Cell;
use std::collections::{HashMap, HashSet};
use std::sync::Arc;
@@ -56,12 +55,24 @@ pub struct Imp<'a> {
pub tag_bytes: usize,
/// The tag budget ran out and the warning was given.
pub tags_capped: bool,
/// What the file lost that the reader kept going past (see [`MAX_IMPORT_WARNINGS`]).
pub warnings: Vec<String>,
/// Whether a part has relationships of its own, by normalized part name: looked up once
/// however many embedded objects point at the part.
pub has_rels: HashMap<String, bool>,
}
/// Import warnings kept; a hostile file with many oversized tables adds no more than this.
const MAX_IMPORT_WARNINGS: usize = 32;
impl Imp<'_> {
/// Record something the file lost, and log it.
pub fn warn(&mut self, msg: String) {
log::warn!("pptx: {msg}");
if self.warnings.len() < MAX_IMPORT_WARNINGS {
self.warnings.push(msg);
}
}
pub fn alloc(&mut self) -> u32 {
self.next = self.next.saturating_add(1);
self.next
@@ -100,22 +111,21 @@ pub struct Part {
pub ns: Vec<(String, String)>,
/// Prefixes bound below the root only (on a wrapper such as `mc:Choice`), each to one namespace.
pub inner_ns: Vec<(String, String)>,
/// Bytes of namespace declarations this part may still copy onto kept fragments.
ns_budget: Cell<usize>,
}
impl Part {
pub fn new(pkg: &Package, name: &str, doc: &Doc) -> Self {
let ns = doc.ns_decls();
Part { name: name.to_string(), rels: pkg.rels(name), inner_ns: inner_decls(&doc.root, &ns), ns, ns_budget: Cell::new(MAX_COPIED_NS_BYTES) }
Part { name: name.to_string(), rels: pkg.rels(name), inner_ns: inner_decls(&doc.root, &ns), ns }
}
/// Serialize an element with the namespace declarations it needs from the part root.
pub fn keep(&self, el: &El) -> String {
keep_with(el, &[&self.ns], Some(&self.ns_budget))
keep_xml(el, &self.ns)
}
/// Like [`Part::keep`], also declaring prefixes that only an enclosing wrapper binds.
pub fn keep_inner(&self, el: &El) -> String {
keep_with(el, &[&self.ns, &self.inner_ns], Some(&self.ns_budget))
let ns: Vec<(String, String)> = self.ns.iter().chain(&self.inner_ns).cloned().collect();
keep_xml(el, &ns)
}
}
@@ -150,81 +160,36 @@ fn inner_decls(root: &El, ns: &[(String, String)]) -> Vec<(String, String)> {
out
}
/// Longest namespace name copied onto a kept fragment or a resolved `mc` branch. Real ones are
/// URIs of well under 100 bytes; a longer one is not copied, so one declaration cannot be
/// multiplied by every fragment that uses its prefix.
const MAX_NS_VALUE: usize = 2048;
/// Namespace declaration bytes one part may copy onto all its kept fragments together; past it a
/// fragment keeps its markup but is not given further declarations.
const MAX_COPIED_NS_BYTES: usize = 8 << 20;
/// Prefixes `el` and its descendants use in element and attribute names, and those listed in
/// `mc:Ignorable` or `mc:MustUnderstand` (which must stay declared for the markup-compatibility rules to hold),
/// and those starting a QName-like attribute value.
fn used_prefixes(el: &El) -> HashSet<&str> {
fn rec<'a>(e: &'a El, used: &mut HashSet<&'a str>, depth: usize) {
/// Serialize `el`, declaring every prefix it uses that the source root declared.
pub fn keep_xml(el: &El, ns: &[(String, String)]) -> String {
let mut used = HashSet::new();
fn rec(e: &El, used: &mut HashSet<String>, depth: usize) {
if depth > crate::xml::MAX_DEPTH {
return;
}
if let Some((p, _)) = e.name.split_once(':') {
used.insert(p);
used.insert(p.to_string());
}
for (k, v) in &e.attrs {
if let Some((p, local)) = k.split_once(':')
for (k, _) in &e.attrs {
if let Some((p, _)) = k.split_once(':')
&& p != "xmlns"
&& p != "xml"
{
used.insert(p);
if local == "Ignorable" || local == "MustUnderstand" {
used.extend(v.split_whitespace());
}
}
// QName-valued attributes (`xsi:type="q:Item"`) need their prefix declared too.
