dirsql 0.4.70__tar.gz → 0.4.71__tar.gz

{dirsql-0.4.70 → dirsql-0.4.71}/Cargo.lock

@@ -552,7 +552,7 @@ checksum = "6184e33543162437515c2e2b48714794e37845ec9851711914eec9d308f6ebe8"
 
 [[package]]
 name = "dirsql"
-version = "0.4.70"
+version = "0.4.71"
 dependencies = [
  "assert_cmd",
  "axum",
@@ -601,7 +601,7 @@ dependencies = [
 
 [[package]]
 name = "dirsql-py-ext"
-version = "0.4.70"
+version = "0.4.71"
 dependencies = [
  "dirsql",
  "pyo3",

{dirsql-0.4.70 → dirsql-0.4.71}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dirsql
-Version: 0.4.70
+Version: 0.4.71
 Requires-Dist: bin-shim>=0.1
 Summary: Ephemeral SQL index over a local directory
 Keywords: sql,filesystem,directory,sqlite,index

{dirsql-0.4.70 → dirsql-0.4.71}/packages/python/Cargo.toml

@@ -4,7 +4,7 @@ name = "dirsql-py-ext"
 # pypi/maturin handler can rewrite it via `write-version` before
 # `maturin build`. `pyproject.toml` declares `dynamic = ["version"]`
 # and maturin reads this field. Mirrors `packages/rust/Cargo.toml`.
-version = "0.4.70"
+version = "0.4.71"
 # abi3 note: the `abi3-py310` pyo3 feature (below) builds ONE stable-ABI
 # `cp310-abi3` wheel per platform that loads on every CPython >= 3.10
 # (matching `pyproject.toml`'s `requires-python = ">=3.10"`), instead of

dirsql-0.4.71/packages/python/e2e-attestations/claude-1267-posix-classes.json

@@ -0,0 +1,7 @@
+{
+  "command": "just test-e2e",
+  "ran_at": 1791224868,
+  "exit_code": 0,
+  "commit": "ca8ccbb55b7f51f9373b8f774665b925ba10ac21",
+  "branch": "claude/1267-posix-classes"
+}

{dirsql-0.4.70 → dirsql-0.4.71}/packages/rust/Cargo.toml

@@ -6,7 +6,7 @@ description = "Ephemeral SQL index over a local directory"
 # version stays for the binding crates (publish = false). The literal
 # intentionally lags the published version: putitoutthere rewrites it to
 # the planned release version at build time and never commits it back.
-version = "0.4.70"
+version = "0.4.71"
 edition.workspace = true
 # Literal `license` (not `license.workspace = true`) because
 # putitoutthere's preflight check reads `[package].license` directly

{dirsql-0.4.70 → dirsql-0.4.71}/packages/rust/src/lib.rs

@@ -33,6 +33,7 @@ pub mod parsed_vtab;
 pub mod path_table;
 #[doc(hidden)]
 pub mod persist;
+mod posix_class;
 #[doc(hidden)]
 pub mod progress;
 #[doc(hidden)]

{dirsql-0.4.70 → dirsql-0.4.71}/packages/rust/src/matcher.rs

@@ -1,3 +1,4 @@
+use crate::posix_class::expand_posix_classes;
 use globset::GlobBuilder;
 use regex::{Regex, RegexBuilder};
 use std::path::Path;
@@ -20,7 +21,9 @@ pub(crate) struct Pattern(Regex);
 impl Pattern {
     /// `*` and `?` stop at `/`; `**` crosses it.
     pub(crate) fn new(glob: &str) -> Result<Self, GlobError> {
-        let glob = GlobBuilder::new(glob).literal_separator(true).build()?;
+        let glob = GlobBuilder::new(&expand_posix_classes(glob))
+            .literal_separator(true)
+            .build()?;
         let regex = RegexBuilder::new(&unicode_regex(glob.regex()))
             .dot_matches_new_line(true)
             .build()?;
@@ -586,4 +589,11 @@ mod tests {
                 .is_empty()
         );
     }
+
+    #[test]
+    fn a_posix_class_in_a_table_glob_matches_its_characters() {
+        let matcher = TableMatcher::new(&[("notes/[[:digit:]].md", "n")], &[]).unwrap();
+        assert_eq!(names(&matcher, "notes/7.md"), vec!["n"]);
+        assert!(matcher.match_all(Path::new("notes/x.md")).is_empty());
+    }
 }

