normalize html title whitespace

Author: mookums

src/browser/browser.zig

@@ -415,7 +415,7 @@ pub const Page = struct {
 
         const ccharset = try arena.dupeZ(u8, charset);
 
-        const html_doc = try parser.documentHTMLParse(reader, ccharset);
+        const html_doc = try parser.documentHTMLParse(arena, reader, ccharset);
         const doc = parser.documentHTMLToDocument(html_doc);
 
         // save a document's pointer in the page.

src/browser/css/libdom_test.zig

@@ -161,7 +161,7 @@ test "matchFirst" {
     for (testcases) |tc| {
         matcher.reset();
 
-        const doc = try parser.documentHTMLParseFromStr(tc.html);
+        const doc = try parser.documentHTMLParseFromStr(alloc, tc.html);
         defer parser.documentHTMLClose(doc) catch {};
 
         const s = css.parse(alloc, tc.q, .{}) catch |e| {

src/browser/dump.zig

@@ -196,7 +196,10 @@ fn testWriteFullHTML(comptime expected: []const u8, src: []const u8) !void {
     var buf = std.ArrayListUnmanaged(u8){};
     defer buf.deinit(testing.allocator);
 
-    const doc_html = try parser.documentHTMLParseFromStr(src);
+    var aa = std.heap.ArenaAllocator.init(testing.allocator);
+    defer aa.deinit();
+
+    const doc_html = try parser.documentHTMLParseFromStr(aa.allocator(), src);
     defer parser.documentHTMLClose(doc_html) catch {};
 
     const doc = parser.documentHTMLToDocument(doc_html);

src/browser/html/document.zig

@@ -29,6 +29,19 @@ const collection = @import("../dom/html_collection.zig");
 const Walker = @import("../dom/walker.zig").WalkerDepthFirst;
 const Cookie = @import("../storage/cookie.zig").Cookie;
 
+pub fn normalizeWhitespace(arena: std.mem.Allocator, title: []const u8) ![]const u8 {
+    var normalized = try std.ArrayListUnmanaged(u8).initCapacity(arena, title.len);
+    var tokens = std.mem.tokenizeAny(u8, title, &std.ascii.whitespace);
+
+    var prepend = false;
+    while (tokens.next()) |token| {
+        if (prepend) normalized.appendAssumeCapacity(' ') else prepend = true;
+        normalized.appendSliceAssumeCapacity(token);
+    }
+
+    return normalized.items;
+}
+
 // WEB IDL https://html.spec.whatwg.org/#the-document-object
 pub const HTMLDocument = struct {
     pub const Self = parser.DocumentHTML;
@@ -94,9 +107,10 @@ pub const HTMLDocument = struct {
         return try parser.documentHTMLGetTitle(self);
     }
 
-    pub fn set_title(self: *parser.DocumentHTML, v: []const u8) ![]const u8 {
-        try parser.documentHTMLSetTitle(self, v);
-        return v;
+    pub fn set_title(self: *parser.DocumentHTML, v: []const u8, state: *SessionState) ![]const u8 {
+        const normalized = try normalizeWhitespace(state.arena, v);
+        try parser.documentHTMLSetTitle(self, normalized);
+        return normalized;
     }
 
     pub fn _getElementsByName(self: *parser.DocumentHTML, name: []const u8, state: *SessionState) !NodeList {

src/browser/netsurf.zig

@@ -29,6 +29,7 @@ const c = @cImport({
 });
 
 const mimalloc = @import("mimalloc.zig");
+const normalizeWhitespace = @import("html/document.zig").normalizeWhitespace;
 
 // init initializes netsurf lib.
 // init starts a mimalloc heap arena for the netsurf session. The caller must
@@ -2152,12 +2153,12 @@ fn parserErr(err: HubbubErr) ParserError!void {
 
 // documentHTMLParseFromStr parses the given HTML string.
 // The caller is responsible for closing the document.
-pub fn documentHTMLParseFromStr(str: []const u8) !*DocumentHTML {
+pub fn documentHTMLParseFromStr(arena: std.mem.Allocator, str: []const u8) !*DocumentHTML {
     var fbs = std.io.fixedBufferStream(str);
-    return try documentHTMLParse(fbs.reader(), "UTF-8");
+    return try documentHTMLParse(arena, fbs.reader(), "UTF-8");
 }
 
-pub fn documentHTMLParse(reader: anytype, enc: ?[:0]const u8) !*DocumentHTML {
+pub fn documentHTMLParse(arena: std.mem.Allocator, reader: anytype, enc: ?[:0]const u8) !*DocumentHTML {
     var parser: ?*c.dom_hubbub_parser = undefined;
     var doc: ?*c.dom_document = undefined;
     var err: c.hubbub_error = undefined;
@@ -2169,7 +2170,11 @@ pub fn documentHTMLParse(reader: anytype, enc: ?[:0]const u8) !*DocumentHTML {
 
     try parseData(parser.?, reader);
 
-    return @as(*DocumentHTML, @ptrCast(doc.?));
+    const html_doc: *DocumentHTML = @ptrCast(doc.?);
+    const old_title = try documentHTMLGetTitle(html_doc);
+    const normalized = try normalizeWhitespace(arena, old_title);
+    try documentHTMLSetTitle(html_doc, normalized);
+    return html_doc;
 }
 
 pub fn documentParseFragmentFromStr(self: *Document, str: []const u8) !*DocumentFragment {

src/browser/xhr/xhr.zig

@@ -704,7 +704,7 @@ pub const XMLHttpRequest = struct {
         }
 
         var fbs = std.io.fixedBufferStream(self.response_bytes.items);
-        const doc = parser.documentHTMLParse(fbs.reader(), ccharset) catch {
+        const doc = parser.documentHTMLParse(self.arena, fbs.reader(), ccharset) catch {
             self.response_obj = .{ .Failure = {} };
             return;
         };

src/testing.zig

@@ -214,11 +214,13 @@ pub const Document = struct {
         parser.deinit();
         try parser.init();
 
+        var arena = std.heap.ArenaAllocator.init(allocator);
+
         var fbs = std.io.fixedBufferStream(html);
-        const html_doc = try parser.documentHTMLParse(fbs.reader(), "utf-8");
+        const html_doc = try parser.documentHTMLParse(arena.allocator(), fbs.reader(), "utf-8");
 
         return .{
-            .arena = std.heap.ArenaAllocator.init(allocator),
+            .arena = arena,
             .doc = parser.documentHTMLToDocument(html_doc),
         };
     }
@@ -410,7 +412,7 @@ pub const JsRunner = struct {
         errdefer self.loop.deinit();
 
         var html = std.io.fixedBufferStream(opts.html);
-        const document = try parser.documentHTMLParse(html.reader(), "UTF-8");
+        const document = try parser.documentHTMLParse(arena, html.reader(), "UTF-8");
 
         self.state = .{
             .arena = arena,