WIP

Cargo.lock

@@ -5,71 +5,3 @@ version = 4
 [[package]]
 name = "parse_powerisa_pdf"
 version = "0.1.0"
-dependencies = [
- "serde",
-]
-
-[[package]]
-name = "proc-macro2"
-version = "1.0.103"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5ee95bc4ef87b8d5ba32e8b7714ccc834865276eab0aed5c9958d00ec45f49e8"
-dependencies = [
- "unicode-ident",
-]
-
-[[package]]
-name = "quote"
-version = "1.0.42"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a338cc41d27e6cc6dce6cefc13a0729dfbb81c262b1f519331575dd80ef3067f"
-dependencies = [
- "proc-macro2",
-]
-
-[[package]]
-name = "serde"
-version = "1.0.228"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9a8e94ea7f378bd32cbbd37198a4a91436180c5bb472411e48b5ec2e2124ae9e"
-dependencies = [
- "serde_core",
- "serde_derive",
-]
-
-[[package]]
-name = "serde_core"
-version = "1.0.228"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "41d385c7d4ca58e59fc732af25c3983b67ac852c1a25000afe1175de458b67ad"
-dependencies = [
- "serde_derive",
-]
-
-[[package]]
-name = "serde_derive"
-version = "1.0.228"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d540f220d3187173da220f885ab66608367b6574e925011a9353e4badda91d79"
-dependencies = [
- "proc-macro2",
- "quote",
- "syn",
-]
-
-[[package]]
-name = "syn"
-version = "2.0.111"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "390cc9a294ab71bdb1aa2e99d13be9c753cd2d7bd6560c77118597410c4d2e87"
-dependencies = [
- "proc-macro2",
- "quote",
- "unicode-ident",
-]
-
-[[package]]
-name = "unicode-ident"
-version = "1.0.22"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9312f7c4f6ff9069b165498234ce8be658059c6728633667c526e27dc2cf1df5"

Cargo.toml

@@ -5,5 +5,4 @@ edition = "2024"
 license = "LGPL-3.0-or-later"
 
 [dependencies]
-serde = { version = "1.0.228", features = ["derive"] }
 

src/lib.rs

@@ -1,2 +1,5 @@
-mod pdf;
+#[doc(hidden)]
-mod util;
+pub use std as __std;
+
+pub mod pdf;
+pub mod util;

src/main.rs

@@ -0,0 +1,45 @@
+use parse_powerisa_pdf::pdf::{Pdf, PdfTrailer};
+use std::{
+    error::Error,
+    io::{IsTerminal, Read},
+    process::ExitCode,
+};
+
+fn main() -> Result<ExitCode, Box<dyn Error>> {
+    let args: Vec<_> = std::env::args_os().collect();
+    if args
+        .iter()
+        .skip(1)
+        .any(|v| v.as_encoded_bytes().starts_with(b"-") && v != "-")
+        || args.len() > 2
+        || (args.len() == 1 && std::io::stdin().is_terminal())
+    {
+        eprintln!(
+            "Usage: {} [<path/to/file.pdf>]\n\
+            Reads the PDF file passed on the command line,\n\
+            Reads stdin if no arguments are passed or if the file name is just a dash `-`.\n\
+            If stdin is a terminal, you have to pass `-` explicitly to read from it.",
+            args[0].display()
+        );
+        return Ok(ExitCode::FAILURE);
+    }
+    let file_path = args.get(1).filter(|v| *v != "-");
+    let input = if let Some(file_path) = file_path {
+        std::fs::read(file_path)?
