Merge pull request #2708 from dotCooCoo/fix/i18n-scan-reads-bad-messages

i18n: the message scan reads errors made with bad() and error variants, and new ones need rows
2026-10-27 11:17:01 -06:00 · 2026-10-22 00:51:36 +07:00
parent d58888a95c a3e7dc9e41
commit 675b759a37
4 changed files with 1459 additions and 26 deletions
@@ -1649,7 +1649,10 @@ fn messages_translate_by_template_and_their_parts() {

 /// Each language that translates status and error messages ([`COMPLETE_MESSAGES`]) has an `@msg`
 /// row for every message in the sources that reach the status bar, so a new one can't ship
-/// untranslated by accident (`VECTORCRAFT_I18N_DUMP_MESSAGES=<file>` lists them).
+/// untranslated by accident (`VECTORCRAFT_I18N_DUMP_MESSAGES=<file>` lists them). A message listed
+/// in `untranslated_messages.txt` passes whether or not it has rows
+/// (`VECTORCRAFT_I18N_DUMP_UNTRANSLATED=<file>` writes that file's header and the messages in the
+/// sources that lack a row).
 #[test]
 fn complete_languages_translate_every_message() {
    let messages = message_literals();
@@ -1657,17 +1660,86 @@ fn complete_languages_translate_every_message() {
    if let Ok(path) = std::env::var("VECTORCRAFT_I18N_DUMP_MESSAGES") {
        std::fs::write(path, messages.iter().map(|m| format!("{m}\n")).collect::<String>()).expect("write dump");
    }
+    let listed: std::collections::HashSet<&str> = UNTRANSLATED_MESSAGES.lines().filter(|l| !l.is_empty() && !l.starts_with('#')).collect();
+    let mut untranslated = std::collections::BTreeSet::new();
+    let mut unlisted = vec![];
    for code in COMPLETE_MESSAGES {
        let lang = Lang::from_code(code).expect("registered");
        let (entries, _) = parse_entries(lang.0.source);
        let rows: std::collections::HashSet<&str> = entries.iter().filter(|(ctx, _, _)| ctx == "@msg").map(|(_, src, _)| src.as_str()).collect();
-        let missing: Vec<_> = messages.iter().filter(|m| !rows.contains(m.as_str())).collect();
+        let missing: Vec<&str> = messages.iter().map(String::as_str).filter(|m| !rows.contains(m)).collect();
+        untranslated.extend(missing.iter().copied());
+        unlisted.push((code, missing.into_iter().filter(|m| !listed.contains(m)).collect::<Vec<_>>()));
+    }
+    if let Ok(path) = std::env::var("VECTORCRAFT_I18N_DUMP_UNTRANSLATED") {
+        let header = UNTRANSLATED_MESSAGES.lines().take_while(|l| l.starts_with('#'));
+        std::fs::write(path, header.chain(untranslated).map(|l| format!("{l}\n")).collect::<String>()).expect("write dump");
+    }
+    for (code, missing) in unlisted {
        assert!(missing.is_empty(), "{code}: {} untranslated messages: {missing:#?}", missing.len());
    }
 }

