diff --git a/crates/i18n/src/collate.rs b/crates/i18n/src/collate.rs index 06d417897..629b3bf75 100644 --- a/crates/i18n/src/collate.rs +++ b/crates/i18n/src/collate.rs @@ -30,6 +30,7 @@ impl Collator { /// Sorting for `locale`, a language code such as `sv`, `pt-br` or `zh-hant` (any case, `-` /// or `_`, glibc forms such as `sv_SE.UTF-8` and `sr_RS@latin`). A BCP 47 collation type /// picks another order where there is one: `zh-hant-u-co-pinyin`, `zh-hans-u-co-stroke`. + #[must_use] pub fn new(locale: &str) -> Collator { let mut prefs: CollatorPreferences = Locale::try_from_str(&bcp47(locale)).unwrap_or(Locale::UNKNOWN).into(); prefs.numeric_ordering = Some(CollationNumericOrdering::True); @@ -46,6 +47,7 @@ impl Collator { /// the collation counts as equal (`Layer 2` and `Layer 02`) is in code point order, so a sort /// doesn't depend on the input order. Without collation data (never with the embedded data) /// this is code point order. + #[must_use] pub fn compare(&self, a: &str, b: &str) -> Ordering { match &self.0 { Some(c) => c.as_borrowed().compare(a, b).then_with(|| a.cmp(b)), diff --git a/crates/i18n/src/fold.rs b/crates/i18n/src/fold.rs index 959c6425a..dd77b6dfd 100644 --- a/crates/i18n/src/fold.rs +++ b/crates/i18n/src/fold.rs @@ -15,7 +15,7 @@ //! `IRMAK` and `ırmak`. Plain `str::to_lowercase` follows no locale and turns `İ` into `i` plus a //! combining dot, which nothing typed would match. //! -//! The Turkish i handling and sigma folding come from PdfCraft's `fold_case` (snipeTR, +//! The Turkish i handling and sigma folding come from `PdfCraft`'s `fold_case` (snipeTR, //! nevicia2010-hub). use std::ops::Range; @@ -34,6 +34,7 @@ pub struct Folder { impl Folder { /// Case- and accent-insensitive folding for `locale`, a language code such as `tr`, `pt-br` /// or `zh-hant` (any case, `-` or `_`). + #[must_use] pub fn new(locale: &str) -> Folder { let language = locale.split(['-', '_', '.', '@']).next().unwrap_or_default(); Folder { turkic: language.eq_ignore_ascii_case("tr") || language.eq_ignore_ascii_case("az"), accents: false } @@ -41,11 +42,13 @@ impl Folder { /// Accents count: `é` no longer matches `e`, and a match can't end on a letter whose accent it /// leaves out. + #[must_use] pub fn keep_accents(self) -> Folder { Folder { accents: true, ..self } } /// `s` folded, with the map back to `s`. + #[must_use] pub fn fold(&self, s: &str) -> Folded { let nfkd = DecomposingNormalizerBorrowed::new_nfkd(); let extend = CodePointSetData::new::(); @@ -75,17 +78,19 @@ impl Folder { } /// `s` folded, without the map: a key to compare or store. + #[must_use] pub fn key(&self, s: &str) -> String { self.fold(s).text } /// Whether `haystack` contains `needle`, both folded. + #[must_use] pub fn contains(&self, haystack: &str, needle: &str) -> bool { self.fold(haystack).find(&self.fold(needle)).is_some() } /// One cluster, already in NFKD. - fn fold_cluster(&self, cluster: &str, out: &mut String) { + fn fold_cluster(self, cluster: &str, out: &mut String) { let mut chars = cluster.chars().peekable(); while let Some(c) = chars.next() { match c { @@ -219,6 +224,7 @@ pub struct Folded { impl Folded { /// The folded text. + #[must_use] pub fn as_str(&self) -> &str { &self.text } @@ -226,6 +232,7 @@ impl Folded { /// The source byte range that folded bytes `range` came from, widened to whole source /// characters