+    } else {
+        let mut buf = Vec::new();
+        std::io::stdin().lock().read_to_end(&mut buf)?;
+        buf
+    };
+    let pdf = Pdf::parse(input)?;
+    if let PdfTrailer::Stream {
+        xref_stream,
+        start_xref,
+    } = pdf.trailer
+    {
+        dbg!(xref_stream.dictionary());
+    }
+
+    todo!();
+    Ok(ExitCode::SUCCESS)
+}

src/pdf/parse.rs

@@ -0,0 +1,953 @@
+use crate::pdf::object::{
+    MaybeArray, PdfInteger, PdfName, PdfNull, PdfObject, PdfObjectDirect, PdfObjectIdentifier,
+    PdfObjectIndirect, PdfObjectNonNull, PdfReal,
+};
+use std::{any::Any, borrow::Cow, fmt, mem, num::NonZero, sync::Arc};
+
+#[derive(Clone, Copy, PartialEq, Eq, Hash, PartialOrd, Ord, Default)]
+pub struct PdfInputPosition(Option<usize>);
+
+impl fmt::Debug for PdfInputPosition {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        f.debug_tuple("PdfInputPosition")
+            .field(&format_args!("{self}"))
+            .finish()
+    }
+}
+
+impl fmt::Display for PdfInputPosition {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        if let Some(pos) = self.0 {
+            write!(f, "{pos:#x}")
+        } else {
+            f.write_str("<unknown>")
+        }
+    }
+}
+
+impl PdfInputPosition {
+    pub const fn new(pos: usize) -> Self {
+        Self(Some(pos))
+    }
+    pub const fn empty() -> PdfInputPosition {
+        Self(None)
+    }
+}
+
+pub trait GetPdfInputPosition {
+    fn get_pdf_input_position(&self) -> PdfInputPosition;
+}
+
+impl<T: ?Sized + GetPdfInputPosition> GetPdfInputPosition for &'_ T {
+    fn get_pdf_input_position(&self) -> PdfInputPosition {
+        T::get_pdf_input_position(self)
+    }
+}
+
+impl<T: ?Sized + GetPdfInputPosition> GetPdfInputPosition for &'_ mut T {
+    fn get_pdf_input_position(&self) -> PdfInputPosition {
+        T::get_pdf_input_position(self)
+    }
+}
+
+impl<T: ?Sized + GetPdfInputPosition> GetPdfInputPosition for Box<T> {
+    fn get_pdf_input_position(&self) -> PdfInputPosition {
+        T::get_pdf_input_position(self)
+    }
+}
+
+impl GetPdfInputPosition for PdfInputPosition {
+    fn get_pdf_input_position(&self) -> PdfInputPosition {
+        *self
+    }
+}
+
+impl GetPdfInputPosition for bool {
+    fn get_pdf_input_position(&self) -> PdfInputPosition {
+        PdfInputPosition::empty()
+    }
+}
+
+impl GetPdfInputPosition for i128 {
+    fn get_pdf_input_position(&self) -> PdfInputPosition {
+        PdfInputPosition::empty()
+    }
+}
+
+impl GetPdfInputPosition for f64 {
+    fn get_pdf_input_position(&self) -> PdfInputPosition {
+        PdfInputPosition::empty()
+    }
+}
+
+#[derive(Clone, Copy, Default)]
+pub struct PdfInputPositionNoCompare(pub PdfInputPosition);
+
+impl PdfInputPositionNoCompare {
+    pub const fn empty() -> Self {
+        Self(PdfInputPosition::empty())
+    }
+    pub const fn new(pos: usize) -> Self {
+        Self(PdfInputPosition::new(pos))
+    }
+}
+
+impl GetPdfInputPosition for PdfInputPositionNoCompare {
+    fn get_pdf_input_position(&self) -> PdfInputPosition {
+        self.0
+    }
+}
+
+impl From<PdfInputPosition> for PdfInputPositionNoCompare {
+    fn from(value: PdfInputPosition) -> Self {
+        Self(value)
+    }
+}
+
+impl fmt::Debug for PdfInputPositionNoCompare {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        f.debug_tuple("PdfInputPositionNoCompare")
+            .field(&format_args!("{self}"))
+            .finish()
+    }
+}
+
+impl fmt::Display for PdfInputPositionNoCompare {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        self.0.fmt(f)
+    }
+}
+
+impl Ord for PdfInputPositionNoCompare {
+    fn cmp(&self, _other: &Self) -> std::cmp::Ordering {
+        std::cmp::Ordering::Equal
+    }
+}
+
+impl PartialOrd for PdfInputPositionNoCompare {
+    fn partial_cmp(&self, other: &Self) -> Option<std::cmp::Ordering> {
+        Some(self.cmp(other))
+    }
+}
+
+impl std::hash::Hash for PdfInputPositionNoCompare {
+    fn hash<H: std::hash::Hasher>(&self, _state: &mut H) {
+        // don't hash anything since Self always compares equal
+    }
+}
+
+impl Eq for PdfInputPositionNoCompare {}
+
+impl PartialEq for PdfInputPositionNoCompare {
+    fn eq(&self, _other: &Self) -> bool {
+        true
+    }
+}
+
+#[derive(Debug)]
+#[non_exhaustive]
+pub enum PdfParseError {
+    Custom(String),
+    InvalidType {
+        pos: PdfInputPosition,
+        ty: &'static str,
+        expected_ty: &'static str,
+    },
+    InvalidName {
+        pos: PdfInputPosition,
+        name: PdfName,
+        expected_ty: &'static str,
+    },
+    NotAPdfFile,
+    TruncatedFile {
+        pos: PdfInputPosition,
+    },
+    InvalidObjectNumber {
+        pos: PdfInputPosition,
+    },
+    InvalidGenerationNumber {
+        pos: PdfInputPosition,
+    },
+    InvalidNumber {
+        pos: PdfInputPosition,
+    },
+    InvalidStringEscape {
+        pos: PdfInputPosition,
+    },
+    InvalidHexStringDigit {
+        pos: PdfInputPosition,