-/// The parameter errors of Align, Distribute Spacing, Reflect and Shear, whose text the scan above
-/// doesn't read, are translated whole in each language of [`COMPLETE_MESSAGES`].
+/// The messages [`complete_languages_translate_every_message`] lets through without rows (see the
+/// file's header).
+const UNTRANSLATED_MESSAGES: &str = include_str!("untranslated_messages.txt");
+
+/// The scan reads the message of an error made with one of the engine's [`ERROR_HELPERS`] or with
+/// an error enum's variant: a literal or a `format!` string, on the call's line or wrapped over
+/// several lines, in the same template form as the catalogs' rows. It does not read a call whose
+/// name ends in a helper's name (`is_bad(`), a method call (`.err(`), `std::io::Error::new` or one
+/// of [`NAME_VARIANTS`]; without the helpers, it reads only the two variants in this source.
+#[test]
+fn message_scan_reads_error_helpers_and_variants() {
+    let code = r#"
+fn width(p: &Value, cmd: &str, key: &str) -> Result<f64> {
+    let w = p.get(key).and_then(Value::as_f64).ok_or_else(|| bad(cmd, format!("`{key}` must be a number")))?;
+    if w < 0.0 {
+        return Err(bad(cmd, "the width can't be negative"));
+    }
+    let e = bad(
+        "object.example",
+        format!(
+            "`{}` must be between {lo:.1} and {hi}, not {w}",
+            key
+        ),
+    );
+    let _ = bad(&format!("{cmd}, {key}"), "a comma in the first argument");
+    let _ = is_bad(cmd, "not an error helper");
+    let _ = r.err("not an error helper either");
+    let _ = err(format!("can't open `{name}`: damaged"));
+    let _ = PsError::Limit(format!("the {what} is nested too deep"));
+    let _ = Error::Manifest("the manifest has no name".into());
+    let _ = std::io::Error::new(kind, "not an error variant");
+    let _ = PsError::Ps("not an error message", op);
+    let _ = EngineError::UnknownCommand("not.an.error.message".into());
+    Err(error("Select only dimensions"))
+}
+"#;
+    let engine: Vec<_> = ERROR_HELPERS.iter().filter(|(c, _, _)| *c == "engine").map(|&(_, call, before)| (call, before)).collect();
+    let found = source_messages(code, &engine);
+    for want in [
+        "`{key}` must be a number",
+        "the width can't be negative",
+        "`{_1}` must be between {lo} and {hi}, not {w}",
+        "a comma in the first argument",
+        "can't open `{name}`: damaged",
+        "the {what} is nested too deep",
+        "the manifest has no name",
+        "Select only dimensions",
+    ] {
+        assert!(found.iter().any(|m| m == want), "{want} not in {found:#?}");
+    }
+    assert!(!found.iter().any(|m| m.contains("not an error") || m.contains("not.an.error")), "{found:#?}");
+    let elsewhere = source_messages(code, &[]);
+    assert!(elsewhere.len() == 2 && elsewhere.iter().all(|m| m.contains("nested") || m.contains("manifest")), "{elsewhere:#?}");
+    // A message read this way matches its row: `{key}` keeps its name.
+    let cs = Lang::from_code("cs").expect("registered");
+    assert_eq!(message(cs, "`width` must be a number"), "`width` musí být číslo");
+}
+
+/// The parameter errors of Align, Distribute Spacing, Reflect and Shear are translated whole in each
+/// language of [`COMPLETE_MESSAGES`].
 #[test]
 fn align_reflect_and_shear_param_errors_are_translated() {
    use serde_json::json;
@@ -1732,8 +1804,7 @@ fn the_data_set_applied_nothing_warning_is_translated() {

 /// The parameter errors of the blend spine commands (Move Spine Point, Reshape Spine Segment, Add
 /// Spine Point and Delete Spine Point, which Direct Selection, the Pen and the Anchor Point tool
-/// send), whose text the scan above doesn't read, are translated whole in each language of
-/// [`COMPLETE_MESSAGES`] and in Bulgarian.
+/// send) are translated whole in each language of [`COMPLETE_MESSAGES`] and in Bulgarian.
 #[test]
 fn blend_spine_param_errors_are_translated() {
    use serde_json::json;
@@ -1878,8 +1949,8 @@ fn unknown_parameter_errors_are_translated() {
    }
 }

-/// The error for an effect `color` the effect can't draw, whose text the scan above doesn't read,
-/// is translated whole in each language of [`COMPLETE_MESSAGES`] and in Bulgarian.
+/// The error for an effect `color` the effect can't draw is translated whole in each language of
+/// [`COMPLETE_MESSAGES`] and in Bulgarian.
 #[test]
 fn effect_color_param_error_is_translated() {
    use serde_json::json;
@@ -1967,9 +2038,8 @@ fn unread_layers_notes_are_translated() {
    }
 }

-/// The parameter error of `select.key` for an id that isn't a selected top-level object, whose
-/// text [`message_literals`] doesn't read, is translated whole in each language of
-/// [`COMPLETE_MESSAGES`].
+/// The parameter error of `select.key` for an id that isn't a selected top-level object is
+/// translated whole in each language of [`COMPLETE_MESSAGES`].
 #[test]
 fn select_key_param_error_is_translated() {
    use serde_json::json;
@@ -2042,6 +2112,17 @@ const MESSAGE_CRATES: &[&str] = &["ui-egui", "engine", "doc", "format", "svg", "
 /// Where a message literal starts: a status, an error value, a `thiserror` message.
 const MESSAGE_MARKERS: &[&str] = &[".status(", "ui.status = ", "status = ", "Other(", "Err(", "ok_or(", "ok_or_else(|| ", "#[error("];