dirsql-0.4.71/packages/rust/src/posix_class.rs

@@ -0,0 +1,313 @@
+//! POSIX character classes (`[[:digit:]]`) for globset, which has no
+//! `[:name:]` syntax: each bracket expression holding one is rewritten into an
+//! equivalent bracket of plain ranges before the glob is compiled.
+
+/// Rewrite every bracket expression in `pattern` that holds a `[:name:]`
+/// class into plain ranges. Brackets without one, unclosed brackets and
+/// backslash-escaped characters are left exactly as written.
+pub(crate) fn expand_posix_classes(pattern: &str) -> String {
+    let chars: Vec<char> = pattern.chars().collect();
+    let mut out = String::with_capacity(pattern.len());
+    let mut i = 0;
+    while i < chars.len() {
+        match chars[i] {
+            '\\' => {
+                let end = (i + 2).min(chars.len());
+                out.extend(&chars[i..end]);
+                i = end;
+            }
+            '[' => match bracket(&chars, i) {
+                Some((end, Some(rewritten))) => {
+                    out.push_str(&rewritten);
+                    i = end;
+                }
+                Some((end, None)) => {
+                    out.extend(&chars[i..end]);
+                    i = end;
+                }
+                None => {
+                    out.push('[');
+                    i += 1;
+                }
+            },
+            c => {
+                out.push(c);
+                i += 1;
+            }
+        }
+    }
+    out
+}
+
+/// Parse the bracket expression opening at `start`. `None` when it never
+/// closes; otherwise the index past its `]`, plus its rewrite when it holds a
+/// class.
+fn bracket(chars: &[char], start: usize) -> Option<(usize, Option<String>)> {
+    let mut j = start + 1;
+    let negated = matches!(chars.get(j), Some('!' | '^'));
+    if negated {
+        j += 1;
+    }
+    let body_start = j;
+    let mut ranges: Vec<(char, char)> = Vec::new();
+    let mut has_class = false;
+    // bash matches nothing at all when a range ends in a class, negated or not.
+    let mut range_into_class = false;
+    loop {
+        let c = *chars.get(j)?;
+        if c == ']' && j > body_start {
+            j += 1;
+            break;
+        }
+        if let Some((name, next)) = class_at(chars, j) {
+            ranges.extend_from_slice(class_ranges(&name));
+            has_class = true;
+            j = next;
+            continue;
+        }
+        match (chars.get(j + 1), chars.get(j + 2)) {
+            (Some('-'), Some(&end)) if end != ']' => {
+                if let Some((_, next)) = class_at(chars, j + 2) {
+                    has_class = true;
+                    range_into_class = true;
+                    j = next;
+                    continue;
+                }
+                if c <= end {
+                    ranges.push((c, end));
+                }
+                j += 3;
+            }
+            _ => {
+                ranges.push((c, c));
+                j += 1;
+            }
+        }
+    }
+    if !has_class {
+        return Some((j, None));
+    }
+    let rewritten = if range_into_class {
+        "[\0]".to_string()
+    } else {
+        render(negated, ranges)
+    };
+    Some((j, Some(rewritten)))
+}
+
+/// The class `[:name:]` starting at `j`, with the index past its `:]`. A `]`
+/// before the `:]` means the `[` is an ordinary character, as in bash.
+fn class_at(chars: &[char], j: usize) -> Option<(String, usize)> {
