From f2882bf85476cd55eaad0594d983fbcf8a1ef778 Mon Sep 17 00:00:00 2001
From: Laurenz <laurmaedje@gmail.com>
Date: Wed, 25 Aug 2021 13:24:30 +0200
Subject: [PATCH] Support decoding of mac roman names

This allows discovery of Apple fonts without unicode name entries.
---
 src/export/pdf.rs |  9 ++-------
 src/font.rs       | 31 ++++++++++++++++++++++---------
 src/util/mac.rs   | 25 +++++++++++++++++++++++++
 src/util/mod.rs   |  2 ++
 4 files changed, 51 insertions(+), 16 deletions(-)
 create mode 100644 src/util/mac.rs

diff --git a/src/export/pdf.rs b/src/export/pdf.rs
index 9b9d0e8e6..65c922ab5 100644
--- a/src/export/pdf.rs
+++ b/src/export/pdf.rs
@@ -15,7 +15,7 @@ use ttf_parser::{name_id, GlyphId, Tag};
 
 use super::subset;
 use crate::color::Color;
-use crate::font::{FaceId, FontStore};
+use crate::font::{find_name, FaceId, FontStore};
 use crate::geom::{self, Em, Length, Size};
 use crate::image::{Image, ImageId, ImageStore};
 use crate::layout::{Element, Frame, Geometry, Paint};
@@ -289,12 +289,7 @@ impl<'a> PdfExporter<'a> {
             let face = self.fonts.get(face_id);
             let ttf = face.ttf();
 
-            let name = ttf
-                .names()
-                .find(|entry| {
-                    entry.name_id() == name_id::POST_SCRIPT_NAME && entry.is_unicode()
-                })
-                .and_then(|entry| entry.to_string())
+            let name = find_name(ttf.names(), name_id::POST_SCRIPT_NAME)
                 .unwrap_or_else(|| "unknown".to_string());
 
             let base_font = format!("ABCDEF+{}", name);
diff --git a/src/font.rs b/src/font.rs
index 633a1a0e9..00bff02ea 100644
--- a/src/font.rs
+++ b/src/font.rs
@@ -6,10 +6,11 @@ use std::path::{Path, PathBuf};
 use std::rc::Rc;
 
 use serde::{Deserialize, Serialize};
-use ttf_parser::{name_id, GlyphId};
+use ttf_parser::{name_id, GlyphId, PlatformId};
 
 use crate::geom::Em;
 use crate::loading::{FileHash, Loader};
+use crate::util::decode_mac_roman;
 
 /// A unique identifier for a loaded font face.
 #[derive(Debug, Copy, Clone, Eq, PartialEq, Hash, Serialize, Deserialize)]
@@ -352,16 +353,11 @@ impl FaceInfo {
         data: &'a [u8],
     ) -> impl Iterator<Item = FaceInfo> + 'a {
         let count = ttf_parser::fonts_in_collection(data).unwrap_or(1);
-        (0 .. count).filter_map(move |index| {
-            fn find_name(face: &ttf_parser::Face, name_id: u16) -> Option<String> {
-                face.names().find_map(|entry| {
-                    (entry.name_id() == name_id).then(|| entry.to_string()).flatten()
-                })
-            }
 
+        (0 .. count).filter_map(move |index| {
             let face = ttf_parser::Face::from_slice(data, index).ok()?;
-            let family = find_name(&face, name_id::TYPOGRAPHIC_FAMILY)
-                .or_else(|| find_name(&face, name_id::FAMILY))?;
+            let family = find_name(face.names(), name_id::TYPOGRAPHIC_FAMILY)
+                .or_else(|| find_name(face.names(), name_id::FAMILY))?;
 
             let variant = FontVariant {
                 style: match (face.is_italic(), face.is_oblique()) {
@@ -383,6 +379,23 @@ impl FaceInfo {
     }
 }
 
+/// Find a decodable entry in a name table iterator.
+pub fn find_name(mut names: ttf_parser::Names<'_>, name_id: u16) -> Option<String> {
+    names.find_map(|entry| {
+        if entry.name_id() == name_id {
+            if let Some(string) = entry.to_string() {
+                return Some(string);
+            }
+
+            if entry.platform_id() == PlatformId::Macintosh && entry.encoding_id() == 0 {
+                return Some(decode_mac_roman(entry.name()));
+            }
+        }
+
+        None
+    })
+}
+
 /// Properties that distinguish a face from other faces in the same family.
 #[derive(Debug, Default, Copy, Clone, Eq, PartialEq, Hash)]
 #[derive(Serialize, Deserialize)]
diff --git a/src/util/mac.rs b/src/util/mac.rs
new file mode 100644
index 000000000..95e8fcd68
--- /dev/null
+++ b/src/util/mac.rs
@@ -0,0 +1,25 @@
+/// Decode mac roman encoded bytes into a string.
+pub fn decode_mac_roman(coded: &[u8]) -> String {
+    coded.iter().copied().map(char_from_mac_roman).collect()
+}
+
+/// Convert a mac roman coded character to a unicode char.
+fn char_from_mac_roman(code: u8) -> char {
+    #[rustfmt::skip]
+    const TABLE: [char; 128] = [
+        'Ä', 'Å', 'Ç', 'É', 'Ñ', 'Ö', 'Ü', 'á', 'à', 'â', 'ä', 'ã', 'å', 'ç', 'é', 'è',
+        'ê', 'ë', 'í', 'ì', 'î', 'ï', 'ñ', 'ó', 'ò', 'ô', 'ö', 'õ', 'ú', 'ù', 'û', 'ü',
+        '†', '°', '¢', '£', '§', '•', '¶', 'ß', '®', '©', '™', '´', '¨', '≠', 'Æ', 'Ø',
+        '∞', '±', '≤', '≥', '¥', 'µ', '∂', '∑', '∏', 'π', '∫', 'ª', 'º', 'Ω', 'æ', 'ø',
+        '¿', '¡', '¬', '√', 'ƒ', '≈', '∆', '«', '»', '…', '\u{a0}', 'À', 'Ã', 'Õ', 'Œ', 'œ',
+        '–', '—', '“', '”', '‘', '’', '÷', '◊', 'ÿ', 'Ÿ', '⁄', '€', '‹', '›', 'fi', 'fl',
+        '‡', '·', '‚', '„', '‰', 'Â', 'Ê', 'Á', 'Ë', 'È', 'Í', 'Î', 'Ï', 'Ì', 'Ó', 'Ô',
+        '\u{f8ff}', 'Ò', 'Ú', 'Û', 'Ù', 'ı', 'ˆ', '˜', '¯', '˘', '˙', '˚', '¸', '˝', '˛', 'ˇ',
+    ];
+
+    if code < 128 {
+        code as char
+    } else {
+        TABLE[(code - 128) as usize]
+    }
+}
diff --git a/src/util/mod.rs b/src/util/mod.rs
index 309c1241d..3608ca767 100644
--- a/src/util/mod.rs
+++ b/src/util/mod.rs
@@ -1,8 +1,10 @@
 //! Utilities.
 
 mod eco;
+mod mac;
 
 pub use eco::EcoString;
+pub use mac::decode_mac_roman;
 
 use std::cell::RefMut;
 use std::cmp::Ordering;