I've now applied the following patch:
commit 32a48ac93621b16a2ea2cd97e71955fb5fe1d140
Author: Gisle Aas <gisle@aas.no>
Date: Mon Nov 17 11:43:22 2008 +0100
Don't hardcode source line numbers [RT#38114]
diff --git a/t/unicode-bom.t b/t/unicode-bom.t
index 34e066f..b7398cf 100644
--- a/t/unicode-bom.t
+++ b/t/unicode-bom.t
@@ -37,13 +37,17 @@ $p->eof;
$p->parse("\xFE\xFF\0\0<head>Hi there</head>");
$p->eof;
+for (@warn) {
+ s/line (\d+)/line ##/g;
+}
+
is(join("", @warn), <<EOT);
-Parsing of undecoded UTF-8 will give garbage when decoding entities at $0 line 21.
-Parsing of undecoded UTF-8 will give garbage when decoding entities at $0 line 25.
-Parsing of undecoded UTF-16 at $0 line 28.
-Parsing of undecoded UTF-16 at $0 line 31.
-Parsing of undecoded UTF-32 at $0 line 34.
-Parsing of undecoded UTF-32 at $0 line 37.
+Parsing of undecoded UTF-8 will give garbage when decoding entities at $0 line ##.
+Parsing of undecoded UTF-8 will give garbage when decoding entities at $0 line ##.
+Parsing of undecoded UTF-16 at $0 line ##.
+Parsing of undecoded UTF-16 at $0 line ##.
+Parsing of undecoded UTF-32 at $0 line ##.
+Parsing of undecoded UTF-32 at $0 line ##.
EOT
@warn = ();