if k != "xmlns" && !k.starts_with("xmlns:") {
for t in v.split_whitespace() {
if let Some((p, _)) = t.split_once(':')
&& !p.is_empty()
{
used.insert(p);
}
}
used.insert(p.to_string());
}
}
for c in e.elements() {
rec(c, used, depth + 1);
}
}
let mut used = HashSet::new();
rec(el, &mut used, 0);
used
}
/// Serialize `el`, declaring every prefix it uses that the source root declared.
pub fn keep_xml(el: &El, ns: &[(String, String)]) -> String {
keep_with(el, &[ns], None)
}
fn keep_with(el: &El, decls: &[&[(String, String)]], budget: Option<&Cell<usize>>) -> String {
let used = used_prefixes(el);
let mut have: HashSet<&str> = el.attrs.iter().map(|(a, _)| a.as_str()).collect();
let mut extra = Vec::new();
for (k, v) in decls.iter().flat_map(|d| d.iter()) {
let mut el = el.clone();
for (k, v) in ns {
let p = k.strip_prefix("xmlns:").unwrap_or("");
if !p.is_empty() && used.contains(p) && v.len() <= MAX_NS_VALUE && !have.contains(k.as_str()) {
let cost = k.len() + v.len();
if let Some(b) = budget {
if b.get() < cost {
continue;
}
b.set(b.get() - cost);
}
have.insert(k);
extra.push((k.clone(), v.clone()));
if !p.is_empty() && used.contains(p) && !el.attrs.iter().any(|(a, _)| a == k) {
el.attrs.push((k.clone(), v.clone()));
}
}
let mut el = el.clone();
el.attrs.extend(extra);
el.to_xml()
}
@@ -373,18 +338,10 @@ fn kept_parts(pkg: &Package, pres_path: &str) -> Vec<deckcraft_model::KeptPart>
// The part's own relationships (customXml item → itemProps) and their targets.
let rp = opc::rels_path(&name);
if let Some(rels_xml) = text(&rp) {
let mut capped = false;
for sub in pkg.rels(&name).items.iter().filter(|s| !s.external && keepable(&s.target)) {
// Checked before the part is read: repeated relationships to one large part
// must not copy it again each time.
if out.iter().any(|k| k.name.eq_ignore_ascii_case(&sub.target)) {
continue;
}
if out.len() >= 400 {
capped = true;
continue;
}
if let Some(x) = text(&sub.target) {
if let Some(x) = text(&sub.target)
&& !out.iter().any(|k| k.name.eq_ignore_ascii_case(&sub.target))
{
out.push(KeptPart {
content_type: pkg.content_type(&sub.target).unwrap_or_default(),
name: sub.target.clone(),
@@ -393,9 +350,7 @@ fn kept_parts(pkg: &Package, pres_path: &str) -> Vec<deckcraft_model::KeptPart>
});
}
}
if !capped && out.len() < 400 {
out.push(KeptPart { name: rp, xml: rels_xml, ..Default::default() });
}
out.push(KeptPart { name: rp, xml: rels_xml, ..Default::default() });
}
}
}
@@ -432,39 +387,15 @@ fn resolve_mc_rels(el: &mut El, depth: usize, rels: Option<&Rels>) {
});
let branch = choice.or_else(|| ac.child("Fallback"));
if let Some(b) = branch {
// The branch's namespace declarations (the first usable one of each name,
// in order), indexed by prefix once: each child then looks up the prefixes
// it uses instead of scanning every declaration.
let mut decls: Vec<(&String, &String)> = Vec::new();
// An oversized declaration still shadows an outer one of its name: its
// content is left unbound rather than bound to the outer namespace.
let mut by_prefix: HashMap<&str, Option<usize>> = HashMap::new();
for (k, v) in b.attrs.iter().chain(ac.attrs.iter()) {
let prefix = if k == "xmlns" { Some("") } else { k.strip_prefix("xmlns:") };
if let Some(p) = prefix
&& !by_prefix.contains_key(p)
{
by_prefix.insert(p, (v.len() <= MAX_NS_VALUE).then_some(decls.len()));
if v.len() <= MAX_NS_VALUE {
decls.push((k, v));
}
}
}
for n in &b.children {
if let Node::El(e) = n {
// Carry the namespace declarations from the branch that its content
// uses (all of them per child made children x declarations copies).
