From cadaef3f076fb0bebeb4b72813361b7e6aa5f186 Mon Sep 17 00:00:00 2001
From: sharkdp <davidpeter@web.de>
Date: Sun, 6 Dec 2020 15:57:33 +0100
Subject: [PATCH] Show error if pattern matches leading dot but --hidden is not
 given, closes #615

---
 src/main.rs         | 13 +++++++++++-
 src/regex_helper.rs | 48 +++++++++++++++++++++++++++++++++++++++++++++
 tests/tests.rs      | 15 ++++++++++++++
 3 files changed, 75 insertions(+), 1 deletion(-)

diff --git a/src/main.rs b/src/main.rs
index cf0f946..5208765 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -30,7 +30,7 @@ use crate::filetypes::FileTypes;
 use crate::filter::OwnerFilter;
 use crate::filter::{SizeFilter, TimeFilter};
 use crate::options::Options;
-use crate::regex_helper::pattern_has_uppercase_char;
+use crate::regex_helper::{pattern_has_uppercase_char, pattern_matches_strings_with_leading_dot};
 
 // We use jemalloc for performance reasons, see https://github.com/sharkdp/fd/pull/481
 // FIXME: re-enable jemalloc on macOS, see comment in Cargo.toml file for more infos
@@ -431,6 +431,17 @@ fn run() -> Result<ExitCode> {
             }),
     };
 
+    if cfg!(unix)
+        && config.ignore_hidden
+        && pattern_matches_strings_with_leading_dot(&pattern_regex)
+    {
+        return Err(anyhow!(
+            "The pattern seems to only match files with a leading dot, but hidden files are \
+            filtered by default. Consider adding -H/--hidden to search hidden files as well \
+            or adjust your search pattern."
+        ));
+    }
+
     let re = RegexBuilder::new(&pattern_regex)
         .case_insensitive(!config.case_sensitive)
         .dot_matches_new_line(true)
diff --git a/src/regex_helper.rs b/src/regex_helper.rs
index 4cadd96..bc7f77a 100644
--- a/src/regex_helper.rs
+++ b/src/regex_helper.rs
@@ -34,6 +34,45 @@ fn hir_has_uppercase_char(hir: &Hir) -> bool {
     }
 }
 
+/// Determine if a regex pattern only matches strings starting with a literal dot (hidden files)
+pub fn pattern_matches_strings_with_leading_dot(pattern: &str) -> bool {
+    let mut parser = ParserBuilder::new().allow_invalid_utf8(true).build();
+
+    parser
+        .parse(pattern)
+        .map(|hir| hir_matches_strings_with_leading_dot(&hir))
+        .unwrap_or(false)
+}
+
+/// See above.
+fn hir_matches_strings_with_leading_dot(hir: &Hir) -> bool {
+    use regex_syntax::hir::*;
+
+    // Note: this only really detects the simplest case where a regex starts with
+    // "^\\.", i.e. a start text anchor and a literal dot character. There are a lot
+    // of other patterns that ONLY match hidden files, e.g. ^(\\.foo|\\.bar) which are
+    // not (yet) detected by this algorithm.
+    match *hir.kind() {
+        HirKind::Concat(ref hirs) => {
+            let mut hirs = hirs.iter();
+            if let Some(hir) = hirs.next() {
+                if *hir.kind() != HirKind::Anchor(Anchor::StartText) {
+                    return false;
+                }
+            } else {
+                return false;
+            }
+
+            if let Some(hir) = hirs.next() {
+                *hir.kind() == HirKind::Literal(Literal::Unicode('.'))
+            } else {
+                false
+            }
+        }
+        _ => false,
+    }
+}
+
 #[test]
 fn pattern_has_uppercase_char_simple() {
     assert!(pattern_has_uppercase_char("A"));
@@ -50,3 +89,12 @@ fn pattern_has_uppercase_char_advanced() {
     assert!(!pattern_has_uppercase_char(r"\Acargo"));
     assert!(!pattern_has_uppercase_char(r"carg\x6F"));
 }
+
+#[test]
+fn matches_strings_with_leading_dot_simple() {
+    assert!(pattern_matches_strings_with_leading_dot("^\\.gitignore"));
+
+    assert!(!pattern_matches_strings_with_leading_dot("^.gitignore"));
+    assert!(!pattern_matches_strings_with_leading_dot("\\.gitignore"));
+    assert!(!pattern_matches_strings_with_leading_dot("^gitignore"));
+}
diff --git a/tests/tests.rs b/tests/tests.rs
index e67eb5b..f57541a 100644
--- a/tests/tests.rs
+++ b/tests/tests.rs
@@ -1699,3 +1699,18 @@ fn test_number_parsing_errors() {
 
     te.assert_failure(&["--max-results=a"]);
 }
+
+/// Print error if search pattern starts with a dot and --hidden is not set
+/// (Unix only, hidden files on Windows work differently)
+#[test]
+#[cfg(unix)]
+fn test_error_if_hidden_not_set_and_pattern_starts_with_dot() {
+    let te = TestEnv::new(&[], &[".gitignore", ".whatever", "non-hidden"]);
+
+    te.assert_failure(&["^\\.gitignore"]);
+    te.assert_failure(&["--glob", ".gitignore"]);
+
+    te.assert_output(&["--hidden", "^\\.gitignore"], ".gitignore");
+    te.assert_output(&["--hidden", "--glob", ".gitignore"], ".gitignore");
+    te.assert_output(&[".gitignore"], "");
+}