+/// Helpers that make an error from a message, by crate, with the number of arguments before the
+/// message: the engine's `bad(cmd, msg)` (`EngineError::BadParams`), `err(msg)` (`cmd/fileio`)
+/// and `error(msg)` (`cmd/dimension.rs`), and the PDF writer's `bad(why)` (`forms.rs`) and
+/// `failed(why)` (`encrypt.rs`, `linearize.rs`).
+const ERROR_HELPERS: &[(&str, &str, usize)] =
+    &[("engine", "bad(", 1), ("engine", "err(", 0), ("engine", "error(", 0), ("pdf", "bad(", 0), ("pdf", "failed(", 0)];
+
+/// Error variants whose first argument is a name, not a message: a command id
+/// (`EngineError::UnknownCommand`) and a PostScript error name (`PsError::Ps`).
+const NAME_VARIANTS: &[&str] = &["UnknownCommand", "Ps"];
+
 /// Messages that are signals, not text for people.
 const NOT_MESSAGES: &[&str] = &["quit", "cancelled"];

@@ -2063,24 +2144,78 @@ pub fn message_literals() -> std::collections::BTreeSet<String> {
                // Control replies are English protocol data, never status-bar messages
                // (see i18n's module contract). Engine messages are scanned in their own crates.
                let text = std::fs::read_to_string(&path).unwrap_or_default();
-                let code = without_test_module(&text);
-                for marker in MESSAGE_MARKERS {
-                    for (at, _) in code.match_indices(marker) {
-                        let rest = code[at + marker.len()..].trim_start();
-                        let rest = rest.strip_prefix("format!(").map_or(rest, str::trim_start);
-                        let Some(lit) = rest.strip_prefix('"').and_then(string_literal) else { continue };
-                        let has_word = lit.as_bytes().windows(3).any(|w| w.iter().all(u8::is_ascii_lowercase));
-                        if has_word && !NOT_MESSAGES.contains(&lit.as_str()) {
-                            out.insert(as_template(&lit));
-                        }
-                    }
-                }
+                let helpers: Vec<_> =
+                    ERROR_HELPERS.iter().filter(|(c, _, _)| path.starts_with(crates.join(c))).map(|&(_, call, before)| (call, before)).collect();
+                out.extend(source_messages(without_test_module(&text), &helpers));
            }
        }
    }
    out
 }