let used = used_prefixes(e);
let have: HashSet<&str> = e.attrs.iter().map(|(a, _)| a.as_str()).collect();
let mut wanted: Vec<usize> = std::iter::once("")
.chain(used.iter().copied())
.filter_map(|p| by_prefix.get(p).copied().flatten())
.filter(|&i| !have.contains(decls[i].0.as_str()))
.collect();
wanted.sort_unstable();
wanted.dedup();
let mut e = e.clone();
e.attrs.extend(wanted.into_iter().map(|i| (decls[i].0.clone(), decls[i].1.clone())));
// Carry namespace declarations from the branch onto its content.
for (k, v) in b.attrs.iter().chain(ac.attrs.iter()) {
if k.starts_with("xmlns") && !e.attrs.iter().any(|(a, _)| a == k) {
e.attrs.push((k.clone(), v.clone()));
}
}
out.push(Node::El(e));
}
}
@@ -543,6 +474,7 @@ pub fn import_theme(bytes: &[u8]) -> Result<(deckcraft_model::theme::Theme, Vec<
tags_by_part: HashMap::new(),
tag_bytes: 0,
tags_capped: false,
warnings: vec![],
has_rels: HashMap::new(),
};
let part = Part::new(&pkg, &theme_path, &doc);
@@ -557,11 +489,22 @@ pub fn import(bytes: &[u8]) -> Result<Presentation, PptxError> {
/// [`import`], plus the parts of the slide list the file names but that couldn't be read (a
/// missing or truncated slide): those slides are left out, so a caller can say so (#203).
pub fn import_checked(bytes: &[u8]) -> Result<(Presentation, Vec<String>), PptxError> {
import_with_chart_points(bytes, chart::MAX_PRESENTATION_CHART_POINTS)
import_report(bytes).map(|(p, lost, _)| (p, lost))
}
/// [`import_checked`], plus warnings for content the file had that was cut to fit a limit (a
/// table past the cell cap keeps its leading rows).
pub fn import_report(bytes: &[u8]) -> Result<(Presentation, Vec<String>, Vec<String>), PptxError> {
import_report_with_chart_points(bytes, chart::MAX_PRESENTATION_CHART_POINTS)
}
/// [`import_checked`] with the presentation's chart point budget set (tests lower it).
pub(crate) fn import_with_chart_points(bytes: &[u8], chart_points_cap: usize) -> Result<(Presentation, Vec<String>), PptxError> {
import_report_with_chart_points(bytes, chart_points_cap).map(|(p, lost, _)| (p, lost))
}
/// [`import_report`] with the presentation's chart point budget set.
fn import_report_with_chart_points(bytes: &[u8], chart_points_cap: usize) -> Result<(Presentation, Vec<String>, Vec<String>), PptxError> {
let mut lost_slides: Vec<String> = vec![];
let pkg = Package::open(bytes)?;
let root_rels = pkg.rels("");
@@ -594,6 +537,7 @@ pub(crate) fn import_with_chart_points(bytes: &[u8], chart_points_cap: usize) ->
tags_by_part: HashMap::new(),
tag_bytes: 0,
tags_capped: false,
warnings: vec![],
has_rels: HashMap::new(),
};
// Slide ids first so hyperlinks anywhere can point at slides.
@@ -625,7 +569,7 @@ pub(crate) fn import_with_chart_points(bytes: &[u8], chart_points_cap: usize) ->
p.media = std::mem::take(&mut imp.media);
p.next_id = imp.next;
p.fix_next_id();
Ok((p, lost_slides))
Ok((p, lost_slides, std::mem::take(&mut imp.warnings)))
}
/// Masters, layouts and slides to read, in file order and without duplicates: (master paths,
@@ -1526,90 +1470,3 @@ fn header_footer(p: &Presentation) -> HeaderFooter {
}
hf
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn mc_branches_carry_only_the_declarations_their_content_uses() {
// 2,000 declarations on the branch and 2,000 children: each child gets the one it uses
// (and the default), not all 2,000.
let decls: String = (0..2000).map(|i| format!(r#" xmlns:n{i}="urn:n{i}""#)).collect();
let kids: String = (0..2000).map(|i| format!("<n{i}:x/>")).collect();
let x = format!(
r#"<r xmlns:mc="http://schemas.openxmlformats.org/markup-compatibility/2006"><mc:AlternateContent><mc:Fallback xmlns="urn:d"{decls}>{kids}</mc:Fallback></mc:AlternateContent></r>"#
);
let mut d = crate::xml::parse(x.as_bytes()).unwrap();
resolve_mc(&mut d.root, 0);
let kids: Vec<&El> = d.root.elements().collect();
assert_eq!(kids.len(), 2000);
assert_eq!(kids[7].attrs, vec![("xmlns".to_string(), "urn:d".to_string()), ("xmlns:n7".to_string(), "urn:n7".to_string())]);
}
/// The branch's own declaration of a prefix wins over the `AlternateContent`'s; an oversized
/// one still shadows it (its content stays unbound); a child's own declaration is left alone.
