pdfboss_encoding/

afm.rs

//! Standard-14 advance-width tables, from the public Adobe Core-14
//! AFM metrics (see ISO 32000-1 Annex D / Adobe's PDF Reference for the
//! standard-14 font names). Widths are in the 1000-unit glyph space
//! used throughout PDF (ISO 32000-1 9.2.4), i.e. thousandths of the
//! font's em square; multiply by `font_size / 1000` for text space.
//!
//! These tables back [`standard_14_width`], used as a fallback advance
//! when a PDF's `/Widths` array is absent or incomplete and the font
//! program itself isn't embedded (font substitution).

/// Strips a PDF font-subset tag (`ISO 32000-1` 9.6.4: six uppercase
/// letters followed by `+`, e.g. `ABCDEF+Times-Bold`) from `base_font`.
fn strip_subset_prefix(base_font: &str) -> &str {
    let bytes = base_font.as_bytes();
    // `> 7`, not `> 6`: the guard must require at least one character after
    // the `+` for the strip to be meaningful, so the degenerate
    // `"ABCDEF+"` (len 7) is left unstripped instead of yielding an empty
    // name. Matches `substitute.rs`'s `strip_subset_prefix` guard exactly.
    if bytes.len() > 7 && bytes[6] == b'+' && bytes[..6].iter().all(u8::is_ascii_uppercase) {
        &base_font[7..]
    } else {
        base_font
    }
}

/// A standard-14 font family (weight/slant resolved separately).
#[derive(Clone, Copy, PartialEq, Eq, Debug)]
enum Family {
    Helvetica,
    Times,
    Courier,
}

/// Recognized family stems, longest first. Ordering matters: a shorter
/// stem that is itself a prefix of a longer one (`times` vs.
/// `timesnewroman` vs. `timesnewromanps`, `courier` vs. `couriernew`)
/// must never shadow the more specific match, or the leftover
/// remainder handed to [`strip_style_tokens`] would contain stem
/// characters instead of style/noise ones. `classify` walks this list
/// in order and takes the first hit.
const FAMILY_STEMS: &[(&str, Family)] = &[
    ("timesnewromanps", Family::Times),
    ("timesnewroman", Family::Times),
    ("couriernew", Family::Courier),
    ("helvetica", Family::Helvetica),
    ("courier", Family::Courier),
    ("arial", Family::Helvetica),
    ("times", Family::Times),
];

/// Style/noise tokens recognized after the family stem, longest first
/// (so, e.g., `psmt` is tried whole rather than stopping at `ps`).
/// `mt`/`ps`/`psmt` are TrueType-alias noise suffixes (`Arial-BoldMT`,
/// `TimesNewRomanPSMT`). Anything left over after greedily stripping
/// these tokens means the name carries a real, metrically distinct
/// qualifier (`Narrow`, `Neue`, `Condensed`, `Light`, `Ten`, ...) and
/// so isn't one of the standard 14.
const STYLE_TOKENS: &[&str] = &[
    "oblique", "regular", "italic", "roman", "bold", "book", "psmt", "ps", "mt",
];

/// Greedily strips [`STYLE_TOKENS`] from the front of `rest`, folding
/// them into `(bold, italic)` flags, until `rest` is fully consumed.
/// Returns `None` the moment a prefix token isn't recognized, i.e. the
/// name has a qualifier this module doesn't tabulate.
fn strip_style_tokens(mut rest: &str) -> Option<(bool, bool)> {
    let mut bold = false;
    let mut italic = false;
    while !rest.is_empty() {
        let tok = *STYLE_TOKENS.iter().find(|&&t| rest.starts_with(t))?;
        match tok {
            "bold" => bold = true,
            "italic" | "oblique" => italic = true,
            _ => {}
        }
        rest = &rest[tok.len()..];
    }
    Some((bold, italic))
}

/// Classifies a (subset-prefix-stripped) `/BaseFont` name into a
/// standard-14 family plus bold/italic flags. The name is lowercased
/// and separator-stripped (`-`, `,`, space) before matching, so
/// `Arial-BoldMT`, `arial-bold`, and `Arial,Bold` all resolve alike.
/// The (separator-stripped) name must then start with one of the
/// [`FAMILY_STEMS`] (recognizing the common PostScript and TrueType
/// aliases: `Arial` -> Helvetica, `TimesNewRoman`/`TimesNewRomanPS` ->
/// Times, `CourierNew` -> Courier), and everything after the stem must
/// fully reduce to [`STYLE_TOKENS`] noise.
///
/// This is deliberately strict: a font that merely shares a standard-14
/// prefix but is metrically distinct — `Helvetica-Narrow`, `Helvetica
/// Neue`, `HelveticaNeue-Bold`, `TimesTen-Bold`, `Helvetica-Light`,
/// `Times New Roman Condensed` — leaves an unrecognized token after the
/// stem and so classifies as `None`, same as Symbol, ZapfDingbats, or
/// any other name that isn't one of the standard 14. Returning `None`
/// here is correct even though it's plausible-looking: the caller then
/// falls back to the substitute face's own `hmtx` widths rather than a
/// wrong Core-14 advance.
fn classify(base_font: &str) -> Option<(Family, bool, bool)> {
    let stripped = strip_subset_prefix(base_font);
    let lower = stripped.to_lowercase();
    let compact: String = lower
        .chars()
        .filter(|c| !matches!(c, '-' | ',' | ' '))
        .collect();
    let (family, rest) = FAMILY_STEMS
        .iter()
        .find_map(|&(stem, family)| compact.strip_prefix(stem).map(|rest| (family, rest)))?;
    let (bold, italic) = strip_style_tokens(rest)?;
    Some((family, bold, italic))
}

/// Binary-searches a `(glyph_name, width)` table sorted by name.
fn lookup(table: &[(&str, u16)], glyph_name: &str) -> Option<f32> {
    table
        .binary_search_by_key(&glyph_name, |&(name, _)| name)
        .ok()
        .map(|i| table[i].1 as f32)
}