+/// The message literals in `code`, as catalog templates (see [`message_literals`]): the literal
+/// after each of [`MESSAGE_MARKERS`], the first argument of an error enum's variant
+/// (`PsError::Limit(…)`, `Error::Manifest(…)`; not [`NAME_VARIANTS`]),
+/// and the message argument of each call in `helpers` (a call and the number of arguments before
+/// its message), a literal or a `format!` string.
+fn source_messages(code: &str, helpers: &[(&str, usize)]) -> Vec<String> {
+    let after_markers = MESSAGE_MARKERS.iter().flat_map(|marker| code.match_indices(marker).map(|(at, _)| &code[at + marker.len()..]));
+    let in_variants = code.match_indices("Error::").filter_map(|(at, m)| {
+        let rest = &code[at + m.len()..];
+        let (variant, args) = rest.split_at(rest.find(|c: char| !c.is_ascii_alphanumeric() && c != '_').unwrap_or(rest.len()));
+        let named = variant.starts_with(|c: char| c.is_ascii_uppercase()) && !NAME_VARIANTS.contains(&variant);
+        if named { args.strip_prefix('(') } else { None }
+    });
+    let in_helpers = helpers.iter().flat_map(|&(call, before)| {
+        code.match_indices(call)
+            .filter(|&(at, _)| !code[..at].ends_with(|c: char| c.is_alphanumeric() || c == '_' || c == '.'))
+            .filter_map(move |(at, _)| skip_args(&code[at + call.len()..], before))
+    });
+    let mut out = vec![];
+    for rest in after_markers.chain(in_variants).chain(in_helpers) {
+        let rest = rest.trim_start();
+        let rest = rest.strip_prefix("format!(").map_or(rest, str::trim_start);
+        let Some(lit) = rest.strip_prefix('"').and_then(string_literal) else { continue };
+        let has_word = lit.as_bytes().windows(3).any(|w| w.iter().all(u8::is_ascii_lowercase));
+        if has_word && !NOT_MESSAGES.contains(&lit.as_str()) {
+            out.push(as_template(&lit));
+        }
+    }
+    out
+}
+
+/// The arguments of a call whose opening parenthesis was just read, past the first `n` of them;
+/// `None` when `n` is 1 or more and the call has `n` arguments or fewer.
+fn skip_args(args: &str, n: usize) -> Option<&str> {
+    if n == 0 {
+        return Some(args);
+    }
+    let (mut left, mut depth, mut chars) = (n, 0usize, args.char_indices());
+    while let Some((at, c)) = chars.next() {
+        match c {
+            '(' | '[' | '{' => depth += 1,
+            ')' | ']' | '}' => depth = depth.checked_sub(1)?,
+            '"' => loop {
+                match chars.next()?.1 {
+                    '\\' => {
+                        chars.next()?;
+                    }
+                    '"' => break,
+                    _ => {}
+                }
+            },
+            ',' if depth == 0 => {
+                left -= 1;
+                if left == 0 {
+                    return Some(&args[at + 1..]);
+                }
+            }
+            _ => {}
+        }
+    }
+    None
+}
+
 /// Source up to its inline test module (`#[cfg(test)] mod tests {`); declarations of test files
 /// (`#[cfg(test)] mod tests_x;`) don't end it.
 fn without_test_module(text: &str) -> &str {
@@ -254,9 +254,14 @@ other `sv-*` locale resolve to it), Georgian (`ka`, complete; `ka-GE` and `ka_GE
  `Couldn't open {name}: {e}`; `{_1}`, `{_2}` … stand for the format string's `{}`, and the values in the
  placeholders are translated in turn (the reason after `: {e}` is often a message too). Czech, German, Greek, Spanish, French, Georgian, Hungarian, Italian, Japanese, Korean, Polish, Brazilian Portuguese, Russian, Simplified Chinese, Traditional Chinese, Swedish, Ukrainian, Vietnamese and Finnish cover every
  message literal the test scan finds (`complete_languages_translate_every_message`, languages listed in
-  `COMPLETE_MESSAGES`): a new `Err("…")`, `Other(…)`, `#[error(…)]` or `status(…)` message needs a row in each of those catalogs (`VECTORCRAFT_I18N_DUMP_MESSAGES=messages.txt cargo test -p vectorcraft-ui-egui
-  complete_languages_translate_every_message` lists them all). Other languages show messages in English
-  until they add `@msg` rows.
+  `COMPLETE_MESSAGES`), except the messages listed in `crates/ui-egui/src/i18n/untranslated_messages.txt`.
+  A new `Err("…")`, `Other(…)`, `#[error(…)]` or `status(…)` message needs a row in each of those catalogs, and so
+  does a new message given to an error enum's variant (`PsError::Limit(…)`), to the engine's `bad(cmd, …)`, `err(…)` or
+  `error(…)`, or to the PDF writer's `bad(…)` or `failed(…)` (`VECTORCRAFT_I18N_DUMP_MESSAGES=messages.txt cargo test -p vectorcraft-ui-egui
+  complete_languages_translate_every_message` lists them all). The list also holds the messages that open pull
+  requests add without those rows. A change that adds the rows for a listed message removes its line from the list
+  (`VECTORCRAFT_I18N_DUMP_UNTRANSLATED=<file>` writes the messages in the sources that lack a row). Other languages
+  show messages in English until they add `@msg` rows.
 - Not translated on purpose: names that are user data (layers, swatches, fonts, documents), the tab
  title's colour mode, command ids and parameter names inside messages. Not done yet: locale-aware number
  and date formats, right-to-left layout. Chinese and Japanese UI text
@@ -10,6 +10,11 @@ Per-language status of VectorCraft's interface. How the catalogs work and how to
 origin/main `01e165af`, except the rows that give a count date, which were counted on `main` at `cb0a7d9` (2026-10-21).
 The reference is the French catalog, which the tests hold complete: 3,416 interface
 strings (menus, command labels, `tl!` literals: panels, dialogs, tooltips, Preferences) plus 604 `@msg` status and
+error messages, 4,020 in all. A language's % is its rows over 4,020 (capped at 100% per part). The
+`complete_menus` flag and `COMPLETE_MESSAGES` list make `i18n::tests` fail if a complete language misses a
+string, so "full" here is enforced by tests. The messages listed in `crates/ui-egui/src/i18n/untranslated_messages.txt`
+are not translated yet, and the tests let them through. No language has help content: there is no translated (or English)
+in-app help, and the docs are English.
 error messages, 4,020 in all. A language's % is its rows over 4,020 (capped at 100% per part).
 `i18n::tests` fail if a language with the `complete_menus` flag misses a menu string or `tl!` literal, or if a
 language in the `COMPLETE_MESSAGES` list misses a status or error message. Every non-English language marked