#[test]
fn mc_branch_declarations_keep_their_precedence() {
let long = "u".repeat(MAX_NS_VALUE + 1);
let x = format!(
r#"<r xmlns:mc="http://schemas.openxmlformats.org/markup-compatibility/2006"><mc:AlternateContent xmlns:p="urn:ac" xmlns:q="urn:q-short"><mc:Fallback xmlns:p="urn:b" xmlns:q="{long}"><p:x/><q:y/><p:z xmlns:p="urn:own"/></mc:Fallback></mc:AlternateContent></r>"#
);
let mut d = crate::xml::parse(x.as_bytes()).unwrap();
resolve_mc(&mut d.root, 0);
let kids: Vec<&El> = d.root.elements().collect();
let attrs = |i: usize| kids[i].attrs.iter().map(|(k, v)| format!("{k}={v}")).collect::<Vec<_>>();
assert_eq!(attrs(0), ["xmlns:p=urn:b"]);
assert!(attrs(1).is_empty(), "{:?}", attrs(1));
assert_eq!(attrs(2), ["xmlns:p=urn:own"]);
}
#[test]
fn mc_branches_keep_prefixes_named_by_must_understand() {
let x = r#"<r xmlns:mc="http://schemas.openxmlformats.org/markup-compatibility/2006"><mc:AlternateContent><mc:Fallback xmlns:q="urn:q"><a mc:MustUnderstand="q"/></mc:Fallback></mc:AlternateContent></r>"#;
let mut d = crate::xml::parse(x.as_bytes()).unwrap();
resolve_mc(&mut d.root, 0);
let kid = d.root.elements().next().unwrap();
assert!(kid.attrs.iter().any(|(k, v)| k == "xmlns:q" && v == "urn:q"), "{:?}", kid.attrs);
}
#[test]
fn mc_branches_keep_prefixes_used_in_qname_values() {
let x = r#"<r xmlns:mc="http://schemas.openxmlformats.org/markup-compatibility/2006"><mc:AlternateContent><mc:Fallback xmlns:q="urn:q" xmlns:z="urn:z"><n xmlns:xsi="urn:xsi" xsi:type="q:Item"/></mc:Fallback></mc:AlternateContent></r>"#;
let mut d = crate::xml::parse(x.as_bytes()).unwrap();
resolve_mc(&mut d.root, 0);
let kid = d.root.elements().next().unwrap();
assert!(kid.attrs.iter().any(|(k, v)| k == "xmlns:q" && v == "urn:q"), "{:?}", kid.attrs);
assert!(!kid.attrs.iter().any(|(k, _)| k == "xmlns:z"), "unused declarations stay out: {:?}", kid.attrs);
}
#[test]
fn kept_fragments_skip_oversized_namespace_names() {
let big = "u".repeat(MAX_NS_VALUE + 1);
let ns = vec![("xmlns:a".to_string(), "urn:a".to_string()), ("xmlns:b".to_string(), big)];
let d = crate::xml::parse(br#"<a:x xmlns:a="urn:a" xmlns:b="urn:b"><b:y/></a:x>"#).unwrap();
let mut el = d.root.clone();
el.attrs.clear();
let out = keep_xml(&el, &ns);
assert!(out.contains(r#"xmlns:a="urn:a""#), "{out}");
assert!(!out.contains("uuuu"), "the oversized declaration is not copied");
}
#[test]
fn copied_namespace_bytes_are_budgeted_per_part() {
let d = crate::xml::parse(br#"<a:x xmlns:a="urn:a"><a:y/></a:x>"#).unwrap();
let mut el = d.root.clone();
el.attrs.clear();
let part = Part {
name: String::new(),
rels: Default::default(),
ns: vec![("xmlns:a".to_string(), "urn:a".to_string())],
inner_ns: vec![],
ns_budget: Cell::new(15),
};
assert!(part.keep(&el).contains(r#"xmlns:a="urn:a""#), "within budget");
assert!(part.ns_budget.get() < 15);
let again = part.keep(&el);
assert!(!again.contains("xmlns:a"), "budget spent, no further declarations: {again}");
}
}
+1 -7
View File
@@ -741,13 +741,7 @@ mod tests {
// 200 `m:t` nested in each other around one word: the word, not 200 copies of it.