+    },
+    DuplicateIndirectObjectDefinition {
+        pos: PdfInputPosition,
+        id: PdfObjectIdentifier,
+    },
+    MissingObj {
+        pos: PdfInputPosition,
+    },
+    MissingEndObj {
+        pos: PdfInputPosition,
+    },
+    InvalidDictionaryClosingDoubleRAngle {
+        pos: PdfInputPosition,
+    },
+    DuplicateDictionaryKey {
+        pos: PdfInputPosition,
+        name: PdfName,
+    },
+    InvalidNameEscape {
+        pos: PdfInputPosition,
+    },
+    InvalidOrMissingEolAfterStreamKeyword {
+        pos: PdfInputPosition,
+    },
+    MissingEndStreamKeyword {
+        pos: PdfInputPosition,
+    },
+    IntegerOutOfRange {
+        pos: PdfInputPosition,
+    },
+    MissingTrailer {
+        pos: PdfInputPosition,
+    },
+    WrongArrayLength {
+        pos: PdfInputPosition,
+        len: usize,
+        expected_len: usize,
+    },
+    MissingStartXRefKeyword {
+        pos: PdfInputPosition,
+    },
+    MissingStartXRefValue {
+        pos: PdfInputPosition,
+    },
+    MissingEofComment {
+        pos: PdfInputPosition,
+    },
+    UnexpectedByte {
+        pos: PdfInputPosition,
+        byte: u8,
+    },
+    InvalidStartXRefValue {
+        pos: PdfInputPosition,
+        start_xref: usize,
+    },
+}
+
+impl From<std::convert::Infallible> for PdfParseError {
+    fn from(value: std::convert::Infallible) -> Self {
+        match value {}
+    }
+}
+
+impl GetPdfInputPosition for PdfParseError {
+    fn get_pdf_input_position(&self) -> PdfInputPosition {
+        match *self {
+            PdfParseError::Custom(_) | PdfParseError::NotAPdfFile => PdfInputPosition::empty(),
+            PdfParseError::InvalidType { pos, .. }
+            | PdfParseError::InvalidName { pos, .. }
+            | PdfParseError::TruncatedFile { pos }
+            | PdfParseError::InvalidObjectNumber { pos }
+            | PdfParseError::InvalidGenerationNumber { pos }
+            | PdfParseError::InvalidNumber { pos }
+            | PdfParseError::InvalidStringEscape { pos }
+            | PdfParseError::InvalidHexStringDigit { pos }
+            | PdfParseError::DuplicateIndirectObjectDefinition { pos, .. }
+            | PdfParseError::MissingObj { pos }
+            | PdfParseError::MissingEndObj { pos }
+            | PdfParseError::InvalidDictionaryClosingDoubleRAngle { pos }
+            | PdfParseError::DuplicateDictionaryKey { pos, .. }
+            | PdfParseError::InvalidNameEscape { pos }
+            | PdfParseError::InvalidOrMissingEolAfterStreamKeyword { pos }
+            | PdfParseError::MissingEndStreamKeyword { pos }
+            | PdfParseError::IntegerOutOfRange { pos }
+            | PdfParseError::MissingTrailer { pos }
+            | PdfParseError::WrongArrayLength { pos, .. }
+            | PdfParseError::MissingStartXRefKeyword { pos }
+            | PdfParseError::MissingStartXRefValue { pos }
+            | PdfParseError::MissingEofComment { pos }
+            | PdfParseError::UnexpectedByte { pos, .. }
+            | PdfParseError::InvalidStartXRefValue { pos, .. } => pos,
+        }
+    }
+}
+
+impl fmt::Display for PdfParseError {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
+        match *self {
+            PdfParseError::Custom(ref v) => f.write_str(v),
+            PdfParseError::InvalidType {
+                pos,
+                ty,
+                expected_ty,
+            } => {
+                write!(
+                    f,
+                    "at {pos}: invalid type: expected {expected_ty}, got {ty}"
+                )
+            }
+            PdfParseError::InvalidName {
+                pos,
+                ref name,
+                expected_ty,
+            } => {
+                write!(
+                    f,
+                    "at {pos}: invalid name: expected a(n) {expected_ty}, got {name}"
+                )
+            }
+            PdfParseError::NotAPdfFile => f.write_str("Not a PDF file"),
+            PdfParseError::TruncatedFile { pos } => {
+                write!(f, "at {pos}: PDF file is truncated too early")
+            }
+            PdfParseError::InvalidObjectNumber { pos } => {
+                write!(f, "at {pos}: PDF object number is invalid")
+            }
+            PdfParseError::InvalidGenerationNumber { pos } => {
+                write!(
+                    f,
+                    "at {pos}: PDF object identifier's generation number is invalid"
+                )
+            }
+            PdfParseError::InvalidNumber { pos } => {
+                write!(f, "at {pos}: invalid number")
+            }
+            PdfParseError::InvalidStringEscape { pos } => {
+                write!(f, "at {pos}: invalid string escape")
+            }
+            PdfParseError::InvalidHexStringDigit { pos } => {
+                write!(f, "at {pos}: invalid hex string digit")
+            }
+            PdfParseError::DuplicateIndirectObjectDefinition { pos, id } => {