with their accents: `ss` in the folding of `Straße` maps back to `ß`, `e` in /// that of `é` to the whole `é`. An empty range maps to an empty one. + #[must_use] pub fn source_range(&self, range: Range) -> Range { // The last cluster starting at or before `at` is the one holding it: clusters that fold to // nothing share their offset with the next one. @@ -241,6 +248,7 @@ impl Folded { /// The first match of `needle` (folded the same way), as a byte range of the source text. /// An empty needle matches at the start. + #[must_use] pub fn find(&self, needle: &Folded) -> Option> { self.matches(needle).next() } @@ -363,10 +371,10 @@ mod tests { #[test] fn matches_map_back_to_the_source() { - let f = Folder::new("en"); fn find<'a>(f: Folder, haystack: &'a str, needle: &str) -> Option<&'a str> { f.fold(haystack).find(&f.fold(needle)).map(|r| &haystack[r]) } + let f = Folder::new("en"); let find = |haystack, needle| find(f, haystack, needle); assert_eq!(find("Große Straße", "strasse"), Some("Straße")); assert_eq!(find("Große Straße", "SS"), Some("ß")); diff --git a/crates/i18n/src/format.rs b/crates/i18n/src/format.rs index 14eb8c102..badfbdfca 100644 --- a/crates/i18n/src/format.rs +++ b/crates/i18n/src/format.rs @@ -4,7 +4,7 @@ use std::fmt; use fixed_decimal::{Decimal, Sign}; -use icu_decimal::{DecimalFormatter, parts}; +use icu_decimal::{DecimalFormatter, options::DecimalFormatterOptions, parts}; use icu_locale_core::Locale; use icu_locale_core::extensions::unicode::{key, value}; use writeable::{Part, PartsWrite, Writeable}; @@ -46,7 +46,8 @@ impl Format { if digits == Digits::Latin { locale.extensions.unicode.keywords.set(key!("nu"), value!("latn")); } - let decimal = data::provider().and_then(|p| DecimalFormatter::try_new_with_buffer_provider(p, (&locale).into(), Default::default()).ok()); + let decimal = data::provider() + .and_then(|p| DecimalFormatter::try_new_with_buffer_provider(p, (&locale).into(), DecimalFormatterOptions::default()).ok()); let mut format = Format { decimal, group: None, point: None, percent: ("", "%"), marks: Marks::default() }; // The locale's own separators and digits, read from a number that shows them all. let probe = format.parts(&Decimal::try_from_str("-1234567.5").unwrap_or_default()); @@ -81,6 +82,7 @@ impl Format { } /// The format for a locale tag (`de-CH`, `ar-EG`), with the locale's own digits. + #[must_use] pub fn from_tag(tag: &str) -> Format { Format::new(&Region::from_tag(tag), Digits::Locale) } @@ -92,6 +94,7 @@ impl Format { /// `value` rounded to `decimals` digits after the decimal separator; infinities and NaN are /// `∞`, `-∞` and `NaN`. + #[must_use] pub fn number(&self, value: f64, decimals: usize) -> String { match to_decimal(value, decimals) { Some(d) => self.write(&d), @@ -101,6 +104,7 @@ impl Format { /// [`Self::number`] without grouping separators, for values such as pixel sizes and years: /// `1920`, `١٩٢٠`, `1234,5`. + #[must_use] pub fn number_ungrouped(&self, value: f64, decimals: usize) -> String { match to_decimal(value, decimals) { Some(d) => { @@ -113,6 +117,7 @@ impl Format { /// `ratio` as a percentage with `decimals` digits after the decimal separator: `0.5` is /// `50%`, `50 %` (French, German), `%50` (Turkish), `٥٠٪؜` (Egyptian Arabic). + #[must_use] pub fn percent(&self, ratio: f64, decimals: usize) -> String { let Some(d) = to_decimal(ratio * 100.0, decimals) else { return format!