Struct EmbeddedFont

Source

pub struct EmbeddedFont { /* private fields */ }

Expand description

A font dictionary this session added, ready to name from a content stream.

EmbeddedFont::object is the /Font resource crate::content::ResourceTable::realize will allocate a name for. EmbeddedFont::encode turns Unicode into the character codes a Tj/TJ of that font expects.

Implementations§

Source §

impl EmbeddedFont

Source

pub fn object(&self) -> ObjRef

The /Font dictionary to name from a page resource.

Source

pub fn encode(&self, text: &str) -> Vec<u8> ⓘ

Character codes for text in this font’s encoding.

Unmappable characters become .notdef (code 0). A composite font writes two-byte big-endian GIDs (Identity-H); a simple or standard font writes one byte.

A font from EditDoc::embed_cid_font writes two-byte big-endian CIDs, and finds them by inverting the caller’s /ToUnicode CMap rather than by consulting the program’s cmap. That is not a convenience: under a caller-supplied CMap the program’s cmap does not describe the file’s code space, and the CMap is the document’s only statement of what its codes mean. A character the CMap does not reach is code 0, as everywhere else.

Source

pub fn encode_checked(&self, text: &str) -> Result<Vec<u8>, MissingGlyph>

Character codes for text, refusing a character this font cannot draw.

The same encoding as EmbeddedFont::encode, with the .notdef fallback replaced by an error. This is what a caller who is placing text wants: a watermark whose degree sign silently became a blank is worse than one that refused to be written.

use pdfrum_edit::{EditDoc, StandardFont};
use pdfrum_parser::{LoadOptions, load};
use std::sync::Arc;

let bytes: Arc<[u8]> = Arc::from(&include_bytes!("../../tests/files/hello.pdf")[..]);
let doc = load(bytes, &LoadOptions::default())?;
let mut edit = EditDoc::new(&doc);
let font = edit.standard_font(StandardFont::Helvetica)?;

assert!(font.encode_checked("Hi").is_ok());
// WinAnsi has no Han: refused rather than drawn blank.
assert_eq!(font.encode_checked("\u{4e00}").unwrap_err().character, '\u{4e00}');