+                write!(f, "at {pos}: duplicate indirect object definition: {id:?}")
+            }
+            PdfParseError::MissingObj { pos } => {
+                write!(
+                    f,
+                    "at {pos}: indirect object definition is missing `obj` keyword"
+                )
+            }
+            PdfParseError::MissingEndObj { pos } => {
+                write!(
+                    f,
+                    "at {pos}: indirect object definition is missing `endobj` keyword"
+                )
+            }
+            PdfParseError::InvalidDictionaryClosingDoubleRAngle { pos } => {
+                write!(f, "at {pos}: dictionary has an invalid closing `>>` symbol")
+            }
+            PdfParseError::DuplicateDictionaryKey { pos, ref name } => {
+                write!(f, "at {pos}: duplicate dictionary key: {name}")
+            }
+            PdfParseError::InvalidNameEscape { pos } => {
+                write!(f, "at {pos}: invalid name escape")
+            }
+            PdfParseError::InvalidOrMissingEolAfterStreamKeyword { pos } => {
+                write!(
+                    f,
+                    "at {pos}: invalid or missing end-of-line after `stream` keyword"
+                )
+            }
+            PdfParseError::MissingEndStreamKeyword { pos } => {
+                write!(f, "at {pos}: missing `endstream` keyword")
+            }
+            PdfParseError::IntegerOutOfRange { pos } => {
+                write!(f, "at {pos}: integer out of range")
+            }
+            PdfParseError::MissingTrailer { pos } => {
+                write!(f, "at {pos}: missing `trailer` keyword")
+            }
+            PdfParseError::WrongArrayLength {
+                pos,
+                len,
+                expected_len,
+            } => {
+                write!(
+                    f,
+                    "at {pos}: wrong array length: expected {expected_len}, got {len}"
+                )
+            }
+            PdfParseError::MissingStartXRefKeyword { pos } => {
+                write!(f, "at {pos}: missing `startxref` keyword")
+            }
+            PdfParseError::MissingStartXRefValue { pos } => {
+                write!(f, "at {pos}: missing `startxref` value")
+            }
+            PdfParseError::MissingEofComment { pos } => {
+                write!(f, "at {pos}: missing `%%EOF` comment")
+            }
+            PdfParseError::UnexpectedByte { pos, byte } => {
+                write!(f, "at {pos}: unexpected byte {}", byte.escape_ascii())
+            }
+            PdfParseError::InvalidStartXRefValue { pos, start_xref } => {
+                write!(
+                    f,
+                    "at {pos}: invalid `startxref` value: {start_xref} ({start_xref:#x})"
+                )
+            }
+        }
+    }
+}
+
+impl std::error::Error for PdfParseError {}
+
+pub trait PdfParse: Sized + 'static {
+    fn type_name() -> Cow<'static, str>;
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError>;
+    fn parse_option(object: PdfObject) -> Result<Option<Self>, PdfParseError> {
+        match object {
+            PdfObject::Null(_) => Ok(None),
+            PdfObject::Indirect(ref v) if v.get().is_null() => Ok(None),
+            PdfObject::Boolean(_)
+            | PdfObject::Integer(_)
+            | PdfObject::Real(_)
+            | PdfObject::String(_)
+            | PdfObject::Name(_)
+            | PdfObject::Array(_)
+            | PdfObject::Dictionary(_)
+            | PdfObject::Stream(_)
+            | PdfObject::Indirect(_) => Self::parse(object).map(Some),
+        }
+    }
+}
+
+impl<T: PdfParse> PdfParse for Option<T> {
+    fn type_name() -> Cow<'static, str> {
+        T::type_name()
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        T::parse_option(object)
+    }
+    fn parse_option(object: PdfObject) -> Result<Option<Self>, PdfParseError> {
+        if matches!(object, PdfObject::Null(_)) {
+            Ok(None)
+        } else {
+            Self::parse(object).map(Some)
+        }
+    }
+}
+
+macro_rules! impl_pdf_parse_prim_int {
+    ($ty:ident) => {
+        impl PdfParse for $ty {
+            fn type_name() -> Cow<'static, str> {
+                Cow::Borrowed(stringify!($ty))
+            }
+            fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+                let v: PdfInteger = PdfParse::parse(object)?;
+                v.value()
+                    .try_into()
+                    .map_err(|_| PdfParseError::IntegerOutOfRange { pos: v.pos() })
+            }
+        }
+        impl PdfParse for NonZero<$ty> {
+            fn type_name() -> Cow<'static, str> {
+                Cow::Borrowed(concat!("NonZero<", stringify!($ty), ">"))
+            }
+            fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+                let v: PdfInteger = PdfParse::parse(object)?;
+                v.value()
+                    .try_into()
+                    .ok()