/// The advance width (1000-unit glyph space) of `glyph_name` in the
/// standard-14 font whose PDF `/BaseFont` (or a recognized alias, or a
/// subsetted variant) is `base_font`, or `None` if `base_font` isn't one
/// of the standard 14 (including Symbol/ZapfDingbats, which this module
/// doesn't tabulate) or `glyph_name` is unknown to that font. Courier's
/// four variants are monospaced (every glyph is 600 units wide per the
/// Adobe Core-14 AFM metrics) and so always return `Some(600.0)`.
/// Oblique variants share their upright counterpart's widths (Helvetica
/// and Helvetica-Oblique are metrically identical, likewise the bold
/// pair), so only upright/bold tables are stored.
pub fn standard_14_width(base_font: &str, glyph_name: &str) -> Option<f32> {
    let (family, bold, italic) = classify(base_font)?;
    match family {
        Family::Courier => Some(600.0),
        Family::Helvetica => {
            let table = if bold { HELVETICA_BOLD } else { HELVETICA };
            lookup(table, glyph_name)
        }
        Family::Times => {
            let table = match (bold, italic) {
                (true, true) => TIMES_BOLD_ITALIC,
                (true, false) => TIMES_BOLD,
                (false, true) => TIMES_ITALIC,
                (false, false) => TIMES_ROMAN,
            };
            lookup(table, glyph_name)
        }
    }
}

/// Whether `base_font` (after subset-prefix stripping and alias
/// normalization) names one of the standard 14 fonts.
pub fn is_standard_14(base_font: &str) -> bool {
    classify(base_font).is_some()
}

/// `Helvetica` glyph widths, sorted by
/// glyph name for binary search (Adobe Core-14 AFM metrics).
#[rustfmt::skip]
const HELVETICA: &[(&str, u16)] = &[
    ("A", 667),
    ("AE", 1000),
    ("Aacute", 667),
    ("Acircumflex", 667),
    ("Adieresis", 667),
    ("Agrave", 667),
    ("Aring", 667),
    ("Atilde", 667),
    ("B", 667),
    ("C", 722),
    ("Ccedilla", 722),
    ("D", 722),
    ("E", 667),
    ("Eacute", 667),
    ("Ecircumflex", 667),
    ("Edieresis", 667),
    ("Egrave", 667),
    ("Eth", 722),
    ("F", 611),
    ("G", 778),
    ("H", 722),
    ("I", 278),
    ("Iacute", 278),
    ("Icircumflex", 278),
    ("Idieresis", 278),
    ("Igrave", 278),
    ("J", 500),
    ("K", 667),
    ("L", 556),
    ("Lslash", 556),
    ("M", 833),
    ("N", 722),
    ("Ntilde", 722),
    ("O", 778),
    ("OE", 1000),
    ("Oacute", 778),
    ("Ocircumflex", 778),
    ("Odieresis", 778),
    ("Ograve", 778),
    ("Oslash", 778),
    ("Otilde", 778),
    ("P", 667),
    ("Q", 778),
    ("R", 722),
    ("S", 667),
    ("Scaron", 667),
    ("T", 611),
    ("Thorn", 667),
    ("U", 722),
    ("Uacute", 722),
    ("Ucircumflex", 722),
    ("Udieresis", 722),
    ("Ugrave", 722),
    ("V", 667),
    ("W", 944),
    ("X", 667),
    ("Y", 667),
    ("Yacute", 667),
    ("Ydieresis", 667),
    ("Z", 611),
    ("Zcaron", 611),
    ("a", 556),
    ("aacute", 556),
    ("acircumflex", 556),
    ("acute", 333),
    ("adieresis", 556),
    ("ae", 889),
    ("agrave", 556),
    ("ampersand", 667),
    ("aring", 556),
    ("asciicircum", 469),
    ("asciitilde", 584),
    ("asterisk", 389),
    ("at", 1015),
    ("atilde", 556),
    ("b", 556),
    ("backslash", 278),
    ("bar", 260),
    ("braceleft", 334),
    ("braceright", 334),
    ("bracketleft", 278),
    ("bracketright", 278),
    ("breve", 333),
    ("brokenbar", 260),
    ("bullet", 350),
    ("c", 500),
    ("caron", 333),
    ("ccedilla", 500),
    ("cedilla", 333),
    ("cent", 556),
    ("circumflex", 333),
    ("colon", 278),
    ("comma", 278),
    ("copyright", 737),
    ("currency", 556),
    ("d", 556),
    ("dagger", 556),
    ("daggerdbl", 556),
    ("degree", 400),
    ("dieresis", 333),
    ("divide", 584),
    ("dollar", 556),
    ("dotaccent", 333),
    ("dotlessi", 278),
    ("e", 556),
    ("eacute", 556),
    ("ecircumflex", 556),
    ("edieresis", 556),
    ("egrave", 556),
    ("eight", 556),
    ("ellipsis", 1000),
    ("emdash", 1000),
    ("endash", 556),
    ("equal", 584),
    ("eth", 556),
    ("exclam", 278),
    ("exclamdown", 333),
    ("f", 278),
    ("fi", 500),
    ("five", 556),
    ("fl", 500),
    ("florin", 556),
    ("four", 556),
    ("fraction", 167),
    ("g", 556),
    ("germandbls", 611),
    ("grave", 333),
    ("greater", 584),
    ("guillemotleft", 556),
    ("guillemotright", 556),
    ("guilsinglleft", 333),
    ("guilsinglright", 333),
    ("h", 556),
    ("hungarumlaut", 333),
    ("hyphen", 333),
    ("i", 222),
    ("iacute", 278),
    ("icircumflex", 278),
    ("idieresis", 278),
    ("igrave", 278),
    ("j", 222),
    ("k", 500),
    ("l", 222),
    ("less", 584),
    ("logicalnot", 584),
    ("lslash", 222),
    ("m", 833),
    ("macron", 333),
    ("minus", 584),
    ("mu", 556),
    ("multiply", 584),
    ("n", 556),
    ("nine", 556),
    ("ntilde", 556),
    ("numbersign", 556),
    ("o", 556),
    ("oacute", 556),
    ("ocircumflex", 556),
    ("odieresis", 556),
    ("oe", 944),
    ("ogonek", 333),
    ("ograve", 556),
    ("one", 556),
    ("onehalf", 834),
    ("onequarter", 834),
    ("onesuperior", 333),
    ("ordfeminine", 370),
    ("ordmasculine", 365),
    ("oslash", 611),
    ("otilde", 556),
    ("p", 556),
    ("paragraph", 537),
    ("parenleft", 333),
    ("parenright", 333),
    ("percent", 889),
    ("period", 278),
    ("periodcentered", 278),
    ("perthousand", 1000),
    ("plus", 584),
    ("plusminus", 584),
    ("q", 556),
    ("question", 556),
    ("questiondown", 611),
    ("quotedbl", 355),
    ("quotedblbase", 333),
    ("quotedblleft", 333),
    ("quotedblright", 333),
    ("quoteleft", 222),
    ("quoteright", 222),
    ("quotesinglbase", 222),
    ("quotesingle", 191),
    ("r", 333),
    ("registered", 737),
    ("ring", 333),
    ("s", 500),
    ("scaron", 500),
    ("section", 556),
    ("semicolon", 278),
    ("seven", 556),
    ("six", 556),
    ("slash", 278),
    ("space", 278),
    ("sterling", 556),
    ("t", 278),
    ("thorn", 556),
    ("three", 556),
    ("threequarters", 834),
    ("threesuperior", 333),
    ("tilde", 333),
    ("trademark", 1000),
    ("two", 556),
    ("twosuperior", 333),
    ("u", 556),
    ("uacute", 556),
    ("ucircumflex", 556),
    ("udieresis", 556),
    ("ugrave", 556),
    ("underscore", 556),
    ("v", 500),
    ("w", 722),
    ("x", 500),
    ("y", 500),
    ("yacute", 500),
    ("ydieresis", 500),
    ("yen", 556),
    ("z", 500),
    ("zcaron", 500),
    ("zero", 556),
];

