From a8877d334c2f23a9b11076a80288c46d617530fc Mon Sep 17 00:00:00 2001
From: Mitchell Hashimoto <mitchell.hashimoto@gmail.com>
Date: Sun, 24 Sep 2023 20:32:35 -0700
Subject: [PATCH] config: unicode range parser is more lenient about whitespace

---
 src/config/Config.zig | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)

diff --git a/src/config/Config.zig b/src/config/Config.zig
index 1327aa4f3..2a422dbdf 100644
--- a/src/config/Config.zig
+++ b/src/config/Config.zig
@@ -1571,7 +1571,7 @@ pub const RepeatableCodepointMap = struct {
     ///   U+1234-5678
     ///   U+1234,U+5678
     ///   U+1234-5678,U+5678
-    ///   U+1234,U+5678-9ABC
+    ///   U+1234,U+5678-U+9ABC
     ///
     /// etc.
     const UnicodeRangeParser = struct {
@@ -1586,12 +1586,16 @@ pub const RepeatableCodepointMap = struct {
             const start = try self.parseCodepoint();
             if (self.eof()) return .{ start, start };
 
+            // We're allowed to have any whitespace here
+            self.consumeWhitespace();
+
             // Otherwise we expect either a range or a comma
             switch (self.input[self.i]) {
                 // Comma means we have another codepoint but in a different
                 // range so we return our current codepoint.
                 ',' => {
                     self.advance();
+                    self.consumeWhitespace();
                     if (self.eof()) return error.InvalidValue;
                     return .{ start, start };
                 },
@@ -1599,10 +1603,14 @@ pub const RepeatableCodepointMap = struct {
                 // Hyphen means we have a range.
                 '-' => {
                     self.advance();
+                    self.consumeWhitespace();
                     if (self.eof()) return error.InvalidValue;
                     const end = try self.parseCodepoint();
+                    self.consumeWhitespace();
                     if (!self.eof() and self.input[self.i] != ',') return error.InvalidValue;
                     self.advance();
+                    self.consumeWhitespace();
+                    if (start > end) return error.InvalidValue;
                     return .{ start, end };
                 },
 
@@ -1610,6 +1618,15 @@ pub const RepeatableCodepointMap = struct {
             }
         }
 
+        fn consumeWhitespace(self: *UnicodeRangeParser) void {
+            while (!self.eof()) {
+                switch (self.input[self.i]) {
+                    ' ', '\t' => self.advance(),
+                    else => return,
+                }
+            }
+        }
+
         fn parseCodepoint(self: *UnicodeRangeParser) !u21 {
             if (self.input[self.i] != 'U') return error.InvalidValue;
             self.advance();