+                    .and_then(NonZero::new)
+                    .ok_or(PdfParseError::IntegerOutOfRange { pos: v.pos() })
+            }
+        }
+    };
+}
+
+impl_pdf_parse_prim_int!(u8);
+impl_pdf_parse_prim_int!(i8);
+impl_pdf_parse_prim_int!(u16);
+impl_pdf_parse_prim_int!(i16);
+impl_pdf_parse_prim_int!(u32);
+impl_pdf_parse_prim_int!(i32);
+impl_pdf_parse_prim_int!(u64);
+impl_pdf_parse_prim_int!(i64);
+impl_pdf_parse_prim_int!(u128);
+impl_pdf_parse_prim_int!(usize);
+impl_pdf_parse_prim_int!(isize);
+
+impl PdfParse for i128 {
+    fn type_name() -> Cow<'static, str> {
+        Cow::Borrowed("i128")
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        let v: PdfInteger = PdfParse::parse(object)?;
+        Ok(v.value().into())
+    }
+}
+
+impl PdfParse for NonZero<i128> {
+    fn type_name() -> Cow<'static, str> {
+        Cow::Borrowed("NonZero<i128>")
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        let v: PdfInteger = PdfParse::parse(object)?;
+        NonZero::new(v.value().into()).ok_or(PdfParseError::IntegerOutOfRange { pos: v.pos() })
+    }
+}
+
+impl PdfParse for f64 {
+    fn type_name() -> Cow<'static, str> {
+        Cow::Borrowed("f64")
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        Ok(<PdfReal as PdfParse>::parse(object)?.value())
+    }
+}
+
+impl PdfParse for f32 {
+    fn type_name() -> Cow<'static, str> {
+        Cow::Borrowed("f32")
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        Ok(<f64 as PdfParse>::parse(object)? as f32)
+    }
+}
+
+impl PdfParse for PdfNull {
+    fn type_name() -> Cow<'static, str> {
+        Cow::Borrowed("null")
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        match PdfObjectDirect::from(object) {
+            PdfObjectDirect::Null(v) => Ok(v),
+            object => Err(PdfParseError::InvalidType {
+                pos: object.pos(),
+                ty: object.type_name(),
+                expected_ty: "null",
+            }),
+        }
+    }
+    fn parse_option(object: PdfObject) -> Result<Option<Self>, PdfParseError> {
+        Self::parse(object).map(Some)
+    }
+}
+
+impl PdfParse for PdfObjectNonNull {
+    fn type_name() -> Cow<'static, str> {
+        Cow::Borrowed("non-null")
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        Option::<PdfObjectNonNull>::from(object).ok_or(PdfParseError::InvalidType {
+            pos: PdfInputPosition::empty(),
+            ty: "null",
+            expected_ty: "non-null",
+        })
+    }
+    fn parse_option(object: PdfObject) -> Result<Option<Self>, PdfParseError> {
+        Ok(object.into())
+    }
+}
+
+impl PdfParse for PdfObjectDirect {
+    fn type_name() -> Cow<'static, str> {
+        Cow::Borrowed("direct object")
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        Ok(object.into())
+    }
+}
+
+impl PdfParse for PdfObject {
+    fn type_name() -> Cow<'static, str> {
+        Cow::Borrowed("object")
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        Ok(object)
+    }
+}
+
+impl PdfParse for PdfObjectIndirect {
+    fn type_name() -> Cow<'static, str> {
+        Cow::Borrowed("indirect object")
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        match object {
+            PdfObject::Indirect(v) => Ok(v),
+            _ => Err(PdfParseError::InvalidType {
+                pos: object.pos(),
+                ty: object.type_name(),
+                expected_ty: "indirect object",
+            }),
+        }
+    }
+    fn parse_option(object: PdfObject) -> Result<Option<Self>, PdfParseError> {
+        match object {
+            PdfObject::Indirect(v) => Ok(Some(v)),
+            PdfObject::Null(_) => Ok(None),
+            _ => Err(PdfParseError::InvalidType {
+                pos: object.pos(),
+                ty: object.type_name(),
+                expected_ty: "indirect object",
+            }),
+        }
+    }
+}
+
+impl<T: PdfParse, const N: usize> PdfParse for [T; N] {
+    fn type_name() -> Cow<'static, str> {
+        Cow::Owned(format!("[{}; {N}]", T::type_name()))
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        match PdfObjectDirect::from(object) {
+            PdfObjectDirect::Array(array) => {
+                let array_pos = array.pos();
+                let elements = array.into_elements();
+                let mut elements: Arc<[PdfObject; N]> =
+                    elements.try_into().map_err(|elements: Arc<[PdfObject]>| {
+                        PdfParseError::WrongArrayLength {
+                            pos: array_pos,
+                            len: elements.len(),
+                            expected_len: N,
+                        }
+                    })?;
+                let elements: Box<[T]> = if let Some(elements) = Arc::get_mut(&mut elements) {
+                    Result::from_iter(elements.iter_mut().map(|v| T::parse(mem::take(v))))?
+                } else {
+                    Result::from_iter(elements.iter().map(|v| T::parse(v.clone())))?
