From d3a85b68347071d8d93ec796a38c707483d7b272 Mon Sep 17 00:00:00 2001
From: Robert Hensing <robert@roberthensing.nl>
Date: Wed, 6 Dec 2023 15:14:41 +0100
Subject: [PATCH] isValidSchemeName: Add function

---
 src/libutil/url.cc        | 17 +++++++++++++++++
 src/libutil/url.hh        |  9 +++++++++
 tests/unit/libutil/url.cc | 18 ++++++++++++++++++
 3 files changed, 44 insertions(+)

diff --git a/src/libutil/url.cc b/src/libutil/url.cc
index 57b64d607..f2d5f1782 100644
--- a/src/libutil/url.cc
+++ b/src/libutil/url.cc
@@ -3,6 +3,7 @@
 #include "util.hh"
 #include "split.hh"
 #include "canon-path.hh"
+#include "string.hh"
 
 namespace nix {
 
@@ -183,4 +184,20 @@ std::string fixGitURL(const std::string & url)
     }
 }
 
+// https://www.rfc-editor.org/rfc/rfc3986#section-3.1
+bool isValidSchemeName(std::string_view s)
+{
+    if (s.empty()) return false;
+    if (!isASCIIAlpha(s[0])) return false;
+    for (auto c : s.substr(1)) {
+        if (isASCIIAlpha(c)) continue;
+        if (isASCIIDigit(c)) continue;
+        if (c == '+') continue;
+        if (c == '-') continue;
+        if (c == '.') continue;
+        return false;
+    }
+    return true;
+}
+
 }
diff --git a/src/libutil/url.hh b/src/libutil/url.hh
index 833f54678..24806bbff 100644
--- a/src/libutil/url.hh
+++ b/src/libutil/url.hh
@@ -55,4 +55,13 @@ ParsedUrlScheme parseUrlScheme(std::string_view scheme);
    changes absolute paths into file:// URLs. */
 std::string fixGitURL(const std::string & url);
 
+/**
+ * Whether a string is valid as RFC 3986 scheme name.
+ * Colon `:` is part of the URI; not the scheme name, and therefore rejected.
+ * See https://www.rfc-editor.org/rfc/rfc3986#section-3.1
+ *
+ * Does not check whether the scheme is understood, as that's context-dependent.
+ */
+bool isValidSchemeName(std::string_view scheme);
+
 }
diff --git a/tests/unit/libutil/url.cc b/tests/unit/libutil/url.cc
index a678dad20..09fa4e218 100644
--- a/tests/unit/libutil/url.cc
+++ b/tests/unit/libutil/url.cc
@@ -344,4 +344,22 @@ namespace nix {
         ASSERT_EQ(percentDecode(e), s);
     }
 
+TEST(nix, isValidSchemeName) {
+    ASSERT_TRUE(isValidSchemeName("http"));
+    ASSERT_TRUE(isValidSchemeName("https"));
+    ASSERT_TRUE(isValidSchemeName("file"));
+    ASSERT_TRUE(isValidSchemeName("file+https"));
+    ASSERT_TRUE(isValidSchemeName("fi.le"));
+    ASSERT_TRUE(isValidSchemeName("file-ssh"));
+    ASSERT_TRUE(isValidSchemeName("file+"));
+    ASSERT_TRUE(isValidSchemeName("file."));
+    ASSERT_TRUE(isValidSchemeName("file1"));
+    ASSERT_FALSE(isValidSchemeName("file:"));
+    ASSERT_FALSE(isValidSchemeName("file/"));
+    ASSERT_FALSE(isValidSchemeName("+file"));
+    ASSERT_FALSE(isValidSchemeName(".file"));
+    ASSERT_FALSE(isValidSchemeName("-file"));
+    ASSERT_FALSE(isValidSchemeName("1file"));
+}
+
 }