let x = format!(r#"<a14:m xmlns:a14="a" xmlns:m="m"><m:oMath>{}word{}</m:oMath></a14:m>"#, "<m:t>".repeat(200), "</m:t>".repeat(200));
let d = parse(x.as_bytes()).unwrap();
let part = Part {
name: "ppt/slides/slide1.xml".into(),
rels: Default::default(),
ns: vec![],
inner_ns: vec![],
ns_budget: std::cell::Cell::new(1 << 20),
};
let part = Part { name: "ppt/slides/slide1.xml".into(), rels: Default::default(), ns: vec![], inner_ns: vec![] };
let mut para = Paragraph::default();
push_math_run(&mut para, &part, &d.root);
assert_eq!(para.runs[0].text, "word");
+3
View File
@@ -785,6 +785,9 @@ fn table(imp: &mut Imp, part: &Part, tbl: &El) -> Table {
let n = t.rows.iter().map(|r| r.cells.len()).max().unwrap_or(0);
t.cols = vec![72.0; n];
}
if t.over_cap() {
imp.warn(format!("a table in {} is over {} cells; its last rows were left out", part.name, deckcraft_model::table::MAX_CELLS));
}
t.normalize();
t
}
+9
View File
@@ -65,6 +65,8 @@ pub struct Exp<'a> {
n_kept: u32,
pub slide_parts: HashMap<SlideId, String>,
pub table_styles: Vec<String>,
/// A table past the cell cap was met; export fails rather than write a trimmed one.
pub table_over_cap: bool,
/// The ids in `table_styles`, for checking membership without a scan.
pub table_style_set: std::collections::HashSet<String>,
placeholder: Option<String>,
@@ -558,6 +560,7 @@ pub fn export(p: &Presentation, main_ct: &str) -> Result<Vec<u8>, PptxError> {
n_kept: 0,
slide_parts: HashMap::new(),
table_styles: vec![],
table_over_cap: false,
table_style_set: Default::default(),
placeholder: None,
};
@@ -766,6 +769,12 @@ pub fn export(p: &Presentation, main_ct: &str) -> Result<Vec<u8>, PptxError> {
root.add_new(&k.rel_type, &k.name, false);
}
x.pkg.rels("", &root);
if x.table_over_cap {
return Err(PptxError::Unsupported(format!(
"a table has more than {} cells; export refused rather than drop rows",
deckcraft_model::table::MAX_CELLS
)));
}
x.pkg.finish()
}
+3 -79
View File
@@ -132,20 +132,7 @@ fn title(w: &mut W, text: &str, c: &Chart) {
w.close("c:title");
}
/// Points one cache writes. Excel and PowerPoint hold 32,000 points per series, so this is far
/// above any chart they make; a longer vector is cut here with a warning, not refused.
const MAX_CACHE_POINTS: usize = 1_000_000;
/// `vals` cut to [`MAX_CACHE_POINTS`].
fn capped<T>(vals: &[T]) -> &[T] {
if vals.len() > MAX_CACHE_POINTS {
log::warn!("pptx: chart data cache cut from {} to {MAX_CACHE_POINTS} points", vals.len());
}
&vals[..vals.len().min(MAX_CACHE_POINTS)]
}
fn str_cache(w: &mut W, f: &str, vals: &[String]) {
let vals = capped(vals);
w.open0("c:strRef");
w.elt("c:f", f);
w.open0("c:strCache");
@@ -160,7 +147,6 @@ fn str_cache(w: &mut W, f: &str, vals: &[String]) {
}
fn num_cache(w: &mut W, f: &str, vals: &[Option<f64>]) {
let vals = capped(vals);
w.open0("c:numRef");
w.elt("c:f", f);
w.open0("c:numCache");
@@ -179,7 +165,6 @@ fn num_cache(w: &mut W, f: &str, vals: &[Option<f64>]) {
/// Numbers written into the chart itself (`c:numLit`), with no worksheet range behind them.