+                };
+                Ok(*Box::<[T; N]>::try_from(elements)
+                    .ok()
+                    .expect("already checked length"))
+            }
+            object => Err(PdfParseError::InvalidType {
+                pos: object.pos(),
+                ty: object.type_name(),
+                expected_ty: "array",
+            }),
+        }
+    }
+}
+
+impl<T: PdfParse> PdfParse for Arc<[T]> {
+    fn type_name() -> Cow<'static, str> {
+        Cow::Owned(format!("Arc<[{}]>", T::type_name()))
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        match PdfObjectDirect::from(object) {
+            PdfObjectDirect::Array(array) => {
+                let mut elements = array.into_elements();
+                if let Some(retval) = <dyn Any>::downcast_ref::<Self>(&elements) {
+                    return Ok(retval.clone());
+                }
+                if let Some(elements) = Arc::get_mut(&mut elements) {
+                    Result::from_iter(elements.iter_mut().map(|v| T::parse(mem::take(v))))
+                } else {
+                    Result::from_iter(elements.iter().map(|v| T::parse(v.clone())))
+                }
+            }
+            PdfObjectDirect::Null(_) => Ok(Self::default()),
+            object => Err(PdfParseError::InvalidType {
+                pos: object.pos(),
+                ty: object.type_name(),
+                expected_ty: "array",
+            }),
+        }
+    }
+}
+
+impl<T: PdfParse> PdfParse for MaybeArray<T> {
+    fn type_name() -> Cow<'static, str> {
+        Cow::Owned(format!("MaybeArray<{}>", T::type_name()))
+    }
+    fn parse(object: PdfObject) -> Result<Self, PdfParseError> {
+        match PdfObjectDirect::from(object) {
+            PdfObjectDirect::Null(_) => Ok(Self::default()),
+            PdfObjectDirect::Array(object) => Ok(Self(PdfParse::parse(object.into())?)),
+            object => Ok(Self(Arc::new([PdfParse::parse(object.into())?]))),
+        }
+    }
+}
+
+#[macro_export]
+macro_rules! pdf_parse {
+    (
+        $(#[$($struct_meta:tt)*])*
+        $struct_vis:vis struct $Struct:ident$(<$($StructParam:ident $(: $StructBound:tt)? $(= $StructParamDefault:ty)?),* $(,)?>)? {
+            $(#[pdf $($pdf_meta:tt)*]
+            $(#[$($field_meta:tt)*])*
+            $field_vis:vis $field_name:ident: $field_ty:ty,)*
+        }
+    ) => {
+        $(#[$($struct_meta)*])*
+        $struct_vis struct $Struct$(<$($StructParam $(: $StructBound)? $(= $StructParamDefault)?),*>)? {
+            $($(#[$($field_meta)*])*
+            $field_vis $field_name: $field_ty,)*
+        }
+
+        $crate::pdf::parse::pdf_parse! {
+            @impl
+            struct $Struct$(<$($StructParam $(: $StructBound)?),*>)? {
+                $(#[pdf $($pdf_meta)*]
+                $(#[$($field_meta)*])*
+                $field_name: $field_ty,)*
+            }
+        }
+    };
+    (
+        @impl
+        struct $Struct:ident$(<$($StructParam:ident $(: $StructBound:tt)?),* $(,)?>)? {
+            $($(#[$($field_meta:tt)*])*
+            $field_name:ident: $field_ty:ty,)*
+        }
+    ) => {
+        impl$(<$($StructParam: $crate::pdf::parse::PdfParse $(+ $StructBound)?),*>)? $crate::pdf::parse::PdfParse for $Struct$(<$($StructParam),*>)? {
+            fn type_name() -> $crate::__std::borrow::Cow<'static, $crate::__std::primitive::str> {
+                let args: &[$crate::__std::borrow::Cow<'static, $crate::__std::primitive::str>] = &[
+                    $($(<$StructParam as $crate::pdf::parse::PdfParse>::type_name()),*)?
