Fix embedding of CID-keyed fonts into PDF

Fixes #2032
Fixes #573

3 files changed, 52 insertions, 21 deletions

diff --git a/crates/typst/src/export/pdf/font.rs b/crates/typst/src/export/pdf/font.rs
index 4b10ead9..13c69118 100644
--- a/crates/typst/src/export/pdf/font.rs
+++ b/crates/typst/src/export/pdf/font.rs
@@ -37,14 +37,8 @@ pub fn write_fonts(ctx: &mut PdfContext) {
 
         // Do we have a TrueType or CFF font?
         //
-        // FIXME 1: CFF2 must be handled differently and requires PDF 2.0
+        // FIXME: CFF2 must be handled differently and requires PDF 2.0
         // (or we have to convert it to CFF).
-        //
-        // FIXME 2: CFF fonts that have a Top DICT that uses CIDFont operators
-        // may not have an identity CID-GID encoding. These are currently not
-        // handled correctly. See also:
-        // - PDF Spec, Section 9.7.4.2
-        // - https://stackoverflow.com/questions/74165171/embedded-opentype-cff-font-in-a-pdf-shows-strange-behaviour-in-some-viewers
         let is_cff = ttf
             .raw_face()
             .table(CFF)
@@ -83,11 +77,15 @@ pub fn write_fonts(ctx: &mut PdfContext) {
         }
 
         // Extract the widths of all glyphs.
-        let num_glyphs = ttf.number_of_glyphs();
-        let mut widths = vec![0.0; num_glyphs as usize];
-        for g in std::iter::once(0).chain(glyph_set.keys().copied()) {
-            let x = ttf.glyph_hor_advance(GlyphId(g)).unwrap_or(0);
-            widths[g as usize] = font.to_em(x).to_font_units();
+        let mut widths = vec![];
+        for gid in std::iter::once(0).chain(glyph_set.keys().copied()) {
+            let width = ttf.glyph_hor_advance(GlyphId(gid)).unwrap_or(0);
+            let units = font.to_em(width).to_font_units();
+            let cid = glyph_cid(font, gid);
+            if usize::from(cid) >= widths.len() {
+                widths.resize(usize::from(cid) + 1, 0.0);
+                widths[usize::from(cid)] = units;
+            }
         }
 
         // Write all non-zero glyph widths.
@@ -203,8 +201,8 @@ fn create_cmap(
     ttf: &ttf_parser::Face,
     glyph_set: &mut BTreeMap<u16, EcoString>,
 ) -> UnicodeCmap {
-    // For glyphs that have codepoints mapping to in the font's cmap table, we
-    // prefer them over pre-existing text mappings from the document. Only
+    // For glyphs that have codepoints mapping to them in the font's cmap table,
+    // we prefer them over pre-existing text mappings from the document. Only
     // things that don't have a corresponding codepoint (or only a private-use
     // one) like the "Th" in Linux Libertine get the text of their first
     // occurrences in the document instead.
@@ -236,3 +234,35 @@ fn create_cmap(
 
     cmap
 }
+
+/// Get the CID for a glyph id.
+///
+/// When writing text into a PDF, we have to specify CIDs (character ids) not
+/// GIDs (glyph IDs).
+///
+/// Most of the time, the mapping between these two is an identity mapping. In
+/// particular, for TrueType fonts, the mapping is an identity mapping because
+/// of this line above:
+/// ```ignore
+/// cid.cid_to_gid_map_predefined(Name(b"Identity"));
+/// ```
+///
+/// However, CID-keyed CFF fonts may have a non-identity mapping defined in
+/// their charset. For those, we must map the glyph IDs in a `TextItem` to CIDs.
+/// The font defines the map through its charset. The charset usually maps
+/// glyphs to SIDs (string ids) specifying the glyph's name. Not for CID-keyed
+/// fonts though! For these, the SIDs are CIDs in disguise. Relevant quote from
+/// the CFF spec:
+///
+/// > The charset data, although in the same format as non-CIDFonts, will
+/// > represent CIDs rather than SIDs, [...]
+///
+/// This function performs the mapping from glyph ID to CID. It also works for
+/// non CID-keyed fonts. Then, it will simply return the glyph ID.
+pub(super) fn glyph_cid(font: &Font, glyph_id: u16) -> u16 {
+    font.ttf()
+        .tables()
+        .cff
+        .and_then(|cff| cff.glyph_cid(ttf_parser::GlyphId(glyph_id)))
+        .unwrap_or(glyph_id)
+}
diff --git a/crates/typst/src/export/pdf/mod.rs b/crates/typst/src/export/pdf/mod.rs
index 818f436f..7a831445 100644
--- a/crates/typst/src/export/pdf/mod.rs
+++ b/crates/typst/src/export/pdf/mod.rs
@@ -253,8 +253,8 @@ where
         });
     }
 
-    fn map(&self, item: T) -> usize {
-        self.to_pdf[&item]
+    fn map(&self, item: &T) -> usize {
+        self.to_pdf[item]
     }
 
     fn pdf_indices<'a>(
diff --git a/crates/typst/src/export/pdf/page.rs b/crates/typst/src/export/pdf/page.rs
index e2f7e319..12d05bcf 100644
--- a/crates/typst/src/export/pdf/page.rs
+++ b/crates/typst/src/export/pdf/page.rs
@@ -236,7 +236,7 @@ impl PageContext<'_, '_> {
         let current_state = self.state.external_graphics_state.as_ref();
         if current_state != Some(graphics_state) {
             self.parent.ext_gs_map.insert(*graphics_state);
-            let name = eco_format!("Gs{}", self.parent.ext_gs_map.map(*graphics_state));
+            let name = eco_format!("Gs{}", self.parent.ext_gs_map.map(graphics_state));
             self.content.set_parameters(Name(name.as_bytes()));
 
             if graphics_state.uses_opacities() {
@@ -288,7 +288,7 @@ impl PageContext<'_, '_> {
     fn set_font(&mut self, font: &Font, size: Abs) {
         if self.state.font.as_ref().map(|(f, s)| (f, *s)) != Some((font, size)) {
             self.parent.font_map.insert(font.clone());
-            let name = eco_format!("F{}", self.parent.font_map.map(font.clone()));
+            let name = eco_format!("F{}", self.parent.font_map.map(font));
             self.content.set_font(Name(name.as_bytes()), size.to_f32());
             self.state.font = Some((font.clone(), size));
         }
@@ -472,8 +472,9 @@ fn write_text(ctx: &mut PageContext, x: f32, y: f32, text: &TextItem) {
             adjustment = Em::zero();
         }
 
-        encoded.push((glyph.id >> 8) as u8);
-        encoded.push((glyph.id & 0xff) as u8);
+        let cid = super::font::glyph_cid(&text.font, glyph.id);
+        encoded.push((cid >> 8) as u8);
+        encoded.push((cid & 0xff) as u8);
 
         if let Some(advance) = text.font.advance(glyph.id) {
             adjustment += glyph.x_advance - advance;
@@ -568,7 +569,7 @@ fn write_path(ctx: &mut PageContext, x: f32, y: f32, path: &geom::Path) {
 /// Encode a vector or raster image into the content stream.
 fn write_image(ctx: &mut PageContext, x: f32, y: f32, image: &Image, size: Size) {
     ctx.parent.image_map.insert(image.clone());
-    let name = eco_format!("Im{}", ctx.parent.image_map.map(image.clone()));
+    let name = eco_format!("Im{}", ctx.parent.image_map.map(image));
     let w = size.x.to_f32();
     let h = size.y.to_f32();
     ctx.content.save_state();