From 8b64b571934b763711b08831a7b5eb60802f172a Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Sandro=20J=C3=A4ckel?= <sandro.jaeckel@gmail.com>
Date: Fri, 19 Jun 2026 22:40:16 +0200
Subject: [PATCH] Partially revert "Add regression test for NuschtOS/search#339
 (#103)"

---
 libixx/src/string_view.rs | 35 +++++------------------------------
 libixx/src/test/search.rs | 27 ---------------------------
 2 files changed, 5 insertions(+), 57 deletions(-)

diff --git a/libixx/src/string_view.rs b/libixx/src/string_view.rs
index 00a7853..4a62ade 100644
--- a/libixx/src/string_view.rs
+++ b/libixx/src/string_view.rs
@@ -35,52 +35,27 @@ impl Display for StringView<'_, '_> {
 impl StringView<'_, '_> {
   pub fn matches(&self, search: &[Vec<&[u8]>]) -> Result<bool, IxxError> {
     let mut self_parts_start = 0;
-    // byte offset into self.parts[self_parts_start] to resume from
     let mut self_parts_start_str_idx = 0;
-    // single-part matches leave self_parts_start on the matched label,
-    // a subsequent multi-part segment needs to start at the next label
-    let mut prev_was_single_part = false;
 
     for segment in search {
-      // for eg: "programs.vim" each dot-component matches a full label exactly
-      if segment.len() > 1 {
-        if prev_was_single_part {
-          self_parts_start += 1;
-          self_parts_start_str_idx = 0;
-
-          prev_was_single_part = false;
-        }
-        for part in segment {
-          if self_parts_start >= self.parts.len() {
-              return Ok(false);
-          }
-          let self_part = self.index.resolve_reference(self.parts[self_parts_start])?;
-          if !eq_ignore_ascii_case(&self_part.data, part) {
-            return Ok(false);
-          }
-          self_parts_start += 1;
-        }
-      } else {
-        // bare word or trailing wildcard: substring should match across remaining labels
-        let part = &segment[0];
+      for part in segment {
         'outer: {
           for (self_part_idx, self_part) in self.parts[self_parts_start..].iter().enumerate() {
             let self_part = self.index.resolve_reference(*self_part)?;
 
             if let Some(idx) = ascii_ignore_case_find(&self_part.data[self_parts_start_str_idx..], part) {
               self_parts_start += self_part_idx;
-              self_parts_start_str_idx = if self_part_idx == 0 {
-                self_parts_start_str_idx + idx
+              if self_part_idx == 0 {
+                self_parts_start_str_idx += idx;
               } else {
-                idx
-              };
+                self_parts_start_str_idx = 0;
+              }
               break 'outer;
             }
             self_parts_start_str_idx = 0;
           }
           return Ok(false);
         }
-        prev_was_single_part = true;
       }
     }
 
diff --git a/libixx/src/test/search.rs b/libixx/src/test/search.rs
index 7351145..ec1b0ae 100644
--- a/libixx/src/test/search.rs
+++ b/libixx/src/test/search.rs
@@ -103,33 +103,6 @@ fn test_exact_search() {
     vec![(1, 0, "programs.nixvim.enable".to_string())]
   );
 
-  assert_eq!(
-    index.search(Some(0), "programs.vim", 10).unwrap(),
-    vec![(2, 0, "programs.vim.enable".to_string())]
-  );
-  assert_eq!(
-    index.search(Some(0), "programs.vim.enable", 10).unwrap(),
-    vec![(2, 0, "programs.vim.enable".to_string())]
-  );
-
-  // regression tests that wildcard matching also works as expected
-  assert_eq!(
-    index.search(Some(0), "programs.vim*", 10).unwrap(),
-    vec![(2, 0, "programs.vim.enable".to_string())]
-  );
-  assert_eq!(
-    index.search(Some(0), "programs.*vim.enable", 10).unwrap(),
-    vec![(2, 0, "programs.vim.enable".to_string())]
-  );
-  assert_eq!(
-    index.search(Some(0), "programs.neovim*", 10).unwrap(),
-    vec![(0, 0, "programs.neovim.enable".to_string())]
-  );
-  assert_eq!(
-    index.search(Some(0), "programs.nixvim*", 10).unwrap(),
-    vec![(1, 0, "programs.nixvim.enable".to_string())]
-  );
-
   // regression test that no out of bounds happen
   assert_eq!(
       index.search(Some(0), "programs.vim.enable.extra", 10).unwrap(),