+                ];
+                if args.is_empty() {
+                    $crate::__std::borrow::Cow::Borrowed($crate::__std::stringify!($Struct))
+                } else {
+                    let mut retval = $crate::__std::string::String::new();
+                    retval.push_str($crate::__std::stringify!($Struct));
+                    retval.push_str("<");
+                    let mut first = true;
+                    for arg in args {
+                        if first {
+                            first = false;
+                        } else {
+                            retval.push_str(", ");
+                        }
+                        retval.push_str(arg);
+                    }
+                    retval.push_str(">");
+                    $crate::__std::borrow::Cow::Owned(retval)
+                }
+            }
+            fn parse(object: $crate::pdf::object::PdfObject) -> $crate::__std::result::Result<Self, $crate::pdf::parse::PdfParseError> {
+                let object = $crate::__std::convert::From::from(object);
+                let $crate::pdf::object::PdfObjectDirect::Dictionary(object) = object else {
+                    return $crate::__std::result::Result::Err($crate::pdf::parse::PdfParseError::InvalidType {
+                        pos: object.pos(),
+                        ty: object.type_name(),
+                        expected_ty: $crate::__std::stringify!($Struct),
+                    });
+                };
+                let pos = object.pos();
+                let mut object = object.into_fields();
+                let object_mut = $crate::__std::sync::Arc::make_mut(&mut object);
+                let _ = object_mut;
+                $($crate::pdf::parse::pdf_parse! {
+                    @impl_struct_field(pos, object, object_mut)
+                    []
+                    $(#[$($field_meta)*])*
+                    $field_name: $field_ty
+                })*
+                $crate::__std::result::Result::Ok(Self {
+                    $($field_name,)*
+                })
+            }
+        }
+    };
+    (
+        @impl_struct_field($pos:ident, $object:ident, $object_mut:ident)
+        [$(#[$($prev_field_meta:tt)*])*]
+        #[pdf $pdf_meta:tt]
+        $(#[$($field_meta:tt)*])*
+        $field_name:ident: $field_ty:ty
+    ) => {
+        $crate::pdf::parse::pdf_parse! {
+            @impl_struct_field($pos, $object, $object_mut, pdf $pdf_meta)
+            [$(#[$($prev_field_meta)*])*]
+            $(#[$($field_meta)*])*
+            $field_name: $field_ty
+        }
+    };
+    (
+        @impl_struct_field($pos:ident, $object:ident, $object_mut:ident $($pdf_meta:tt)*)
+        [$(#[$($prev_field_meta:tt)*])*]
+        #[$($next_field_meta:tt)*]
+        $(#[$($field_meta:tt)*])*
+        $field_name:ident: $field_ty:ty
+    ) => {
+        $crate::pdf::parse::pdf_parse! {
+            @impl_struct_field($pos, $object, $object_mut $($pdf_meta)*)
+            [$(#[$($prev_field_meta)*])* #[$($next_field_meta)*]]
+            $(#[$($field_meta)*])*
+            $field_name: $field_ty
+        }
+    };
+    (
+        @impl_struct_field($pos:ident, $object:ident, $object_mut:ident, pdf(flatten))
+        [$(#[$($field_meta:tt)*])*]
+        $field_name:ident: $field_ty:ty
+    ) => {
+        let $field_name = <$field_ty as $crate::pdf::parse::PdfParse>::parse(
+            $crate::pdf::object::PdfObject::Dictionary(
+                $crate::pdf::object::PdfDictionary::from_fields($pos, $object),
+            ),
+        )?;
+    };
+    (
+        @impl_struct_field($pos:ident, $object:ident, $object_mut:ident, pdf(name = $name:expr))
+        [$(#[$($field_meta:tt)*])*]
+        $field_name:ident: $field_ty:ty
+    ) => {
+        let $field_name = $crate::pdf::object::PdfName::new_static(
+            $crate::__std::convert::AsRef::<[u8]>::as_ref($name),
+        );
+        let $field_name = <$field_ty as $crate::pdf::parse::PdfParse>::parse(
+            $object_mut
+                .remove(&$field_name)
+                .unwrap_or($crate::pdf::object::PdfObject::Null($crate::pdf::object::PdfNull::new($pos))),
+        )?;
+    };
+    (
+        $(#[$($enum_meta:tt)*])*
+        $enum_vis:vis enum $Enum:ident {
+            $(#[pdf $($pdf_meta:tt)*]
+            $(#[$($variant_meta:tt)*])*
+            $VariantName:ident $(($($variant_paren_body:tt)*))? $({$($variant_brace_body:tt)*})?,)*
+        }
+    ) => {
+        $(#[$($enum_meta)*])*
+        $enum_vis enum $Enum {
+            $($(#[$($variant_meta)*])*
+            $VariantName $(($($variant_paren_body)*))? $({$($variant_brace_body)*})?,)*
+        }
+
+        $crate::pdf::parse::pdf_parse! {
+            @impl
+            $(#[$($enum_meta)*])*
+            enum $Enum {
+                $(#[pdf $($pdf_meta)*]
+                $(#[$($variant_meta)*])*
+                $VariantName $(($($variant_paren_body)*))? $({$($variant_brace_body)*})?,)*
+            }
+        }
+    };
+    (
+        @impl
+        $(#[$($enum_meta:tt)*])*
+        enum $Enum:ident {
+            $(#[pdf(name = $name:expr)]
+            $(#[$($variant_meta:tt)*])*
+            $VariantName:ident,)*
+            $(#[pdf(other)]
+            $(#[$($variant_meta_other:tt)*])*
+            $VariantNameOther:ident($($PdfName:tt)*),)?
+        }
+    ) => {
+        impl $crate::__std::convert::From<$Enum> for $crate::pdf::object::PdfName {
+            fn from(value: $Enum) -> Self {
+                match value {
+                    $($Enum::$VariantName => $crate::pdf::object::PdfName::new_static(
+                        $crate::__std::convert::AsRef::<[u8]>::as_ref($name),
+                    ),)*
+                    $($Enum::$VariantNameOther(v) => $crate::__std::convert::Into::into(v),)?
