Ensure hyphen/minus is treated as literal in regexes

Each U+002D HYPHEN-MINUS in [0-9A-Za-z_-/.\] is treated as a member of a
range. The third one, which appears to have been intended to be a
literal, is part of an empty range because the starting point
U+005F LOW LINE is greater than the endpoint U+002F SOLIDUS, resulting
in at least some grep implementations not considering U+002D, U+002F
or U+005F to match the pattern. This resulted in one of the
dbus-launch tests being unintentionally skipped when it used a
regex based on the one in the spec.

regex(7) suggests "To include a literal '-' [in a bracketed character
set], make it the first or last character".

Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=100686

diff --git a/doc/dbus-specification.xml b/doc/dbus-specification.xml
index 7110be7..0f0ca5a 100644 (file)
--- a/doc/dbus-specification.xml
+++ b/doc/dbus-specification.xml
@@ -3097,7 +3097,7 @@
         <listitem>
           <para>
             The set of optionally-escaped bytes is:
-            <literal>[0-9A-Za-z_-/.\]</literal>. To escape, each
+            <literal>[-0-9A-Za-z_/.\]</literal>. To escape, each
             <emphasis>byte</emphasis> (note, not character) which is not in the
             set of optionally-escaped bytes must be replaced with an ASCII
             percent (<literal>%</literal>) and the value of the byte in hex.

diff --git a/test/test-dbus-launch-x11.sh b/test/test-dbus-launch-x11.sh
index be0f4f3..8f04910 100755 (executable)
--- a/test/test-dbus-launch-x11.sh
+++ b/test/test-dbus-launch-x11.sh
@@ -198,7 +198,7 @@ test_xdg_runtime_dir () {
     export XDG_RUNTIME_DIR="$workdir"
     fake_uuid="ffffffffffffffffffffffffffffffff"
 
-    if echo "$workdir" | grep '[^0-9A-Za-z_-/.]'; then
+    if echo "$workdir" | grep '[^-0-9A-Za-z_/.]'; then
         test_num=$(($test_num + 1))
         echo "ok ${test_num} # SKIP - $workdir would need escaping"
         return