fn num_lit(w: &mut W, vals: &[Option<f64>]) {
let vals = capped(vals);
w.open0("c:numLit");
w.val("c:ptCount", vals.len());
for (i, v) in vals.iter().enumerate() {
@@ -261,9 +246,6 @@ fn point_fills(w: &mut W, s: &Series, g: Group) {
}
}
/// Data points (categories and values counted per series) one chart group writes to its caches.
const MAX_EXPORT_POINTS: usize = 4_000_000;
fn series(w: &mut W, c: &Chart, s: &Series, i: usize, kind: ChartType, g: Group) {
w.open0("c:ser");
w.val("c:idx", i);
@@ -323,7 +305,7 @@ fn series(w: &mut W, c: &Chart, s: &Series, i: usize, kind: ChartType, g: Group)
trendline(w, t);
}
}
let n = c.categories.len().max(s.values.len()).min(MAX_CACHE_POINTS);
let n = c.categories.len().max(s.values.len());
let last = n + 1;
if let Some(e) = &s.error_bars {
error_bars(w, e, g);
@@ -509,24 +491,6 @@ fn axes(w: &mut W, c: &Chart, g: Group) {
w.close("c:valAx");
}
/// Points `series` writes into its caches for group `g`, synthesized vectors included.
fn series_points(c: &Chart, s: &Series, g: Group) -> usize {
let v = s.values.len();
if matches!(g, Group::Scatter | Group::Bubble) {
let xs = if s.x_values.is_empty() { v } else { s.x_values.len() };
let sizes = if g != Group::Bubble {
0
} else if s.sizes.is_empty() {
v
} else {
s.sizes.len()
};
xs.saturating_add(v).saturating_add(sizes)
} else {
c.categories.len().saturating_add(v)
}
}
fn plot_group(w: &mut W, c: &Chart, kind: ChartType, list: &[(usize, &Series)]) {
let g = group_of(kind);
let grouping = if kind.percent() {
@@ -565,17 +529,7 @@ fn plot_group(w: &mut W, c: &Chart, kind: ChartType, list: &[(usize, &Series)])
if g != Group::Stock {
w.val("c:varyColors", vary);
}
// Each series repeats the categories in its own cache, so the output grows with series times
// categories: stop adding series once a group's caches pass the budget (the first is always written).
// The first series is always written; each of its caches is cut at MAX_CACHE_POINTS.
let mut left = MAX_EXPORT_POINTS;
for (n, (i, s)) in list.iter().enumerate() {
let cost = series_points(c, s, g);
if n > 0 && cost > left {
log::warn!("pptx: chart export stopped after {n} series; the data caches passed {MAX_EXPORT_POINTS} points");
break;
}
left = left.saturating_sub(cost);
for (i, s) in list {
series(w, c, s, *i, kind, g);
}
if c.data_labels {
@@ -1107,21 +1061,7 @@ mod region_map_tests {
mod tests {
use deckcraft_color::{ColorTransform, Rgba};
use deckcraft_model::ColorRef;
use deckcraft_model::chart::{Chart, ChartType, Series};
#[test]
fn export_cost_counts_the_caches_each_group_writes() {
let s = Series { values: vec![Some(1.0); 600], ..Default::default() };
let mut c = Chart::sample(ChartType::Column);
c.categories = vec!["a".into(); 1000];
// Category and value caches: 1,000 + 600, however the series is stored.
assert_eq!(super::series_points(&c, &s, super::Group::Bar), 1600);
// A bubble series without x values or sizes still writes three value-length caches.
assert_eq!(super::series_points(&c, &s, super::Group::Bubble), 1800);
assert_eq!(super::series_points(&c, &s, super::Group::Scatter), 1200);
let s = Series { values: vec![Some(1.0); 600], x_values: vec![Some(1.0); 5], sizes: vec![Some(1.0); 7], ..Default::default() };
assert_eq!(super::series_points(&c, &s, super::Group::Bubble), 612);
}
use deckcraft_model::chart::{Chart, ChartType};
/// The chart's text colour is written where PowerPoint keeps it (`c:txPr` on the axes, legend,
/// data labels and chart, the title's run defaults) and read back as the same colour.
@@ -1182,19 +1122,3 @@ mod tests {
assert!(!crate::read::chart::read_chart_xml(&super::chart_xml(&c).0).unwrap().cat_reversed);
}
}
#[cfg(test)]
mod cache_cap_tests {
use super::*;
/// A cache longer than the cap is cut, never refused; anything a real chart holds (32,000
/// points per series) is far under it. Zero-sized elements, so the test allocates nothing.