+                }
+            }
+        }
+
+        $crate::pdf::parse::pdf_parse! {
+            @impl_try_from
+            $(#[$($enum_meta)*])*
+            enum $Enum {
+                $(#[pdf(name = $name)]
+                $(#[$($variant_meta)*])*
+                $VariantName,)*
+                $(#[pdf(other)]
+                $(#[$($variant_meta_other)*])*
+                $VariantNameOther($($PdfName)*),)?
+            }
+        }
+
+        impl $crate::pdf::parse::PdfParse for $Enum {
+            fn type_name() -> $crate::__std::borrow::Cow<'static, $crate::__std::primitive::str> {
+                $crate::__std::borrow::Cow::Borrowed($crate::__std::stringify!($Struct))
+            }
+            fn parse(object: $crate::pdf::object::PdfObject) -> $crate::__std::result::Result<Self, $crate::pdf::parse::PdfParseError> {
+                let object = $crate::__std::convert::From::from(object);
+                let $crate::pdf::object::PdfObjectDirect::Name(name) = object else {
+                    return $crate::__std::result::Result::Err($crate::pdf::parse::PdfParseError::InvalidType {
+                        pos: object.pos(),
+                        ty: object.type_name(),
+                        expected_ty: $crate::__std::stringify!($Struct),
+                    });
+                };
+                $crate::__std::result::Result::Ok($crate::__std::convert::TryInto::<$Enum>::try_into(name)?)
+            }
+        }
+    };
+    (
+        @impl_try_from
+        $(#[$($enum_meta:tt)*])*
+        enum $Enum:ident {
+            $(#[pdf(name = $name:expr)]
+            $(#[$($variant_meta:tt)*])*
+            $VariantName:ident,)*
+            #[pdf(other)]
+            $(#[$($variant_meta_other:tt)*])*
+            $VariantNameOther:ident(PdfName),
+        }
+    ) => {
+        impl $crate::__std::convert::From<$crate::pdf::object::PdfName> for $Enum {
+            fn from(name: $crate::pdf::object::PdfName) -> Self {
+                $(if name == $crate::pdf::object::PdfName::new_static(
+                    $crate::__std::convert::AsRef::<[u8]>::as_ref($name),
+                ) {
+                    $Enum::$VariantName
+                } else)* {
+                    $Enum::$VariantNameOther(name)
+                }
+            }
+        }
+    };
+    (
+        @impl_try_from
+        $(#[$($enum_meta:tt)*])*
+        enum $Enum:ident {
+            $(#[pdf(name = $name:expr)]
+            $(#[$($variant_meta:tt)*])*
+            $VariantName:ident,)*
+            #[pdf(other)]
+            $(#[$($variant_meta_other:tt)*])*
+            $VariantNameOther:ident($PdfName:ty),
+        }
+    ) => {
+        impl $crate::__std::convert::TryFrom<$crate::pdf::object::PdfName> for $Enum {
+            type Error = $crate::pdf::parse::PdfParseError;
+
+            fn try_from(name: $crate::pdf::object::PdfName) -> $crate::__std::result::Result<Self, $crate::pdf::parse::PdfParseError> {
+                $(if name == $crate::pdf::object::PdfName::new_static(
+                    $crate::__std::convert::AsRef::<[u8]>::as_ref($name),
+                ) {
+                    $crate::__std::result::Result::Ok($Enum::$VariantName)
+                } else)* {
+                    $crate::__std::result::Result::Ok($Enum::$VariantNameOther($crate::__std::convert::TryInto::<$PdfName>::try_into(name)?))
+                }
+            }
+        }
+    };
+    (
+        @impl_try_from
+        $(#[$($enum_meta:tt)*])*
+        enum $Enum:ident {
+            $(#[pdf(name = $name:expr)]
+            $(#[$($variant_meta:tt)*])*
+            $VariantName:ident,)*
+        }
+    ) => {
+        impl $crate::__std::convert::TryFrom<$crate::pdf::object::PdfName> for $Enum {
+            type Error = $crate::pdf::parse::PdfParseError;
+
+            fn try_from(name: $crate::pdf::object::PdfName) -> $crate::__std::result::Result<Self, $crate::pdf::parse::PdfParseError> {
+                $(if name == $crate::pdf::object::PdfName::new_static(
+                    $crate::__std::convert::AsRef::<[u8]>::as_ref($name),
+                ) {
+                    $crate::__std::result::Result::Ok($Enum::$VariantName)
+                } else)* {
+                    $crate::__std::result::Result::Err($crate::pdf::parse::PdfParseError::InvalidName {
+                        pos: name.pos(),
+                        name,
+                        expected_ty: $crate::__std::stringify!($Struct),
+                    })
+                }
+            }
+        }
+    };
+}
+
+pub use pdf_parse;

src/util.rs

@@ -5,7 +5,7 @@ use std::{
     sync::Arc,
 };
 
-pub(crate) enum ArcOrRef<'a, T: ?Sized> {
+pub enum ArcOrRef<'a, T: ?Sized> {
     Arc(Arc<T>),
     Ref(&'a T),
 }