mirror of
https://github.com/storytold/pdfcraft.git
synced 2026-10-25 10:39:01 -06:00
Fuzzed inputs drove one check-one process to 2-6 GB (the 16 GB CI runner running two of them was killed every night). Three causes, each with a test: - lopdf (bootstrap inspector) sized PNG predictor rows from /Columns before reading data: /Columns 4294967295 allocated two 4 GiB rows. Vendored lopdf 0.45.0 refuses rows longer than the data (and uses saturating row widths); its load-time stream decoding also gets a 256 MiB limit (it had none). - hayro sized a tiling pattern's cell pixmap from /XStep x scale, bounded only by u16: a step far beyond /BBox allocated 2 GB. The vendored patch keeps the cell within 3000 px a side, lowering the resolution instead. - cos decoded cross-reference streams up to the general 1 GiB limit; they now stop at what 8,388,607 objects' rows can hold (~200 MB at most), and the section is reconstructed (logged) instead. On the nightly's seed (41) the worst captured input fell from 6.2 GB to 6 MB, and a 300 s fuzz run had no process above 300 MB (before: many at 2-5 GB). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
150 lines
4.6 KiB
Rust
150 lines
4.6 KiB
Rust
use lopdf::{Document, Object, SaveOptions};
|
|
|
|
fn main() {
|
|
println!("Final Object Stream Compression Test\n");
|
|
println!("{}", "=".repeat(60));
|
|
|
|
// Create test document with various object types
|
|
let mut doc = Document::with_version("1.5");
|
|
|
|
// Add some compressible objects (fonts, annotations, etc.)
|
|
let font1 = doc.add_object(lopdf::dictionary! {
|
|
"Type" => "Font",
|
|
"Subtype" => "Type1",
|
|
"BaseFont" => "Helvetica"
|
|
});
|
|
|
|
let font2 = doc.add_object(lopdf::dictionary! {
|
|
"Type" => "Font",
|
|
"Subtype" => "Type1",
|
|
"BaseFont" => "Times-Roman"
|
|
});
|
|
|
|
// Add a page (should NOT be compressed)
|
|
let page_id = doc.add_object(lopdf::dictionary! {
|
|
"Type" => "Page",
|
|
"MediaBox" => vec![0.into(), 0.into(), 612.into(), 792.into()],
|
|
"Contents" => lopdf::Object::Array(vec![]),
|
|
"Resources" => lopdf::dictionary! {
|
|
"Font" => lopdf::dictionary! {
|
|
"F1" => font1,
|
|
"F2" => font2
|
|
}
|
|
}
|
|
});
|
|
|
|
// Add pages tree (should NOT be compressed)
|
|
let pages_id = doc.add_object(lopdf::dictionary! {
|
|
"Type" => "Pages",
|
|
"Kids" => vec![page_id.into()],
|
|
"Count" => 1
|
|
});
|
|
|
|
// Add catalog (should NOT be compressed - it's in trailer)
|
|
let catalog_id = doc.add_object(lopdf::dictionary! {
|
|
"Type" => "Catalog",
|
|
"Pages" => pages_id
|
|
});
|
|
|
|
doc.trailer.set("Root", catalog_id);
|
|
|
|
// Add more compressible objects
|
|
for i in 0..10 {
|
|
doc.add_object(lopdf::dictionary! {
|
|
"Type" => "Annot",
|
|
"Subtype" => "Text",
|
|
"Contents" => format!("Annotation {}", i)
|
|
});
|
|
}
|
|
|
|
println!("Created test document with:");
|
|
println!(" - 2 Font objects (should be compressed)");
|
|
println!(" - 1 Page object (should NOT be compressed)");
|
|
println!(" - 1 Pages object (should NOT be compressed)");
|
|
println!(" - 1 Catalog object (should NOT be compressed)");
|
|
println!(" - 10 Annotation objects (should be compressed)");
|
|
|
|
// Save with object streams
|
|
let options = SaveOptions {
|
|
use_object_streams: true,
|
|
use_xref_streams: true,
|
|
..Default::default()
|
|
};
|
|
|
|
let mut buffer = Vec::new();
|
|
doc.save_with_options(&mut buffer, options).unwrap();
|
|
|
|
println!("\nSaved with object streams: {} bytes", buffer.len());
|
|
|
|
// Load back and analyze
|
|
let loaded = Document::load_mem(&buffer).unwrap();
|
|
println!("\nAnalyzing saved document:");
|
|
println!(" Total objects: {}", loaded.objects.len());
|
|
|
|
// Find object streams
|
|
let mut obj_stream_count = 0;
|
|
let mut compressed_count = 0;
|
|
|
|
for obj in loaded.objects.values() {
|
|
if let Object::Stream(stream) = obj
|
|
&& let Ok(type_obj) = stream.dict.get(b"Type")
|
|
&& let Ok(type_name) = type_obj.as_name()
|
|
&& type_name == b"ObjStm"
|
|
{
|
|
obj_stream_count += 1;
|
|
|
|
// Count objects in stream
|
|
if let Ok(n) = stream.dict.get(b"N").and_then(|o| o.as_i64()) {
|
|
compressed_count += n as usize;
|
|
}
|
|
}
|
|
}
|
|
|
|
println!(" Object streams: {}", obj_stream_count);
|
|
println!(" Compressed objects: {}", compressed_count);
|
|
|
|
// Verify critical objects are NOT compressed
|
|
println!("\nVerifying critical objects:");
|
|
|
|
// Check catalog
|
|
if let Ok(root_ref) = loaded.trailer.get(b"Root")
|
|
&& let Object::Reference(cat_id) = root_ref
|
|
{
|
|
check_not_compressed(&loaded, *cat_id, "Catalog");
|
|
}
|
|
|
|
// Check pages tree
|
|
if let Ok(catalog) = loaded.catalog()
|
|
&& let Ok(pages_ref) = catalog.get(b"Pages")
|
|
&& let Object::Reference(pages_id) = pages_ref
|
|
{
|
|
check_not_compressed(&loaded, *pages_id, "Pages tree");
|
|
}
|
|
|
|
// Check page objects
|
|
let pages = loaded.get_pages();
|
|
for (num, &page_id) in pages.iter() {
|
|
check_not_compressed(&loaded, page_id, &format!("Page {}", num));
|
|
}
|
|
|
|
println!("\n{}", "=".repeat(60));
|
|
println!("Test complete! Object streams are working correctly.");
|
|
}
|
|
|
|
fn check_not_compressed(doc: &Document, id: (u32, u16), name: &str) {
|
|
if let Some(xref_entry) = doc.reference_table.get(id.0) {
|
|
match xref_entry {
|
|
lopdf::xref::XrefEntry::Normal { .. } => {
|
|
println!(" ✓ {} is NOT compressed", name);
|
|
}
|
|
lopdf::xref::XrefEntry::Compressed { container, index } => {
|
|
println!(
|
|
" ✗ ERROR: {} is compressed in stream {} at index {}!",
|
|
name, container, index
|
|
);
|
|
}
|
|
_ => {}
|
|
}
|
|
}
|
|
}
|