/// `Helvetica Bold` glyph widths, sorted by
/// glyph name for binary search (Adobe Core-14 AFM metrics).
#[rustfmt::skip]
const HELVETICA_BOLD: &[(&str, u16)] = &[
    ("A", 722),
    ("AE", 1000),
    ("Aacute", 722),
    ("Acircumflex", 722),
    ("Adieresis", 722),
    ("Agrave", 722),
    ("Aring", 722),
    ("Atilde", 722),
    ("B", 722),
    ("C", 722),
    ("Ccedilla", 722),
    ("D", 722),
    ("E", 667),
    ("Eacute", 667),
    ("Ecircumflex", 667),
    ("Edieresis", 667),
    ("Egrave", 667),
    ("Eth", 722),
    ("F", 611),
    ("G", 778),
    ("H", 722),
    ("I", 278),
    ("Iacute", 278),
    ("Icircumflex", 278),
    ("Idieresis", 278),
    ("Igrave", 278),
    ("J", 556),
    ("K", 722),
    ("L", 611),
    ("Lslash", 611),
    ("M", 833),
    ("N", 722),
    ("Ntilde", 722),
    ("O", 778),
    ("OE", 1000),
    ("Oacute", 778),
    ("Ocircumflex", 778),
    ("Odieresis", 778),
    ("Ograve", 778),
    ("Oslash", 778),
    ("Otilde", 778),
    ("P", 667),
    ("Q", 778),
    ("R", 722),
    ("S", 667),
    ("Scaron", 667),
    ("T", 611),
    ("Thorn", 667),
    ("U", 722),
    ("Uacute", 722),
    ("Ucircumflex", 722),
    ("Udieresis", 722),
    ("Ugrave", 722),
    ("V", 667),
    ("W", 944),
    ("X", 667),
    ("Y", 667),
    ("Yacute", 667),
    ("Ydieresis", 667),
    ("Z", 611),
    ("Zcaron", 611),
    ("a", 556),
    ("aacute", 556),
    ("acircumflex", 556),
    ("acute", 333),
    ("adieresis", 556),
    ("ae", 889),
    ("agrave", 556),
    ("ampersand", 722),
    ("aring", 556),
    ("asciicircum", 584),
    ("asciitilde", 584),
    ("asterisk", 389),
    ("at", 975),
    ("atilde", 556),
    ("b", 611),
    ("backslash", 278),
    ("bar", 280),
    ("braceleft", 389),
    ("braceright", 389),
    ("bracketleft", 333),
    ("bracketright", 333),
    ("breve", 333),
    ("brokenbar", 280),
    ("bullet", 350),
    ("c", 556),
    ("caron", 333),
    ("ccedilla", 556),
    ("cedilla", 333),
    ("cent", 556),
    ("circumflex", 333),
    ("colon", 333),
    ("comma", 278),
    ("copyright", 737),
    ("currency", 556),
    ("d", 611),
    ("dagger", 556),
    ("daggerdbl", 556),
    ("degree", 400),
    ("dieresis", 333),
    ("divide", 584),
    ("dollar", 556),
    ("dotaccent", 333),
    ("dotlessi", 278),
    ("e", 556),
    ("eacute", 556),
    ("ecircumflex", 556),
    ("edieresis", 556),
    ("egrave", 556),
    ("eight", 556),
    ("ellipsis", 1000),
    ("emdash", 1000),
    ("endash", 556),
    ("equal", 584),
    ("eth", 611),
    ("exclam", 333),
    ("exclamdown", 333),
    ("f", 333),
    ("fi", 611),
    ("five", 556),
    ("fl", 611),
    ("florin", 556),
    ("four", 556),
    ("fraction", 167),
    ("g", 611),
    ("germandbls", 611),
    ("grave", 333),
    ("greater", 584),
    ("guillemotleft", 556),
    ("guillemotright", 556),
    ("guilsinglleft", 333),
    ("guilsinglright", 333),
    ("h", 611),
    ("hungarumlaut", 333),
    ("hyphen", 333),
    ("i", 278),
    ("iacute", 278),
    ("icircumflex", 278),
    ("idieresis", 278),
    ("igrave", 278),
    ("j", 278),
    ("k", 556),
    ("l", 278),
    ("less", 584),
    ("logicalnot", 584),
    ("lslash", 278),
    ("m", 889),
    ("macron", 333),
    ("minus", 584),
    ("mu", 611),
    ("multiply", 584),
    ("n", 611),
    ("nine", 556),
    ("ntilde", 611),
    ("numbersign", 556),
    ("o", 611),
    ("oacute", 611),
    ("ocircumflex", 611),
    ("odieresis", 611),
    ("oe", 944),
    ("ogonek", 333),
    ("ograve", 611),
    ("one", 556),
    ("onehalf", 834),
    ("onequarter", 834),
    ("onesuperior", 333),
    ("ordfeminine", 370),
    ("ordmasculine", 365),
    ("oslash", 611),
    ("otilde", 611),
    ("p", 611),
    ("paragraph", 556),
    ("parenleft", 333),
    ("parenright", 333),
    ("percent", 889),
    ("period", 278),
    ("periodcentered", 278),
    ("perthousand", 1000),
    ("plus", 584),
    ("plusminus", 584),
    ("q", 611),
    ("question", 611),
    ("questiondown", 611),
    ("quotedbl", 474),
    ("quotedblbase", 500),
    ("quotedblleft", 500),
    ("quotedblright", 500),
    ("quoteleft", 278),
    ("quoteright", 278),
    ("quotesinglbase", 278),
    ("quotesingle", 238),
    ("r", 389),
    ("registered", 737),
    ("ring", 333),
    ("s", 556),
    ("scaron", 556),
    ("section", 556),
    ("semicolon", 333),
    ("seven", 556),
    ("six", 556),
    ("slash", 278),
    ("space", 278),
    ("sterling", 556),
    ("t", 333),
    ("thorn", 611),
    ("three", 556),
    ("threequarters", 834),
    ("threesuperior", 333),
    ("tilde", 333),
    ("trademark", 1000),
    ("two", 556),
    ("twosuperior", 333),
    ("u", 611),
    ("uacute", 611),
    ("ucircumflex", 611),
    ("udieresis", 611),
    ("ugrave", 611),
    ("underscore", 556),
    ("v", 556),
    ("w", 778),
    ("x", 556),
    ("y", 556),
    ("yacute", 556),
    ("ydieresis", 556),
    ("yen", 556),
    ("z", 500),
    ("zcaron", 500),
    ("zero", 556),
];

