From 2e379710ca4ca1739788beb502a55d8bd596dc16 Mon Sep 17 00:00:00 2001
From: Jonas Maier <jonas@x77.dev>
Date: Sat, 6 Jun 2026 22:55:27 +0200
Subject: regex: submatches now go into variables

---
 src/lib.rs                             | 27 +++++++++++++++++++++++++--
 src/run/mod.rs                         | 13 ++++++++++---
 test-cases/case4_submatches/script.sh  |  8 ++++++++
 test-cases/case4_submatches/stdout.txt |  1 +
 tests/scripts.rs                       |  5 +++++
 5 files changed, 49 insertions(+), 5 deletions(-)
 create mode 100644 test-cases/case4_submatches/script.sh
 create mode 100644 test-cases/case4_submatches/stdout.txt

diff --git a/src/lib.rs b/src/lib.rs
index 825e216..1b5a814 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -18,6 +18,7 @@ use std::sync::{Arc, Mutex};
 
 pub mod ansi;
 pub mod basedir;
+mod bitset;
 pub mod completion;
 pub mod consts;
 pub mod ctrlc;
@@ -31,6 +32,7 @@ pub mod panic;
 pub mod parse;
 pub mod ps1;
 pub mod raw;
+pub mod regex;
 pub mod reload;
 pub mod run;
 pub mod rw;
@@ -38,8 +40,6 @@ pub mod serialization;
 pub mod syntax_highlighting;
 pub mod variants;
 pub mod wait;
-pub mod regex;
-mod bitset;
 
 use raw::*;
 
@@ -49,6 +49,8 @@ use crate::history::HistoryEntry;
 use crate::line::Line;
 use crate::parse::{Block, ExpString, Parse, PostExpansion};
 use crate::ps1::Prompt;
+use crate::regex::bc::RegexCompilationError;
+use crate::regex::{CompiledPattern, Pattern, RegexEngine};
 
 macro_rules! print {
     ($($x:tt)*) => {{
@@ -123,6 +125,8 @@ pub struct Session {
     prompt: Prompt,
 
     terminal_input: Option<ansi::TerminalInput>,
+
+    regex_cache: HashMap<Pattern, Result<Arc<CompiledPattern>, RegexCompilationError>>,
 }
 
 impl Session {
@@ -148,6 +152,7 @@ impl Session {
             prompt,
             vars,
             terminal_input: None,
+            regex_cache: HashMap::new(),
         }
     }
 }
@@ -345,6 +350,24 @@ impl Session {
         self.line.highlight_syntax(&mut self.highlighter)?;
         Ok(())
     }
+
+    fn cached_regex(
+        se: &Mutex<Self>,
+        pat: Pattern,
+    ) -> Result<Arc<CompiledPattern>, RegexCompilationError> {
+        let mut this = se.lock().unwrap();
+        if let Some(re) = this.regex_cache.get(&pat) {
+            re.clone()
+        } else {
+            let res = CompiledPattern::compile(pat.clone());
+            let res = match res {
+                Ok(re) => Ok(Arc::new(re)),
+                Err(e) => Err(e),
+            };
+            this.regex_cache.insert(pat, res.clone());
+            res
+        }
+    }
 }
 
 const DEFAULT_PROFILE: &[u8] = include_bytes!("profile");
diff --git a/src/run/mod.rs b/src/run/mod.rs
index 64522d9..3420c30 100644
--- a/src/run/mod.rs
+++ b/src/run/mod.rs
@@ -457,10 +457,17 @@ impl Executor {
         stdout: OutputWriter,
     ) -> SpawnedCmd {
         for branch in c.branches.into_iter() {
-            // TODO: do not compile every time
-            if let Ok(compiled) = branch.pattern.try_compile()
-                && compiled.matches(&c.discriminant)
+            if let Ok(compiled) = Session::cached_regex(self.se.as_ref(), branch.pattern)
+                && let Some(mtch) = compiled.run(&c.discriminant)
             {
+                // put submatches into variables
+                if !mtch.submatches.is_empty() {
+                    let mut se = self.se.lock().unwrap();
+                    for (var, match_loc) in mtch.submatches.into_iter() {
+                        let val = c.discriminant[match_loc].to_vec();
+                        se.vars.set(var, val);
+                    }
+                }
                 return self.execute_block(branch.block, stdin, stdout);
             }
         }
diff --git a/test-cases/case4_submatches/script.sh b/test-cases/case4_submatches/script.sh
new file mode 100644
index 0000000..75e7f6f
--- /dev/null
+++ b/test-cases/case4_submatches/script.sh
@@ -0,0 +1,8 @@
+set file_name = 'taxes.pdf'
+
+case $file_name {
+    (?<base>.*)[.](?<ext>.*) { }
+    .* { echo format error }
+}
+
+echo "file base is $base, file extension is $ext"
diff --git a/test-cases/case4_submatches/stdout.txt b/test-cases/case4_submatches/stdout.txt
new file mode 100644
index 0000000..570da9d
--- /dev/null
+++ b/test-cases/case4_submatches/stdout.txt
@@ -0,0 +1 @@
+file base is taxes, file extension is pdf
diff --git a/tests/scripts.rs b/tests/scripts.rs
index 2a7a6f8..69a81ce 100644
--- a/tests/scripts.rs
+++ b/tests/scripts.rs
@@ -21,6 +21,11 @@ fn case3_negative_lookahead() {
     common::test_case("case3_negative_lookahead", include_bytes!("../test-cases/case3_negative_lookahead/script.sh"), include_bytes!("../test-cases/case3_negative_lookahead/stdout.txt"));
 }
 
+#[test]
+fn case4_submatches() {
+    common::test_case("case4_submatches", include_bytes!("../test-cases/case4_submatches/script.sh"), include_bytes!("../test-cases/case4_submatches/stdout.txt"));
+}
+
 #[test]
 fn comment() {
     common::test_case("comment", include_bytes!("../test-cases/comment/script.sh"), include_bytes!("../test-cases/comment/stdout.txt"));
-- 
cgit v1.2.3