aleleba/pdfcraft
1
0
Fork 0
mirror of https://github.com/storytold/pdfcraft.git synced 2026-10-27 11:37:59 -06:00
Code Issues Packages Projects Releases Wiki Activity

Merge PR #3259: fix(content): accept inline image EI without preceding whitespace (#3108)

Browse Source
This commit is contained in:
aiongg
2026-10-22 01:44:52 +00:00
parent ffaf312358 9451e0652c
commit da90eb93fb
2 changed files with 72 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files
crates/content/src/lib.rs
+27 -7
View File
@@ -219,19 +219,38 @@ fn inline_image(data: &[u8], pos: usize) -> Result<(Dict, Vec<u8>, usize), usize
return Ok((dict, Vec::new(), data.len()));
}
// Without an `EI` after it, the image runs to the end of the stream.
return Err(end_of_image(data, at).map_or(data.len(), |(_, end)| end));
return Err(end_of_image(data, at, b"").map_or(data.len(), |(_, end)| end));
};
dict.set(k, v);
}
// One whitespace byte separates ID from the data.
let begin = (lx.pos + 1).min(data.len());
let end = unfiltered_image_end(data, begin, &dict).or_else(|| end_of_image(data, begin));
let end = unfiltered_image_end(data, begin, &dict).or_else(|| end_of_image(data, begin, filter_eod(&dict)));
Ok(match end {
Some((end_data, end)) => (dict, data[begin..end_data].to_vec(), end),
None => (dict, data[begin..].to_vec(), data.len()),
})
}
/// The end-of-data marker (`>` for ASCIIHex, `~>` for ASCII85) of the outer stream filter, if any.
fn filter_eod(dict: &Dict) -> &'static [u8] {
for key in [b"Filter".as_slice(), b"F"] {
if let Some(filter) = dict.get(key) {
let name = match filter {
Object::Name(name) => Some(name.as_slice()),
Object::Array(filters) => filters.first().and_then(Object::as_name),
_ => None,
};
return match name {
Some(b"ASCIIHexDecode" | b"AHx") => b">",
Some(b"ASCII85Decode" | b"A85") => b"~>",
_ => b"",
};
}
}
b""
}
/// An unfiltered image's samples determine its length, including byte padding per row.
/// Binary samples may themselves contain whitespace + `EI` + whitespace, so use the known
/// length when the actual terminator follows it. Unknown or damaged layouts keep the heuristic.
@@ -293,23 +312,24 @@ fn unfiltered_image_end(data: &[u8], begin: usize, dict: &Dict) -> Option<(usize
let row = width.checked_mul(components)?.checked_mul(bits)?.div_ceil(8);
let end_data = begin.checked_add(row.checked_mul(height)?)?;
let after = data.get(end_data..)?;
if !after.first().is_some_and(|b| is_ws(*b)) {
let separator = after.iter().position(|b| !is_ws(*b))?;
if separator == 0 && data.get(end_data.checked_sub(1)?).is_some_and(|b| is_ws(*b)) {
return None;
}
let separator = after.iter().position(|b| !is_ws(*b))?;
let ei = end_data.checked_add(separator)?;
let end = ei.checked_add(2)?;
if data.get(ei..end) == Some(b"EI") && data.get(end).is_none_or(|b| is_ws(*b) || is_delim(*b)) { Some((end_data, end)) } else { None }
}
/// Where inline image data starting at `begin` ends: (end of the data, position after `EI`). The
/// data ends at whitespace + `EI` + (whitespace, a delimiter or the end).
fn end_of_image(data: &[u8], begin: usize) -> Option<(usize, usize)> {
/// data ends at whitespace (or `eod`: `>` for ASCIIHex, `~>` for ASCII85) + `EI` + (whitespace,
/// a delimiter or the end).