/// `Times Roman` glyph widths, sorted by
/// glyph name for binary search (Adobe Core-14 AFM metrics).
#[rustfmt::skip]
const TIMES_ROMAN: &[(&str, u16)] = &[
    ("A", 722),
    ("AE", 889),
    ("Aacute", 722),
    ("Acircumflex", 722),
    ("Adieresis", 722),
    ("Agrave", 722),
    ("Aring", 722),
    ("Atilde", 722),
    ("B", 667),
    ("C", 667),
    ("Ccedilla", 667),
    ("D", 722),
    ("E", 611),
    ("Eacute", 611),
    ("Ecircumflex", 611),
    ("Edieresis", 611),
    ("Egrave", 611),
    ("Eth", 722),
    ("F", 556),
    ("G", 722),
    ("H", 722),
    ("I", 333),
    ("Iacute", 333),
    ("Icircumflex", 333),
    ("Idieresis", 333),
    ("Igrave", 333),
    ("J", 389),
    ("K", 722),
    ("L", 611),
    ("Lslash", 611),
    ("M", 889),
    ("N", 722),
    ("Ntilde", 722),
    ("O", 722),
    ("OE", 889),
    ("Oacute", 722),
    ("Ocircumflex", 722),
    ("Odieresis", 722),
    ("Ograve", 722),
    ("Oslash", 722),
    ("Otilde", 722),
    ("P", 556),
    ("Q", 722),
    ("R", 667),
    ("S", 556),
    ("Scaron", 556),
    ("T", 611),
    ("Thorn", 556),
    ("U", 722),
    ("Uacute", 722),
    ("Ucircumflex", 722),
    ("Udieresis", 722),
    ("Ugrave", 722),
    ("V", 722),
    ("W", 944),
    ("X", 722),
    ("Y", 722),
    ("Yacute", 722),
    ("Ydieresis", 722),
    ("Z", 611),
    ("Zcaron", 611),
    ("a", 444),
    ("aacute", 444),
    ("acircumflex", 444),
    ("acute", 333),
    ("adieresis", 444),
    ("ae", 667),
    ("agrave", 444),
    ("ampersand", 778),
    ("aring", 444),
    ("asciicircum", 469),
    ("asciitilde", 541),
    ("asterisk", 500),
    ("at", 921),
    ("atilde", 444),
    ("b", 500),
    ("backslash", 278),
    ("bar", 200),
    ("braceleft", 480),
    ("braceright", 480),
    ("bracketleft", 333),
    ("bracketright", 333),
    ("breve", 333),
    ("brokenbar", 200),
    ("bullet", 350),
    ("c", 444),
    ("caron", 333),
    ("ccedilla", 444),
    ("cedilla", 333),
    ("cent", 500),
    ("circumflex", 333),
    ("colon", 278),
    ("comma", 250),
    ("copyright", 760),
    ("currency", 500),
    ("d", 500),
    ("dagger", 500),
    ("daggerdbl", 500),
    ("degree", 400),
    ("dieresis", 333),
    ("divide", 564),
    ("dollar", 500),
    ("dotaccent", 333),
    ("dotlessi", 278),
    ("e", 444),
    ("eacute", 444),
    ("ecircumflex", 444),
    ("edieresis", 444),
    ("egrave", 444),
    ("eight", 500),
    ("ellipsis", 1000),
    ("emdash", 1000),
    ("endash", 500),
    ("equal", 564),
    ("eth", 500),
    ("exclam", 333),
    ("exclamdown", 333),
    ("f", 333),
    ("fi", 556),
    ("five", 500),
    ("fl", 556),
    ("florin", 500),
    ("four", 500),
    ("fraction", 167),
    ("g", 500),
    ("germandbls", 500),
    ("grave", 333),
    ("greater", 564),
    ("guillemotleft", 500),
    ("guillemotright", 500),
    ("guilsinglleft", 333),
    ("guilsinglright", 333),
    ("h", 500),
    ("hungarumlaut", 333),
    ("hyphen", 333),
    ("i", 278),
    ("iacute", 278),
    ("icircumflex", 278),
    ("idieresis", 278),
    ("igrave", 278),
    ("j", 278),
    ("k", 500),
    ("l", 278),
    ("less", 564),
    ("logicalnot", 564),
    ("lslash", 278),
    ("m", 778),
    ("macron", 333),
    ("minus", 564),
    ("mu", 500),
    ("multiply", 564),
    ("n", 500),
    ("nine", 500),
    ("ntilde", 500),
    ("numbersign", 500),
    ("o", 500),
    ("oacute", 500),
    ("ocircumflex", 500),
    ("odieresis", 500),
    ("oe", 722),
    ("ogonek", 333),
    ("ograve", 500),
    ("one", 500),
    ("onehalf", 750),
    ("onequarter", 750),
    ("onesuperior", 300),
    ("ordfeminine", 276),
    ("ordmasculine", 310),
    ("oslash", 500),
    ("otilde", 500),
    ("p", 500),
    ("paragraph", 453),
    ("parenleft", 333),
    ("parenright", 333),
    ("percent", 833),
    ("period", 250),
    ("periodcentered", 250),
    ("perthousand", 1000),
    ("plus", 564),
    ("plusminus", 564),
    ("q", 500),
    ("question", 444),
    ("questiondown", 444),
    ("quotedbl", 408),
    ("quotedblbase", 444),
    ("quotedblleft", 444),
    ("quotedblright", 444),
    ("quoteleft", 333),
    ("quoteright", 333),
    ("quotesinglbase", 333),
    ("quotesingle", 180),
    ("r", 333),
    ("registered", 760),
    ("ring", 333),
    ("s", 389),
    ("scaron", 389),
    ("section", 500),
    ("semicolon", 278),
    ("seven", 500),
    ("six", 500),
    ("slash", 278),
    ("space", 250),
    ("sterling", 500),
    ("t", 278),
    ("thorn", 500),
    ("three", 500),
    ("threequarters", 750),
    ("threesuperior", 300),
    ("tilde", 333),
    ("trademark", 980),
    ("two", 500),
    ("twosuperior", 300),
    ("u", 500),
    ("uacute", 500),
    ("ucircumflex", 500),
    ("udieresis", 500),
    ("ugrave", 500),
    ("underscore", 500),
    ("v", 500),
    ("w", 722),
    ("x", 500),
    ("y", 500),
    ("yacute", 500),
    ("ydieresis", 500),
    ("yen", 500),
    ("z", 444),
    ("zcaron", 444),
    ("zero", 500),
];