("{}{}{}", self.percent.0, non_finite(ratio), self.percent.1); @@ -212,7 +217,7 @@ impl PartsWrite for Sink<'_> { _ => return result, }; if slot.is_empty() { - *slot = written.to_owned(); + written.clone_into(slot); } } result @@ -370,7 +375,7 @@ mod tests { for (tag, want) in [("en", "50%"), ("de", "50\u{a0}%"), ("fr", "12,5\u{a0}%"), ("tr", "%50"), ("ar-EG", "٥٠٪\u{61c}"), ("fa", "۵۰٪"), ("hi", "50%")] { - let decimals = if tag == "fr" { 1 } else { 0 }; + let decimals = usize::from(tag == "fr"); let ratio = if tag == "fr" { 0.125 } else { 0.5 }; assert_eq!(Format::from_tag(tag).percent(ratio, decimals), want, "{tag}"); } diff --git a/crates/i18n/src/lib.rs b/crates/i18n/src/lib.rs index 53bbffe9e..ce473d84f 100644 --- a/crates/i18n/src/lib.rs +++ b/crates/i18n/src/lib.rs @@ -113,7 +113,9 @@ impl Locale { rows.entry(e.context).or_default().insert(e.source, e.translation); } } - problems.iter().for_each(|p| log::warn!("translation catalog: {p}; row skipped")); + for p in &problems { + log::warn!("translation catalog: {p}; row skipped"); + } Parsed { rows, problems } }) } @@ -156,6 +158,7 @@ impl I18n { /// to left, say). Locales that can't be used are skipped and come back as problems for the app /// to log. A locale's bad rows are logged (`log::warn!`) when it is first looked up in, and /// [`I18n::check_all`] lists them all. + #[must_use] pub fn new(embedded: &'static Embedded, offered: &[&str]) -> (I18n, Vec) { let mut problems = Vec::new(); let mut locales = vec![bare("en"), bare("pseudo")]; @@ -417,6 +420,7 @@ impl I18n { /// value that already is one isolate keeps it; pass numbers, paths and codes through [`ltr`]. /// Other text gets no marks, and a value's own isolate is taken off. Unknown placeholders stay as written, /// and inserted values are never read as templates, so a file name containing `{n}` stays intact. +#[must_use] pub fn fmt(template: &str, args: &[(&str, &str)]) -> String { fill(template, args, true) } @@ -424,6 +428,7 @@ pub fn fmt(template: &str, args: &[(&str, &str)]) -> String { /// [`fmt`] without isolates, for text that goes where the marks would get in the way: file /// names, logs, terminals, text compared or parsed later. Every isolate mark is taken off, in /// the template and in the values. +#[must_use] pub fn fmt_plain(template: &str, args: &[(&str, &str)]) -> String { strip_isolates(&fill(template, args, false)) } @@ -444,6 +449,7 @@ const ALM: char = '\u{061C}'; /// `value` as a first-strong isolate (FSI…PDI): it takes the direction of its first strong /// letter and can't reorder the text around it. What [`fmt`] does to each value in text with a /// right-to-left letter. +#[must_use] pub fn isolate(value: &str) -> String { wrap(FSI, value) } @@ -452,12 +458,14 @@ pub fn isolate(value: &str) -> String { /// codes, which read left to right in every language. In an RTL first-strong isolate, the /// segments of `/home/مستخدم/file.txt` would run right to left. [`fmt`] takes the isolate off /// again in text with no right-to-left letter. +#[must_use] pub fn ltr(value: &str) -> String { wrap(LRI, value) } /// `text` without isolate marks (LRI, RLI, FSI, PDI): a [`fmt`] or [`I18n::trn`] result made fit /// for a file name, a log or a comparison. +#[must_use] pub fn strip_isolates(text: &str) -> String { text.chars().filter(|c| !matches!