aleleba/pdfcraft
1
0
Fork 0
mirror of https://github.com/storytold/pdfcraft.git synced 2026-10-24 10:08:00 -06:00
Code Issues Packages Projects Releases Wiki Activity

fix(render): keep embedded-file keys within the text budget

Browse Source 
The embedded-files name tree copied every key, and an attachment
without a file name copied its key again as the name, so many entries
sharing one long key string multiplied it. Keys, and a key kept again
as the name, now spend the 16 MiB link and attachment text budget; an
embedded file whose key does not fit is not listed.
This commit is contained in:
aiongg
2026-10-22 22:03:43 +00:00
parent a0a66b982b
commit 7a73c2ec3a
1 changed files with 47 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files
crates/render/src/inspect.rs
+47 -10
View File
@@ -42,10 +42,11 @@ const MAX_REMOTE_DESTINATION_NAME_BYTES: usize = 1024;
/// are not listed: real comments stay far below, many small ones sharing one long string do not.
const MAX_ANNOTATION_BYTES: usize = 64 << 20;
/// Web-address text the link listing, and name and description text the attachment listing, keep
/// together. Many links or attachments can share one long string, which each would otherwise copy.
/// Past it, a link to a web address lists as a URI action without its address, and an attachment
/// under its generic name without a description.
/// Web-address text the link listing, and name, description and embedded-file key text the
/// attachment listing, keep together. Many links or attachments can share one long string, which
/// each would otherwise copy. Past it, a link to a web address lists as a URI action without its
/// address, an attachment under its generic name without a description, and an embedded file whose
/// key does not fit is not listed.
const MAX_LINK_ATTACHMENT_TEXT_BYTES: usize = 16 << 20;
/// The bytes a listed annotation keeps.
@@ -745,8 +746,9 @@ impl<'a> Inspector<'a> {
if let Some(names) = catalog.get(b"Names").ok().and_then(|o| self.dict(o)) {
if let Some(ef) = names.get(b"EmbeddedFiles").ok().and_then(|o| self.dict(o)) {
let mut seen = HashSet::new();
for (name, spec) in self.name_tree(ef, &mut seen, 0) {
let fallback = |_: &mut usize| Some(name.clone());
for (name, spec) in self.name_tree(ef, &mut seen, 0, &mut text_budget) {
// The key is kept a second time when it is also the name.
let fallback = |budget: &mut usize| spend(name.as_bytes(), budget, |_| name.clone());
info.attachments.push(self.attachment(spec, AttachmentSource::Document { key: name.clone() }, &mut text_budget, fallback));
}
}
@@ -955,7 +957,15 @@ impl<'a> Inspector<'a> {
}
}
fn name_tree(&self, node: &'a Dictionary, seen: &mut HashSet<*const Dictionary>, depth: u32) -> Vec<(String, &'a Object)> {
/// The entries of an embedded-files name tree. Each string key is spent from `text_budget`
/// (see [`MAX_LINK_ATTACHMENT_TEXT_BYTES`]), and an entry whose key does not fit is left out.
fn name_tree(
&self,
node: &'a Dictionary,
seen: &mut HashSet<*const Dictionary>,
depth: u32,
text_budget: &mut usize,
) -> Vec<(String, &'a Object)> {
let mut out = Vec::new();
if depth > 32 || !seen.insert(node as *const _) {
return out;
@@ -964,7 +974,10 @@ impl<'a> Inspector<'a> {
for pair in pairs.chunks(2) {
if let [k, v] = pair {
let key = match self.resolve(k) {
Object::String(bytes, _) => text_string(bytes),
Object::String(bytes, _) => match spend(bytes, text_budget, text_string) {
Some(key) => key,
None => continue,
},
_ => String::new(),
};
out.push((key, v));
@@ -974,7 +987,7 @@ impl<'a> Inspector<'a> {
if let Ok(Object::Array(kids)) = node.get(b"Kids").map(|o| self.resolve(o)) {
for k in kids {
if let Some(d) = self.dict(k) {
out.extend(self.name_tree(d, seen, depth + 1));
out.extend(self.name_tree(d, seen, depth + 1, text_budget));
}
}
}
@@ -1790,7 +1803,8 @@ pub fn attachment_data(bytes: &[u8], password: Option<&str>, att: &Attachment) -
let names = doc.catalog().ok().and_then(|c| c.get(b"Names").ok()).and_then(|o| insp.dict(o)).ok_or("no /Names")?;
let ef = names.get(b"EmbeddedFiles").ok().and_then(|o| insp.dict(o)).ok_or("no /EmbeddedFiles")?;
let mut seen = HashSet::new();
insp.name_tree(ef, &mut seen, 0).into_iter().find(|(k, _)| k == key).map(|(_, v)| v).ok_or("attachment not found")?
let mut budget = MAX_LINK_ATTACHMENT_TEXT_BYTES;
insp.name_tree(ef, &mut seen, 0, &mut budget).into_iter().find(|(k, _)| k == key).map(|(_, v)| v).ok_or("attachment not found")?
}
AttachmentSource::Annotation { page, index } => {
let (&pid, _) = insp.page_index.iter().find(|(_, i)| **i == *page).ok_or("page not found")?;
@@ -2797,6 +2811,29 @@ trailer << /Root 1 0 R >>
assert!(matches!(&info.links[0].target, LinkTarget::Uri(u) if *u == uri));
}
/// Embedded files whose name-tree keys share one long string keep it, as key and as name, only
/// within `MAX_LINK_ATTACHMENT_TEXT_BYTES`.
#[test]
fn embedded_file_keys_sharing_a_long_string_keep_it_within_a_budget() {
let entries = "4 0 R << /Type /Filespec >> ".repeat(20);
let pdf = format!(
"%PDF-1.7\n1 0 obj << /Type /Catalog /Pages 2 0 R /Names << /EmbeddedFiles << /Names [{entries}] >> >> >> endobj\n\
2 0 obj << /Type /Pages /Kids [] /Count 0 >> endobj\n\
4 0 obj ({}) endobj\n\
trailer << /Root 1 0 R >>\n%%EOF",
"x".repeat(1 << 20),
);
let doc = Document::load_mem(pdf.as_bytes()).expect("loads");
let mut info = DocInfo::default();
Inspector::new(&doc).fill(&mut info);
let kept: usize =
info.attachments.iter().map(|a| a.name.len() + if let AttachmentSource::Document { key } = &a.source { key.len() } else { 0 }).sum();
assert!(!info.attachments.is_empty());
// Past the budget an attachment still has the generic name.
let generic = info.attachments.len() * "attachment".len();
assert!(kept <= MAX_LINK_ATTACHMENT_TEXT_BYTES + generic, "{kept} bytes kept");
}
/// Links and attachments sharing one long string each keep a copy only within
/// `MAX_LINK_ATTACHMENT_TEXT_BYTES`, and every one is still listed.
#[test]