/// `Times Bold` glyph widths, sorted by
/// glyph name for binary search (Adobe Core-14 AFM metrics).
#[rustfmt::skip]
const TIMES_BOLD: &[(&str, u16)] = &[
    ("A", 722),
    ("AE", 1000),
    ("Aacute", 722),
    ("Acircumflex", 722),
    ("Adieresis", 722),
    ("Agrave", 722),
    ("Aring", 722),
    ("Atilde", 722),
    ("B", 667),
    ("C", 722),
    ("Ccedilla", 722),
    ("D", 722),
    ("E", 667),
    ("Eacute", 667),
    ("Ecircumflex", 667),
    ("Edieresis", 667),
    ("Egrave", 667),
    ("Eth", 722),
    ("F", 611),
    ("G", 778),
    ("H", 778),
    ("I", 389),
    ("Iacute", 389),
    ("Icircumflex", 389),
    ("Idieresis", 389),
    ("Igrave", 389),
    ("J", 500),
    ("K", 778),
    ("L", 667),
    ("Lslash", 667),
    ("M", 944),
    ("N", 722),
    ("Ntilde", 722),
    ("O", 778),
    ("OE", 1000),
    ("Oacute", 778),
    ("Ocircumflex", 778),
    ("Odieresis", 778),
    ("Ograve", 778),
    ("Oslash", 778),
    ("Otilde", 778),
    ("P", 611),
    ("Q", 778),
    ("R", 722),
    ("S", 556),
    ("Scaron", 556),
    ("T", 667),
    ("Thorn", 611),
    ("U", 722),
    ("Uacute", 722),
    ("Ucircumflex", 722),
    ("Udieresis", 722),
    ("Ugrave", 722),
    ("V", 722),
    ("W", 1000),
    ("X", 722),
    ("Y", 722),
    ("Yacute", 722),
    ("Ydieresis", 722),
    ("Z", 667),
    ("Zcaron", 667),
    ("a", 500),
    ("aacute", 500),
    ("acircumflex", 500),
    ("acute", 333),
    ("adieresis", 500),
    ("ae", 722),
    ("agrave", 500),
    ("ampersand", 833),
    ("aring", 500),
    ("asciicircum", 581),
    ("asciitilde", 520),
    ("asterisk", 500),
    ("at", 930),
    ("atilde", 500),
    ("b", 556),
    ("backslash", 278),
    ("bar", 220),
    ("braceleft", 394),
    ("braceright", 394),
    ("bracketleft", 333),
    ("bracketright", 333),
    ("breve", 333),
    ("brokenbar", 220),
    ("bullet", 350),
    ("c", 444),
    ("caron", 333),
    ("ccedilla", 444),
    ("cedilla", 333),
    ("cent", 500),
    ("circumflex", 333),
    ("colon", 333),
    ("comma", 250),
    ("copyright", 747),
    ("currency", 500),
    ("d", 556),
    ("dagger", 500),
    ("daggerdbl", 500),
    ("degree", 400),
    ("dieresis", 333),
    ("divide", 570),
    ("dollar", 500),
    ("dotaccent", 333),
    ("dotlessi", 278),
    ("e", 444),
    ("eacute", 444),
    ("ecircumflex", 444),
    ("edieresis", 444),
    ("egrave", 444),
    ("eight", 500),
    ("ellipsis", 1000),
    ("emdash", 1000),
    ("endash", 500),
    ("equal", 570),
    ("eth", 500),
    ("exclam", 333),
    ("exclamdown", 333),
    ("f", 333),
    ("fi", 556),
    ("five", 500),
    ("fl", 556),
    ("florin", 500),
    ("four", 500),
    ("fraction", 167),
    ("g", 500),
    ("germandbls", 556),
    ("grave", 333),
    ("greater", 570),
    ("guillemotleft", 500),
    ("guillemotright", 500),
    ("guilsinglleft", 333),
    ("guilsinglright", 333),
    ("h", 556),
    ("hungarumlaut", 333),
    ("hyphen", 333),
    ("i", 278),
    ("iacute", 278),
    ("icircumflex", 278),
    ("idieresis", 278),
    ("igrave", 278),
    ("j", 333),
    ("k", 556),
    ("l", 278),
    ("less", 570),
    ("logicalnot", 570),
    ("lslash", 278),
    ("m", 833),
    ("macron", 333),
    ("minus", 570),
    ("mu", 556),
    ("multiply", 570),
    ("n", 556),
    ("nine", 500),
    ("ntilde", 556),
    ("numbersign", 500),
    ("o", 500),
    ("oacute", 500),
    ("ocircumflex", 500),
    ("odieresis", 500),
    ("oe", 722),
    ("ogonek", 333),
    ("ograve", 500),
    ("one", 500),
    ("onehalf", 750),
    ("onequarter", 750),
    ("onesuperior", 300),
    ("ordfeminine", 300),
    ("ordmasculine", 330),
    ("oslash", 500),
    ("otilde", 500),
    ("p", 556),
    ("paragraph", 540),
    ("parenleft", 333),
    ("parenright", 333),
    ("percent", 1000),
    ("period", 250),
    ("periodcentered", 250),
    ("perthousand", 1000),
    ("plus", 570),
    ("plusminus", 570),
    ("q", 556),
    ("question", 500),
    ("questiondown", 500),
    ("quotedbl", 555),
    ("quotedblbase", 500),
    ("quotedblleft", 500),
    ("quotedblright", 500),
    ("quoteleft", 333),
    ("quoteright", 333),
    ("quotesinglbase", 333),
    ("quotesingle", 278),
    ("r", 444),
    ("registered", 747),
    ("ring", 333),
    ("s", 389),
    ("scaron", 389),
    ("section", 500),
    ("semicolon", 333),
    ("seven", 500),
    ("six", 500),
    ("slash", 278),
    ("space", 250),
    ("sterling", 500),
    ("t", 333),
    ("thorn", 556),
    ("three", 500),
    ("threequarters", 750),
    ("threesuperior", 300),
    ("tilde", 333),
    ("trademark", 1000),
    ("two", 500),
    ("twosuperior", 300),
    ("u", 556),
    ("uacute", 556),
    ("ucircumflex", 556),
    ("udieresis", 556),
    ("ugrave", 556),
    ("underscore", 500),
    ("v", 500),
    ("w", 722),
    ("x", 500),
    ("y", 500),
    ("yacute", 500),
    ("ydieresis", 500),
    ("yen", 500),
    ("z", 444),
    ("zcaron", 444),
    ("zero", 500),
];

