Struct CMap

Source

pub struct CMap { /* private fields */ }

Expand description

A CMap: a byte decoder plus a charcode→CID map (ISO 32000-1 §9.7.5).

Build one with from_encoding_name, predefined or parse_embedded.

Implementations§

Source §

impl CMap

Source

pub fn decode<'a>( &'a self, bytes: &'a [u8], ) -> impl Iterator<Item = (CharCode, Cid)> + 'a

Decode a string into (character code, CID) pairs — the entry point every text path shares.

Iteration ends when the string is exhausted. Damage never stops it early: a code no codespace range accepts comes back as CharCode(0), and a code cut off by the end of the string does too.

use pdfrum_cmap::{CharCode, Cid, from_encoding_name};
use pdfrum_common::Diagnostics;
use pdfrum_object::Name;

let mut diags = Diagnostics::default();
let cmap = from_encoding_name(&Name::from("GB-EUC-H"), &mut diags);

// 0x41 is not a lead byte, so it is a one-byte code; 0xA1 0xA1 is a pair.
let codes: Vec<CharCode> = cmap.decode(&[0x41, 0xA1, 0xA1]).map(|(c, _)| c).collect();
assert_eq!(codes, vec![CharCode(0x41), CharCode(0xA1A1)]);
assert_eq!(cmap.cid(CharCode(0xA1A1)), Cid(0x0060));

Source

pub fn next_char(&self, bytes: &[u8], offset: &mut usize) -> CharCode

Decode one character code, advancing offset.

Returns CharCode(0) rather than failing when the bytes are damaged. offset may land past a truncated code’s start without reaching the end of the string, which is how iteration terminates.

Source

pub fn cid(&self, code: CharCode) -> Cid

The CID a character code maps to. Unmapped codes give Cid(0), which is .notdef.

A CMap that inherits another through usecmap or /UseCMap answers from its own tables first and asks its parent only where it maps nothing — the child-wins rule of ISO 32000-1 §9.7.5.3.

Source

pub fn charcode_from_cid(&self, cid: Cid) -> CharCode

The character code that maps to cid, or CharCode(0) when none does.

Defined only for a predefined CMap backed by the built-in tables; every other kind returns 0. Even for those it is partial: the scan does not consult the four-byte tables, so a CID that exists only above 0x1_0000 is unreachable in reverse. This is used by CID fonts to find a code for a character they were asked to draw by Unicode.

Source

pub fn char_size(&self, code: CharCode) -> u8

How many bytes a code of this value occupies.

Derived from the value alone, not from the bytes it was decoded from and not from the codespace ranges. That makes it disagree with append_char in exactly one case: under a mixed-two-byte scheme a code below 0x100 whose value is itself a lead byte reports width 1 but is written as two bytes. Code that needs the encoded width should measure what append_char produced.

Source

pub fn count_chars(&self, bytes: &[u8]) -> usize

How many character codes a byte string holds. Always equal to the number of pairs decode yields for the same string.

Source

pub fn append_char(&self, out: &mut Vec<u8>, code: CharCode)

Append a character code to a byte string in this CMap’s encoding — the inverse of next_char.

Source

pub fn is_vertical(&self) -> bool

Whether this CMap selects vertical writing mode (ISO 32000-1 §9.7.4.3).

Decided by the last byte of the name as written, before any suffix handling, so Identity-V is vertical and so is any name ending in V.

Source

pub fn is_loaded(&self) -> bool

Whether the name resolved to real tables.

false covers two different failures that behave the same way: a name matching no row at all, and a name whose row was found but whose built-in table was not. Either way the CMap still decodes and still maps codes; it just maps them to themselves.

Source

pub fn coding(&self) -> CidCoding

The legacy encoding family this CMap’s codes belong to.

Source

pub fn charset(&self) -> CidSet

The character collection its CIDs index.

Source

pub fn coding_scheme(&self) -> CodingScheme

How its bytes split into character codes.

Source

pub fn has_no_direct_table(&self) -> bool

Whether this CMap has no dense charcode→CID table — true for every predefined CMap and false for every embedded one, whatever the program contained. CID fonts branch on it when choosing how to find a glyph.

Source