#[test]
fn caches_are_cut_at_the_cap_not_refused() {
const { assert!(MAX_CACHE_POINTS >= 32_000) };
let over = vec![(); MAX_CACHE_POINTS + 1];
assert_eq!(capped(&over).len(), MAX_CACHE_POINTS);
let real = vec![(); 32_000];
assert_eq!(capped(&real).len(), 32_000);
}
}
+2 -2
View File
@@ -721,8 +721,8 @@ fn frame_close(w: &mut W) {
}
fn table(w: &mut W, x: &mut Exp, o: &mut Out, t: &Table) {
let mut t = t.clone();
t.normalize();
x.table_over_cap |= t.over_cap();
let t = t.normalized();
w.open0("a:tbl");
let (guid, _) = crate::tables::guid_for_style(&t.style);
if x.table_style_set.insert(t.style.clone()) {
+38
View File
@@ -288,6 +288,44 @@ fn unreadable_slide_is_reported() {
assert_eq!(deckcraft_pptx::import(&broken).expect("import").slides.len(), 1);
}
fn table_deck(rows: usize, cols: usize) -> deckcraft_model::Presentation {
use deckcraft_model::table::Row;
use deckcraft_model::{Shape, ShapeId, ShapeKind, Table, Xfrm};
let mut p = defaults::new_presentation(None);
let id = ShapeId(p.alloc_id().expect("ids"));
let t = Table { cols: vec![1.0; cols], rows: vec![Row { height: 1.0, cells: vec![Default::default(); cols] }; rows], ..Default::default() };
let slide = std::sync::Arc::make_mut(&mut p.slides[0]);
slide.shapes.push(Shape { id, name: "T".into(), xfrm: Some(Xfrm::new(0.0, 0.0, 10.0, 10.0)), kind: ShapeKind::Table(t), ..Default::default() });
p
}
/// A table past the cell cap opens with its leading rows and an import warning; saving one that
/// is over the cap is an error, never a silent trim.
#[test]
fn oversized_table_warns_on_import_and_refuses_export() {
let max = deckcraft_model::table::MAX_CELLS;
let cols = 1000;
let rows = max / cols;
// A table exactly at the cap round-trips with no warning.
let bytes = deckcraft_pptx::export(&table_deck(rows, cols)).expect("export at the cap");
let (_, _, warnings) = deckcraft_pptx::import_report(&bytes).expect("import");
assert!(warnings.is_empty(), "{warnings:?}");
// One more row in the file: kept truncated, and reported.
let mut es = entries(&bytes);
let slide = es.iter_mut().find(|(n, _)| n == "ppt/slides/slide1.xml").expect("slide1");
let xml = String::from_utf8(std::mem::take(&mut slide.1)).expect("utf8");
let start = xml.rfind("<a:tr ").expect("row");
let end = xml.rfind("</a:tr>").expect("row end") + "</a:tr>".len();
slide.1 = format!("{}{}{}", &xml[..end], &xml[start..end], &xml[end..]).into_bytes();
let (q, _, warnings) = deckcraft_pptx::import_report(&rezip(&es)).expect("import");
assert_eq!(warnings.len(), 1, "{warnings:?}");
let t = q.slides[0].shapes.iter().find_map(|s| if let deckcraft_model::ShapeKind::Table(t) = &s.kind { Some(t) } else { None }).expect("table");
assert_eq!((t.rows.len(), t.cols.len()), (rows, cols));
// A model table past the cap is refused on export.
let err = deckcraft_pptx::export(&table_deck(rows + 1, cols)).expect_err("over the cap");
assert!(err.to_string().contains("table"), "{err}");
}
/// Many slides naming one notes part: it is read for the first slide only, not copied onto every
/// slide that names it.
#[test]
+1 -7
View File
@@ -166,17 +166,11 @@ impl Decoder {
if frames.len() > 1 {
self.stats.superframes += 1;
}
let queued = self.pending.len();
for f in frames {
if f.is_empty() {
continue;
}
if let Err(e) = self.decode_frame(f, pts) {
// A failed sample outputs nothing: drop what its earlier subframes queued so a
// later drain does not expand each repeat into its own picture.
self.pending.truncate(queued);
return Err(e);
}
self.decode_frame(f, pts)?;
}
Ok(self.drain(false))
}