/// `Times Italic` glyph widths, sorted by
/// glyph name for binary search (Adobe Core-14 AFM metrics).
#[rustfmt::skip]
const TIMES_ITALIC: &[(&str, u16)] = &[
    ("A", 611),
    ("AE", 889),
    ("Aacute", 611),
    ("Acircumflex", 611),
    ("Adieresis", 611),
    ("Agrave", 611),
    ("Aring", 611),
    ("Atilde", 611),
    ("B", 611),
    ("C", 667),
    ("Ccedilla", 667),
    ("D", 722),
    ("E", 611),
    ("Eacute", 611),
    ("Ecircumflex", 611),
    ("Edieresis", 611),
    ("Egrave", 611),
    ("Eth", 722),
    ("F", 611),
    ("G", 722),
    ("H", 722),
    ("I", 333),
    ("Iacute", 333),
    ("Icircumflex", 333),
    ("Idieresis", 333),
    ("Igrave", 333),
    ("J", 444),
    ("K", 667),
    ("L", 556),
    ("Lslash", 556),
    ("M", 833),
    ("N", 667),
    ("Ntilde", 667),
    ("O", 722),
    ("OE", 944),
    ("Oacute", 722),
    ("Ocircumflex", 722),
    ("Odieresis", 722),
    ("Ograve", 722),
    ("Oslash", 722),
    ("Otilde", 722),
    ("P", 611),
    ("Q", 722),
    ("R", 611),
    ("S", 500),
    ("Scaron", 500),
    ("T", 556),
    ("Thorn", 611),
    ("U", 722),
    ("Uacute", 722),
    ("Ucircumflex", 722),
    ("Udieresis", 722),
    ("Ugrave", 722),
    ("V", 611),
    ("W", 833),
    ("X", 611),
    ("Y", 556),
    ("Yacute", 556),
    ("Ydieresis", 556),
    ("Z", 556),
    ("Zcaron", 556),
    ("a", 500),
    ("aacute", 500),
    ("acircumflex", 500),
    ("acute", 333),
    ("adieresis", 500),
    ("ae", 667),
    ("agrave", 500),
    ("ampersand", 778),
    ("aring", 500),
    ("asciicircum", 422),
    ("asciitilde", 541),
    ("asterisk", 500),
    ("at", 920),
    ("atilde", 500),
    ("b", 500),
    ("backslash", 278),
    ("bar", 275),
    ("braceleft", 400),
    ("braceright", 400),
    ("bracketleft", 389),
    ("bracketright", 389),
    ("breve", 333),
    ("brokenbar", 275),
    ("bullet", 350),
    ("c", 444),
    ("caron", 333),
    ("ccedilla", 444),
    ("cedilla", 333),
    ("cent", 500),
    ("circumflex", 333),
    ("colon", 333),
    ("comma", 250),
    ("copyright", 760),
    ("currency", 500),
    ("d", 500),
    ("dagger", 500),
    ("daggerdbl", 500),
    ("degree", 400),
    ("dieresis", 333),
    ("divide", 675),
    ("dollar", 500),
    ("dotaccent", 333),
    ("dotlessi", 278),
    ("e", 444),
    ("eacute", 444),
    ("ecircumflex", 444),
    ("edieresis", 444),
    ("egrave", 444),
    ("eight", 500),
    ("ellipsis", 889),
    ("emdash", 889),
    ("endash", 500),
    ("equal", 675),
    ("eth", 500),
    ("exclam", 333),
    ("exclamdown", 389),
    ("f", 278),
    ("fi", 500),
    ("five", 500),
    ("fl", 500),
    ("florin", 500),
    ("four", 500),
    ("fraction", 167),
    ("g", 500),
    ("germandbls", 500),
    ("grave", 333),
    ("greater", 675),
    ("guillemotleft", 500),
    ("guillemotright", 500),
    ("guilsinglleft", 333),
    ("guilsinglright", 333),
    ("h", 500),
    ("hungarumlaut", 333),
    ("hyphen", 333),
    ("i", 278),
    ("iacute", 278),
    ("icircumflex", 278),
    ("idieresis", 278),
    ("igrave", 278),
    ("j", 278),
    ("k", 444),
    ("l", 278),
    ("less", 675),
    ("logicalnot", 675),
    ("lslash", 278),
    ("m", 722),
    ("macron", 333),
    ("minus", 675),
    ("mu", 500),
    ("multiply", 675),
    ("n", 500),
    ("nine", 500),
    ("ntilde", 500),
    ("numbersign", 500),
    ("o", 500),
    ("oacute", 500),
    ("ocircumflex", 500),
    ("odieresis", 500),
    ("oe", 667),
    ("ogonek", 333),
    ("ograve", 500),
    ("one", 500),
    ("onehalf", 750),
    ("onequarter", 750),
    ("onesuperior", 300),
    ("ordfeminine", 276),
    ("ordmasculine", 310),
    ("oslash", 500),
    ("otilde", 500),
    ("p", 500),
    ("paragraph", 523),
    ("parenleft", 333),
    ("parenright", 333),
    ("percent", 833),
    ("period", 250),
    ("periodcentered", 250),
    ("perthousand", 1000),
    ("plus", 675),
    ("plusminus", 675),
    ("q", 500),
    ("question", 500),
    ("questiondown", 500),
    ("quotedbl", 420),
    ("quotedblbase", 556),
    ("quotedblleft", 556),
    ("quotedblright", 556),
    ("quoteleft", 333),
    ("quoteright", 333),
    ("quotesinglbase", 333),
    ("quotesingle", 214),
    ("r", 389),
    ("registered", 760),
    ("ring", 333),
    ("s", 389),
    ("scaron", 389),
    ("section", 500),
    ("semicolon", 333),
    ("seven", 500),
    ("six", 500),
    ("slash", 278),
    ("space", 250),
    ("sterling", 500),
    ("t", 278),
    ("thorn", 500),
    ("three", 500),
    ("threequarters", 750),
    ("threesuperior", 300),
    ("tilde", 333),
    ("trademark", 980),
    ("two", 500),
    ("twosuperior", 300),
    ("u", 500),
    ("uacute", 500),
    ("ucircumflex", 500),
    ("udieresis", 500),
    ("ugrave", 500),
    ("underscore", 500),
    ("v", 444),
    ("w", 667),
    ("x", 444),
    ("y", 444),
    ("yacute", 444),
    ("ydieresis", 444),
    ("yen", 500),
    ("z", 389),
    ("zcaron", 389),
    ("zero", 500),
];