(*c, LRI | RLI | FSI | PDI)).collect() } @@ -483,14 +491,13 @@ fn wrap(initiator: char, value: &str) -> String { PDF if open.last() == Some(&false) => { open.pop(); } - PDF => continue, PDI if isolates > 0 => { while open.pop() == Some(false) { out.push(PDF); } isolates -= 1; } - PDI => continue, + PDF | PDI => continue, _ => {} } out.push(c); @@ -713,11 +720,11 @@ fn plural_rules(code: &str) -> Option<(PluralRules, Vec)> { categories.push(c); } } - categories.sort_by_key(rank); + categories.sort_by_key(|c| rank(*c)); Some((rules, categories)) } -fn rank(c: &PluralCategory) -> u8 { +fn rank(c: PluralCategory) -> u8 { match c { PluralCategory::Zero => 0, PluralCategory::One => 1, diff --git a/crates/i18n/src/number.rs b/crates/i18n/src/number.rs index 1120fbe88..211d9dfe0 100644 --- a/crates/i18n/src/number.rs +++ b/crates/i18n/src/number.rs @@ -80,6 +80,12 @@ impl Format { /// /// Percent signs, units and the space before a unit (`50 %`, `12 px`) are not read; strip them /// first. Anything else is an error rather than a guess. + /// + /// # Errors + /// + /// Returns `NumberError::Empty` for an empty input or one reduced to spaces and a sign, + /// `Unexpected` for an invalid character, `MixedDigits` for mixed digit sets, `Separator` + /// for misplaced marks, and `TooLarge` when the parsed value is not finite. pub fn parse(&self, text: &str) -> Result { let marks = self.marks; let text: String = text.chars().filter(|c| !is_bidi_mark(*c)).collect(); @@ -129,6 +135,11 @@ impl Format { /// The number in `text` as `locale` (a language tag, `de-CH`, `ar_EG.UTF-8`) writes it: /// [`Format::parse`] for [`Format::from_tag`]. Build the [`Format`] once to read many numbers. A /// locale CLDR doesn't know reads as CLDR's root does (`1,234.5`). +/// +/// # Errors +/// +/// Returns the same `NumberError` as [`Format::parse`] for invalid digits, characters, or +/// separators, or when the value is not finite. pub fn parse_number(text: &str, locale: &str) -> Result { Format::from_tag(locale).parse(text) } diff --git a/crates/i18n/src/region.rs b/crates/i18n/src/region.rs index fcdffe8af..205444c25 100644 --- a/crates/i18n/src/region.rs +++ b/crates/i18n/src/region.rs @@ -73,6 +73,7 @@ impl Region { /// `nu` stays in the number and time locales (`ar-EG-u-nu-latn`). `C` and `POSIX` are the C /// locale: English, A4, metric, a `.` decimal and no grouping. An empty or unreadable tag is /// `en-US`; a bare `en` is the United States, any other bare language has no territory. + #[must_use] pub fn from_tag(tag: &str) -> Region { match Tag::parse(tag) { Some(tag) => Region::from_parts(&tag), @@ -85,6 +86,7 @@ impl Region { /// The region for an app's saved region setting: `""` or `auto` (any case) is /// [`Region::system`], anything else a locale tag for [`Region::from_tag`]. + #[must_use] pub fn from_setting(setting: &str) -> Region { let setting = setting.trim(); if setting.is_empty() || setting.eq_ignore_ascii_case("auto") { Region::system() } else { Region::from_tag(setting) } @@ -99,6 +101,7 @@ impl Region { /// - Linux and the BSDs: `LC_ALL`, else each of `LC_NUMERIC`, `LC_TIME`, `LC_MEASUREMENT` and /// `LC_PAPER` on its own, else `LANG`. /// - The web and Android: the preferred language, the only locale they expose. + #[must_use] pub fn system() -> Region { static SYSTEM: OnceLock = OnceLock::new(); SYSTEM.get_or_init(read_system).clone() @@ -298,6 +301,7 @@ fn from_apple(identifier: &str, metric: bool, decimal: &str, grouping: &str, hou /// A grouping separator the system sets: none at all when empty (Windows' `sThousand` the user /// cleared), else one character. #[cfg(any(test, windows, target_vendor = "apple"))] +#[expect(clippy::option_option, reason = "The public setting distinguishes following the locale, no grouping, and a custom separator")] fn grouping(s: &str) -> Option> { if s.is_empty() { Some(None) } else { single_char(s).map(Some) } } diff --git a/crates/i18n/src/scripts.rs b/crates/i18n/src/scripts.rs index 1ae0f29bc..e16342e26 100644 --- a/crates/i18n/src/scripts.rs +++ b/crates/i18n/src/scripts.rs @@ -39,6 +39,7 @@ pub enum Script { impl Script { /// The ISO 15924 code, `Latn`. + #[must_use] pub fn code(self) -> &'static str { match self { Script::Arab => "Arab", @@ -75,6 +76,7 @@ impl Script { } /// Whether text in this script runs right to left. + #[must_use] pub fn rtl(self) -> bool { matches!(self, Script::Arab | Script::Hebr | Script::Syrc | Script::Thaa) } @@ -112,8 +114,12 @@ pub enum Han { /// where it is the language's own script. A language not in the table is taken to be written in /// Latin. // ponytail: the languages a Craft app ships or could plausibly ship; add a row with the locale. +#[must_use] pub fn scripts(tag: &str) -> &'static [Script] { - use Script::*; + use Script::{ + Arab, Armn, Beng, Cyrl, Deva, Ethi, Geor, Grek, Gujr, Guru, Hang, Hani, Hebr, Hira, Kana, Khmr, Knda, Laoo, Latn, Mlym, Mymr, Orya, Sinh, + Syrc, Taml, Telu, Thaa, Thai, Tibt, + }; let tag = Subtags::parse(tag); if let Some(s) = tag.script.as_deref().and_then(subtag_scripts) { return s; @@ -154,6 +160,7 @@ pub fn scripts(tag: &str) -> &'static [Script] { } /// The direction of a language tag's text: right to left when its main script is. +#[must_use] pub fn direction(tag: &str) -> Direction { match scripts(tag).first() { Some(s) if s.rtl() => Direction::Rtl, @@ -165,6 +172,7 @@ pub fn direction(tag: &str) -> Direction { /// P3): the contents of isolates (LRI, RLI or FSI up to the matching PDI) are skipped, and the /// search stops at the first paragraph separator. `None` when the paragraph has no strong /// character; P3 then reads it as left to right, and an app may use the interface direction. +#[must_use] pub fn base_direction(text: &str) -> Option { match unicode_bidi::get_base_direction(text) { unicode_bidi::Direction::Ltr => Some(Direction::Ltr), @@ -175,9 +183,10 @@ pub fn base_direction(text: &str) -> Option { /// The forms a language tag's Han characters take: `zh-CN`, `zh-SG`, `zh-Hans` and plain `zh` → /// Simplified; `zh-TW`, `zh-Hant` and the other regions CLDR writes in Traditional characters -/// (`zh-US`, `zh-AU`) → Traditional; `zh-HK`, `zh-MO`, `zh-Hant-HK` and `yue` → HongKong; `ja` → +/// (`zh-US`, `zh-AU`) → Traditional; `zh-HK`, `zh-MO`, `zh-Hant-HK` and `yue` → `HongKong`; `ja` → /// Japanese; `ko` → Korean. `None` for a language not written with Han, which leaves the choice /// to the font stack. +#[must_use] pub fn han(tag: &str) -> Option { let tag = Subtags::parse(tag); let hong_kong = matches!