+    if chars.get(j) != Some(&'[') || chars.get(j + 1) != Some(&':') {
+        return None;
+    }
+    let mut k = j + 2;
+    loop {
+        match (chars.get(k)?, chars.get(k + 1)) {
+            (':', Some(']')) => return Some((chars[j + 2..k].iter().collect(), k + 2)),
+            (']', _) => return None,
+            _ => k += 1,
+        }
+    }
+}
+
+/// The ASCII members of a class; an unknown name has none, as in bash.
+fn class_ranges(name: &str) -> &'static [(char, char)] {
+    match name {
+        "alnum" => &[('0', '9'), ('A', 'Z'), ('a', 'z')],
+        "alpha" => &[('A', 'Z'), ('a', 'z')],
+        "ascii" => &[('\0', '\x7f')],
+        "blank" => &[(' ', ' '), ('\t', '\t')],
+        "cntrl" => &[('\0', '\x1f'), ('\x7f', '\x7f')],
+        "digit" => &[('0', '9')],
+        "graph" => &[('!', '~')],
+        "lower" => &[('a', 'z')],
+        "print" => &[(' ', '~')],
+        "punct" => &[('!', '/'), (':', '@'), ('[', '`'), ('{', '~')],
+        "space" => &[('\t', '\r'), (' ', ' ')],
+        "upper" => &[('A', 'Z')],
+        "word" => &[('0', '9'), ('A', 'Z'), ('a', 'z'), ('_', '_')],
+        "xdigit" => &[('0', '9'), ('A', 'F'), ('a', 'f')],
+        _ => &[],
+    }
+}
+
+/// Characters globset reads as syntax at some position inside a bracket, so
+/// they are pulled out of ranges and placed where they are literal.
+const POSITIONAL: [char; 4] = [']', '-', '!', '^'];
+
+/// A globset bracket matching exactly `ranges` (or everything else when
+/// `negated`). A bracket never matches the separator, so `/` is dropped from a
+/// positive set and added to a negated one. NUL, which no path holds, stands
+/// in when nothing else may open the bracket.
+fn render(negated: bool, ranges: Vec<(char, char)>) -> String {
+    let mut set: Vec<(u32, u32)> = ranges
+        .into_iter()
+        .map(|(a, b)| (a as u32, b as u32))
+        .collect();
+    set = without(set, '/');
+    if negated {
+        set.push(('/' as u32, '/' as u32));
+    }
+    let present = |c: char| set.iter().any(|&(a, b)| a <= c as u32 && c as u32 <= b);
+    let specials: Vec<char> = POSITIONAL.into_iter().filter(|&c| present(c)).collect();
+    let mut plain = POSITIONAL.into_iter().fold(set.clone(), without);
+    plain.sort_unstable();
+    if plain.is_empty() && !specials.contains(&']') {
+        plain.push((0, 0));
+    }
+
+    let mut out = String::from("[");
+    if negated {
+        out.push('!');
+    }
+    if specials.contains(&']') {
+        out.push(']');
+    }
+    for (a, b) in plain {
+        out.extend(char::from_u32(a));
+        if a != b {
+            out.push('-');
+            out.extend(char::from_u32(b));
+        }
+    }
+    for c in ['!', '^', '-'] {
+        if specials.contains(&c) {
+            out.push(c);
+        }
+    }
+    out.push(']');
+    out
+}
+
+/// `set` with `c` removed, splitting any range that spans it. `c` is never
+/// NUL, so `c - 1` cannot underflow.
+fn without(set: Vec<(u32, u32)>, c: char) -> Vec<(u32, u32)> {
+    let c = c as u32;
+    set.into_iter()
+        .flat_map(|(a, b)| [(a, b.min(c - 1)), (a.max(c + 1), b)])
+        .filter(|&(a, b)| a <= b)
+        .collect()
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn a_class_becomes_its_ranges() {
+        assert_eq!(expand_posix_classes("[[:digit:]].md"), "[0-9].md");
+        assert_eq!(expand_posix_classes("[[:upper:]]*"), "[A-Z]*");
+    }
+
+    #[test]
+    fn every_named_class_expands() {
+        let cases = [