/// `Times Bold Italic` glyph widths, sorted by
/// glyph name for binary search (Adobe Core-14 AFM metrics).
#[rustfmt::skip]
const TIMES_BOLD_ITALIC: &[(&str, u16)] = &[
    ("A", 667),
    ("AE", 944),
    ("Aacute", 667),
    ("Acircumflex", 667),
    ("Adieresis", 667),
    ("Agrave", 667),
    ("Aring", 667),
    ("Atilde", 667),
    ("B", 667),
    ("C", 667),
    ("Ccedilla", 667),
    ("D", 722),
    ("E", 667),
    ("Eacute", 667),
    ("Ecircumflex", 667),
    ("Edieresis", 667),
    ("Egrave", 667),
    ("Eth", 722),
    ("F", 667),
    ("G", 722),
    ("H", 778),
    ("I", 389),
    ("Iacute", 389),
    ("Icircumflex", 389),
    ("Idieresis", 389),
    ("Igrave", 389),
    ("J", 500),
    ("K", 667),
    ("L", 611),
    ("Lslash", 611),
    ("M", 889),
    ("N", 722),
    ("Ntilde", 722),
    ("O", 722),
    ("OE", 944),
    ("Oacute", 722),
    ("Ocircumflex", 722),
    ("Odieresis", 722),
    ("Ograve", 722),
    ("Oslash", 722),
    ("Otilde", 722),
    ("P", 611),
    ("Q", 722),
    ("R", 667),
    ("S", 556),
    ("Scaron", 556),
    ("T", 611),
    ("Thorn", 611),
    ("U", 722),
    ("Uacute", 722),
    ("Ucircumflex", 722),
    ("Udieresis", 722),
    ("Ugrave", 722),
    ("V", 667),
    ("W", 889),
    ("X", 667),
    ("Y", 611),
    ("Yacute", 611),
    ("Ydieresis", 611),
    ("Z", 611),
    ("Zcaron", 611),
    ("a", 500),
    ("aacute", 500),
    ("acircumflex", 500),
    ("acute", 333),
    ("adieresis", 500),
    ("ae", 722),
    ("agrave", 500),
    ("ampersand", 778),
    ("aring", 500),
    ("asciicircum", 570),
    ("asciitilde", 570),
    ("asterisk", 500),
    ("at", 832),
    ("atilde", 500),
    ("b", 500),
    ("backslash", 278),
    ("bar", 220),
    ("braceleft", 348),
    ("braceright", 348),
    ("bracketleft", 333),
    ("bracketright", 333),
    ("breve", 333),
    ("brokenbar", 220),
    ("bullet", 350),
    ("c", 444),
    ("caron", 333),
    ("ccedilla", 444),
    ("cedilla", 333),
    ("cent", 500),
    ("circumflex", 333),
    ("colon", 333),
    ("comma", 250),
    ("copyright", 747),
    ("currency", 500),
    ("d", 500),
    ("dagger", 500),
    ("daggerdbl", 500),
    ("degree", 400),
    ("dieresis", 333),
    ("divide", 570),
    ("dollar", 500),
    ("dotaccent", 333),
    ("dotlessi", 278),
    ("e", 444),
    ("eacute", 444),
    ("ecircumflex", 444),
    ("edieresis", 444),
    ("egrave", 444),
    ("eight", 500),
    ("ellipsis", 1000),
    ("emdash", 1000),
    ("endash", 500),
    ("equal", 570),
    ("eth", 500),
    ("exclam", 389),
    ("exclamdown", 389),
    ("f", 333),
    ("fi", 556),
    ("five", 500),
    ("fl", 556),
    ("florin", 500),
    ("four", 500),
    ("fraction", 167),
    ("g", 500),
    ("germandbls", 500),
    ("grave", 333),
    ("greater", 570),
    ("guillemotleft", 500),
    ("guillemotright", 500),
    ("guilsinglleft", 333),
    ("guilsinglright", 333),
    ("h", 556),
    ("hungarumlaut", 333),
    ("hyphen", 333),
    ("i", 278),
    ("iacute", 278),
    ("icircumflex", 278),
    ("idieresis", 278),
    ("igrave", 278),
    ("j", 278),
    ("k", 500),
    ("l", 278),
    ("less", 570),
    ("logicalnot", 606),
    ("lslash", 278),
    ("m", 778),
    ("macron", 333),
    ("minus", 606),
    ("mu", 576),
    ("multiply", 570),
    ("n", 556),
    ("nine", 500),
    ("ntilde", 556),
    ("numbersign", 500),
    ("o", 500),
    ("oacute", 500),
    ("ocircumflex", 500),
    ("odieresis", 500),
    ("oe", 722),
    ("ogonek", 333),
    ("ograve", 500),
    ("one", 500),
    ("onehalf", 750),
    ("onequarter", 750),
    ("onesuperior", 300),
    ("ordfeminine", 266),
    ("ordmasculine", 300),
    ("oslash", 500),
    ("otilde", 500),
    ("p", 500),
    ("paragraph", 500),
    ("parenleft", 333),
    ("parenright", 333),
    ("percent", 833),
    ("period", 250),
    ("periodcentered", 250),
    ("perthousand", 1000),
    ("plus", 570),
    ("plusminus", 570),
    ("q", 500),
    ("question", 500),
    ("questiondown", 500),
    ("quotedbl", 555),
    ("quotedblbase", 500),
    ("quotedblleft", 500),
    ("quotedblright", 500),
    ("quoteleft", 333),
    ("quoteright", 333),
    ("quotesinglbase", 333),
    ("quotesingle", 278),
    ("r", 389),
    ("registered", 747),
    ("ring", 333),
    ("s", 389),
    ("scaron", 389),
    ("section", 500),
    ("semicolon", 333),
    ("seven", 500),
    ("six", 500),
    ("slash", 278),
    ("space", 250),
    ("sterling", 500),
    ("t", 278),
    ("thorn", 500),
    ("three", 500),
    ("threequarters", 750),
    ("threesuperior", 300),
    ("tilde", 333),
    ("trademark", 1000),
    ("two", 500),
    ("twosuperior", 300),
    ("u", 556),
    ("uacute", 556),
    ("ucircumflex", 556),
    ("udieresis", 556),
    ("ugrave", 556),
    ("underscore", 500),
    ("v", 444),
    ("w", 667),
    ("x", 500),
    ("y", 444),
    ("yacute", 444),
    ("ydieresis", 444),
    ("yen", 500),
    ("z", 389),
    ("zcaron", 389),
    ("zero", 500),
];