fn end_of_image(data: &[u8], begin: usize, eod: &[u8]) -> Option<(usize, usize)> {
let mut i = begin;
while i + 2 <= data.len() {
if data[i] == b'E'
&& data.get(i + 1) == Some(&b'I')
&& (i == begin || is_ws(data[i - 1]))
&& (i == begin || is_ws(data[i - 1]) || (!eod.is_empty() && data[begin..i].ends_with(eod)))
&& data.get(i + 2).is_none_or(|b| is_ws(*b) || is_delim(*b))
{
let end_data = if i > begin && is_ws(data[i - 1]) { i - 1 } else { i };
crates/content/src/tests.rs
+45
View File
@@ -388,3 +388,48 @@ fn piece_lookup_preserves_exact_boundary_and_empty_piece_mapping() {
assert_eq!(empty.piece_of(0), 0);
assert_eq!(empty.piece_of(usize::MAX), 0);
}
#[test]
fn inline_image_terminator_without_preceding_whitespace() {
let cases: &[(&[u8], &[u8])] = &[
(b"BI /W 4 /H 1 /BPC 8 /CS /G /F /AHx ID 2060A0E0>EI Q BT /F1 18 Tf (After) Tj ET", b"2060A0E0>"),
(b"BI /W 4 /H 1 /BPC 8 /CS /G /Filter /ASCIIHexDecode ID 2060A0E0>EI Q BT /F1 18 Tf (After) Tj ET", b"2060A0E0>"),
(b"BI /W 4 /H 1 /BPC 8 /CS /G /F [/AHx] ID 2060A0E0>EI Q BT /F1 18 Tf (After) Tj ET", b"2060A0E0>"),
(b"BI /W 4 /H 1 /BPC 8 /CS /G /F [/AHx /RL] ID 2060A0E0>EI Q BT /F1 18 Tf (After) Tj ET", b"2060A0E0>"),
(b"BI /W 4 /H 1 /BPC 8 /CS /G /F /A85 ID 5s,^g~>EI Q BT /F1 18 Tf (After) Tj ET", b"5s,^g~>"),
(b"BI /W 4 /H 1 /BPC 8 /CS /G /Filter /ASCII85Decode ID 5s,^g~>EI Q BT /F1 18 Tf (After) Tj ET", b"5s,^g~>"),
(b"BI /W 4 /H 1 /BPC 8 /CS /G /F [/A85 /RL] ID 5s,^g~>EI Q BT /F1 18 Tf (After) Tj ET", b"5s,^g~>"),
(b"BI /W 4 /H 1 /BPC 8 /CS /G ID ab()EI Q BT /F1 18 Tf (After) Tj ET", b"ab()"),
(b"BI /W 4 /H 1 /BPC 8 /CS /G ID ab[]EI Q BT /F1 18 Tf (After) Tj ET", b"ab[]"),
(b"BI /W 4 /H 1 /BPC 8 /CS /G ID \x20\x60\xa0\xe0EI Q BT /F1 18 Tf (After) Tj ET", b"\x20\x60\xa0\xe0"),
];
for (src, expected_data) in cases {
let p = parse(src);
assert_eq!(p.skipped, 0, "stream should have 0 skipped bytes: {}", String::from_utf8_lossy(src));
let names: Vec<&[u8]> = p.ops.iter().map(|o| o.op.as_slice()).collect();
assert_eq!(
names,
[&b"BI"[..], b"Q", b"BT", b"Tf", b"Tj", b"ET"],
"operators after inline image must be preserved: {}",
String::from_utf8_lossy(src)
);
let (_, img) = p.ops[0].inline.as_ref().unwrap();
assert_eq!(img, expected_data, "inline image data must match: {}", String::from_utf8_lossy(src));
}
}
#[test]
fn ascii85_inline_image_does_not_terminate_at_gt_ei_inside_payload() {
for src in [
b"BI /W 4 /H 1 /BPC 8 /CS /G /F /A85 ID !!!!>EI !!~>\nEI Q BT /F1 18 Tf (After) Tj ET".as_slice(),
b"BI /W 4 /H 1 /BPC 8 /CS /G /F [/A85 /AHx] ID !!!!>EI !!~>\nEI Q BT /F1 18 Tf (After) Tj ET".as_slice(),
] {
let p = parse(src);
assert_eq!(p.skipped, 0);
let names: Vec<&[u8]> = p.ops.iter().map(|o| o.op.as_slice()).collect();
assert_eq!(names, [&b"BI"[..], b"Q", b"BT", b"Tf", b"Tj", b"ET"]);
let (_, img) = p.ops[0].inline.as_ref().unwrap();
assert_eq!(img, b"!!!!>EI !!~>");
}
}