+            ("alnum", "[0-9A-Za-z]"),
+            ("alpha", "[A-Za-z]"),
+            ("ascii", "[]\0- \"-,.0-\\_-\x7f!^-]"),
+            ("blank", "[\t ]"),
+            ("cntrl", "[\0-\x1f\x7f]"),
+            ("digit", "[0-9]"),
+            ("graph", "[]\"-,.0-\\_-~!^-]"),
+            ("lower", "[a-z]"),
+            ("print", "[] \"-,.0-\\_-~!^-]"),
+            ("punct", "[]\"-,.:-@[-\\_-`{-~!^-]"),
+            ("space", "[\t-\r ]"),
+            ("upper", "[A-Z]"),
+            ("word", "[0-9A-Z_a-z]"),
+            ("xdigit", "[0-9A-Fa-f]"),
+        ];
+        for (name, want) in cases {
+            assert_eq!(
+                expand_posix_classes(&format!("[[:{name}:]]")),
+                want,
+                "{name}"
+            );
+        }
+    }
+
+    #[test]
+    fn a_negated_class_also_excludes_the_separator() {
+        assert_eq!(expand_posix_classes("[![:digit:]]"), "[!/0-9]");
+        assert_eq!(expand_posix_classes("[^[:digit:]]"), "[!/0-9]");
+    }
+
+    #[test]
+    fn classes_and_characters_combine() {
+        assert_eq!(expand_posix_classes("[a[:digit:]]"), "[0-9a]");
+        assert_eq!(expand_posix_classes("[[:digit:][:upper:]]"), "[0-9A-Z]");
+        assert_eq!(expand_posix_classes("[[:digit:]x-z]"), "[0-9x-z]");
+    }
+
+    #[test]
+    fn a_dash_after_a_class_is_literal() {
+        assert_eq!(expand_posix_classes("[[:digit:]-z]"), "[0-9z-]");
+        assert_eq!(expand_posix_classes("[[:digit:]-]"), "[0-9-]");
+        assert_eq!(expand_posix_classes("[[:digit:]x-]"), "[0-9x-]");
+    }
+
+    #[test]
+    fn a_leading_close_bracket_is_a_member() {
+        assert_eq!(expand_posix_classes("[][:digit:]]"), "[]0-9]");
+    }
+
+    #[test]
+    fn an_unknown_class_contributes_nothing() {
+        assert_eq!(expand_posix_classes("[[:foo:]a]"), "[a]");
+        assert_eq!(expand_posix_classes("[[:DIGIT:]]"), "[\0]");
+        assert_eq!(expand_posix_classes("[![:foo:]]"), "[!/]");
+    }
+
+    #[test]
+    fn a_lone_bang_is_kept_off_the_negation_position() {
+        assert_eq!(expand_posix_classes("[[:foo:]!]"), "[\0!]");
+        assert_eq!(expand_posix_classes("[[:foo:]^]"), "[\0^]");
+    }
+
+    #[test]
+    fn a_range_ending_in_a_class_matches_nothing() {
+        assert_eq!(expand_posix_classes("[x-[:digit:]]"), "[\0]");
+        assert_eq!(expand_posix_classes("[!x-[:digit:]]"), "[\0]");
+    }
+
+    #[test]
+    fn a_reversed_range_beside_a_class_is_empty() {
+        assert_eq!(expand_posix_classes("[z-a[:digit:]]"), "[0-9]");
+    }
+
+    #[test]
+    fn a_bracket_without_a_class_is_untouched() {
+        for pattern in [
+            "[a-c]*",
+            "[!x].md",
+            "[]]",
+            "[[:a]",
+            "{a,b}/[xy]",
+            "[a:digit:]]",
+            "[[:a][:digit:]]",
+        ] {
+            assert_eq!(expand_posix_classes(pattern), pattern);
+        }
+    }
+
+    #[test]
+    fn an_unclosed_bracket_is_untouched() {
+        assert_eq!(expand_posix_classes("[[:digit:]"), "[[:digit:]");
+        assert_eq!(expand_posix_classes("a["), "a[");
+    }
+
+    #[test]
+    fn an_escaped_bracket_is_untouched() {
+        assert_eq!(expand_posix_classes("\\[[:digit:]]"), "\\[[:digit:]]");
+        assert_eq!(expand_posix_classes("a\\"), "a\\");
+    }
+
+    #[test]
+    fn text_around_brackets_is_kept() {
+        assert_eq!(
+            expand_posix_classes("**/v[[:digit:]]/*.md"),
+            "**/v[0-9]/*.md"
+        );
+    }
+}