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn standard_14_known_widths() {
        assert_eq!(standard_14_width("Helvetica", "space"), Some(278.0));
        assert_eq!(standard_14_width("Helvetica", "A"), Some(667.0));
        assert_eq!(standard_14_width("Helvetica-Bold", "A"), Some(722.0));
        assert_eq!(standard_14_width("Times-Roman", "A"), Some(722.0));
        assert_eq!(standard_14_width("Times-Italic", "a"), Some(500.0));
        // Courier is monospaced: every glyph is 600 units wide.
        assert_eq!(standard_14_width("Courier", "A"), Some(600.0));
        assert_eq!(standard_14_width("Courier-BoldOblique", "m"), Some(600.0));
        // Oblique variants share their upright counterpart's widths.
        assert_eq!(
            standard_14_width("Helvetica-Oblique", "A"),
            standard_14_width("Helvetica", "A")
        );
        assert_eq!(
            standard_14_width("Helvetica-BoldOblique", "A"),
            standard_14_width("Helvetica-Bold", "A")
        );
    }

    #[test]
    fn standard_14_aliases_and_non_members() {
        assert_eq!(standard_14_width("Arial", "space"), Some(278.0)); // Arial -> Helvetica
        assert_eq!(standard_14_width("ABCDEF+Times-Bold", "A"), Some(722.0)); // subset prefix stripped
        assert_eq!(standard_14_width("TimesNewRomanPSMT", "A"), Some(722.0)); // TrueType alias -> Times-Roman
        assert_eq!(standard_14_width("CourierNewPS-BoldMT", "A"), Some(600.0)); // TrueType alias -> Courier-Bold
        assert!(standard_14_width("Symbol", "alpha").is_none()); // out of scope
        assert!(standard_14_width("SomeEmbeddedFont", "A").is_none());
        assert!(is_standard_14("Courier-Bold") && !is_standard_14("Comic Sans"));
        assert!(is_standard_14("Arial-BoldMT"));
        assert!(!is_standard_14("ZapfDingbats"));
    }

    // Review-fix regression: `classify()` must reject metrically-distinct
    // siblings that merely share a standard-14 prefix (see module docs on
    // `classify`), rather than silently returning a wrong-but-plausible
    // Core-14 width.
    #[test]
    fn standard_14_review_fix_positive() {
        assert_eq!(standard_14_width("Helvetica-Bold", "A"), Some(722.0));
        assert_eq!(standard_14_width("Arial-BoldMT", "space"), Some(278.0));
        assert!(standard_14_width("TimesNewRomanPS-BoldItalicMT", "A").is_some());
        assert_eq!(standard_14_width("CourierNew-Bold", "x"), Some(600.0));
        // Internal spaces are handled (stripped before stem matching).
        assert!(standard_14_width("TimesNewRoman", "A").is_some());
        assert!(standard_14_width("Times New Roman", "A").is_some());
    }

    #[test]
    fn standard_14_rejects_metrically_distinct_siblings() {
        assert!(standard_14_width("Helvetica-Narrow", "A").is_none());
        assert!(standard_14_width("HelveticaNeue-Bold", "A").is_none());
        assert!(standard_14_width("Helvetica Neue", "A").is_none());
        assert!(standard_14_width("TimesTen-Bold", "A").is_none());
        assert!(standard_14_width("Helvetica-Light", "A").is_none());
        assert!(standard_14_width("Times New Roman Condensed", "A").is_none());
    }

    #[test]
    fn standard_14_classify_case_folds() {
        // classify() lowercases internally, same as
        // `FaceRequest::from_font_dict`'s normalization.
        assert_eq!(standard_14_width("arial-bold", "A"), Some(722.0));
    }
}