(tag.region.as_deref(), Some("hk" | "mo")); @@ -202,8 +211,14 @@ pub fn han(tag: &str) -> Option { // ponytail: mn-CN → Mong is left out until vertical Mongolian layout is supported. pub(crate) fn region_script(language: &str, region: &str) -> Option<&'static str> { Some(match (language, region) { - ("az", "iq" | "ir") | ("ha", "cm" | "sd") | ("kk", "af" | "cn" | "ir" | "mn") | ("ku", "iq" | "ir" | "lb") => "arab", - ("ky", "cn") | ("ms", "cc") | ("pa", "pk") | ("tg", "pk") | ("uz", "af") => "arab", + ("az", "iq" | "ir") + | ("ha", "cm" | "sd") + | ("kk", "af" | "cn" | "ir" | "mn") + | ("ku", "iq" | "ir" | "lb") + | ("ky", "cn") + | ("ms", "cc") + | ("pa" | "tg", "pk") + | ("uz", "af") => "arab", ("az", "ru") | ("ku", "am" | "az" | "ge" | "tm") | ("ug", "kz" | "mn") | ("uz", "cn") => "cyrl", ("ky", "tr") | ("sr", "me" | "ro" | "tr") => "latn", ("sd", "in") => "deva", @@ -216,7 +231,10 @@ pub(crate) fn region_script(language: &str, region: &str) -> Option<&'static str /// The scripts a script subtag stands for, including the CLDR unions (`jpan`, `kore`). `None` for /// one not in [`Script`], so the language's own scripts apply. fn subtag_scripts(subtag: &str) -> Option<&'static [Script]> { - use Script::*; + use Script::{ + Arab, Armn, Beng, Cyrl, Deva, Ethi, Geor, Grek, Gujr, Guru, Hang, Hani, Hebr, Hira, Kana, Khmr, Knda, Laoo, Latn, Mlym, Mong, Mymr, Orya, + Sinh, Syrc, Taml, Telu, Thaa, Thai, Tibt, + }; Some(match subtag { "hans" | "hant" | "hani" => &[Hani], "jpan" => &[Hani, Hira, Kana], diff --git a/crates/i18n/src/size.rs b/crates/i18n/src/size.rs index 830c09573..b316360a5 100644 --- a/crates/i18n/src/size.rs +++ b/crates/i18n/src/size.rs @@ -33,6 +33,7 @@ impl SizeUnit { /// // `12 MB`, `3.4 MB`, `640 KB`: decimal steps, whole kilobytes, no decimals from 10 MB. /// let style = FileSize::DECIMAL.smallest(SizeUnit::Kilo).largest(SizeUnit::Mega).unit_decimals(SizeUnit::Kilo, 0).whole_from(10); /// ``` +#[must_use] #[derive(Clone, Copy, Debug, PartialEq, Eq, Hash)] pub struct FileSize { step: u64, @@ -104,6 +105,10 @@ impl FileSize { } /// The unit `bytes` is written in, the number in it and its digits after the point. + #[expect( + clippy::cast_precision_loss, + reason = "File sizes are displayed as rounded f64 values; exact u64 arithmetic is not required for presentation" + )] fn parts(&self, bytes: u64) -> (SizeUnit, f64, usize) { let (smallest, largest) = (self.smallest.min(self.largest), self.largest); let mut unit = smallest; @@ -137,6 +142,7 @@ impl FileSize { } } + #[expect(clippy::cast_precision_loss, reason = "Unit scales are displayed as f64 values; exact u64 arithmetic is not required for presentation")] fn scale(&self, unit: SizeUnit) -> f64 { (self.step as f64).powi(unit as i32) } @@ -160,6 +166,7 @@ impl FileSize { impl I18n { /// `bytes` as a file size in `lang`, written the way `style` says (`1.5 MB`, `512 bytes`), the /// number as [`I18n::format`] writes it (`1,5 Mo`). + #[must_use] pub fn file_size(&self, lang: Lang, bytes: u64, style: FileSize) -> String { let (unit, n, decimals) = style.parts(bytes); if unit == SizeUnit::Byte && !style.byte_symbol { @@ -171,6 +178,7 @@ impl I18n { /// `value` with `decimals` digits after the decimal separator in the unit template `unit`, an /// English catalog key with the number as `{n}` (`{n} px`, `{n} dB`), the number as /// [`I18n::format`] writes it. Percentages are [`crate::Format::percent`]. + #[must_use] pub fn measure(&self, lang: Lang, value: f64, decimals: usize, unit: &str) -> String { self.unit_text(lang, unit, &self.format().number(value, decimals)) }