From 7567d211daa3826bcdb0e7613b55c33976938327 Mon Sep 17 00:00:00 2001
From: firest <enlanubo@gmail.com>
Date: Mon, 7 Aug 2023 17:40:59 +0800
Subject: [PATCH] fix(placeholder): porting fix to support utf8 key in
 placeholder

---
 apps/emqx_utils/src/emqx_placeholder.erl      |  8 ++-
 .../test/emqx_placeholder_SUITE.erl           | 50 +++++++++++++++++++
 2 files changed, 56 insertions(+), 2 deletions(-)

diff --git a/apps/emqx_utils/src/emqx_placeholder.erl b/apps/emqx_utils/src/emqx_placeholder.erl
index edf4123e4..0f677236d 100644
--- a/apps/emqx_utils/src/emqx_placeholder.erl
+++ b/apps/emqx_utils/src/emqx_placeholder.erl
@@ -48,9 +48,13 @@
 
 -define(PH_VAR_THIS, '$this').
 
--define(EX_PLACE_HOLDER, "(\\$\\{[a-zA-Z0-9\\._]+\\})").
+%% To match any pattern starts with '$' and followed by '{', and closed by a '}' char:
+%% e.g. for string "a${abc}bb", "${abc}" will be matched.
+%% Note this is non-greedy matching
+%% e.g. if "${{abc}}" is given, the "${{abc}" should be matched, NOT "${{abc}}".
+-define(EX_PLACE_HOLDER, "(\\$\\{[^}]+\\})").
 
--define(EX_PLACE_HOLDER_DOUBLE_QUOTE, "(\\$\\{[a-zA-Z0-9\\._]+\\}|\"\\$\\{[a-zA-Z0-9\\._]+\\}\")").
+-define(EX_PLACE_HOLDER_DOUBLE_QUOTE, "(\\$\\{[^}]+\\}|\"\\$\\{[^}]+\\}\")").
 
 %% Space and CRLF
 -define(EX_WITHE_CHARS, "\\s").
diff --git a/apps/emqx_utils/test/emqx_placeholder_SUITE.erl b/apps/emqx_utils/test/emqx_placeholder_SUITE.erl
index 81bf0853a..f813656f2 100644
--- a/apps/emqx_utils/test/emqx_placeholder_SUITE.erl
+++ b/apps/emqx_utils/test/emqx_placeholder_SUITE.erl
@@ -206,3 +206,53 @@ t_preproc_tmpl_deep(_) ->
         #{<<"${a}">> => [<<"1">>, "c", 2, 3.0, '${d}', {[<<"1.0">>], 0}]},
         emqx_placeholder:proc_tmpl_deep(Tmpl1, Selected)
     ).
+
+t_proc_tmpl_arbitrary_var_name(_) ->
+    Selected = #{
+        <<"中"/utf8>> => <<"1">>,
+        <<"中-1"/utf8>> => <<"1-1">>,
+        <<"-_+=<>,/?:;\"'\\[]|">> => 1,
+        <<"-_+=<>,">> => #{<<"/?:;\"'\\[]|">> => 2},
+        <<"!@#$%^&*()">> => 1.0,
+        <<"d">> => #{
+            <<"$ff">> => <<"oo">>,
+            <<"${f">> => <<"hi">>,
+            <<"${f}">> => <<"qq">>
+        }
+    },
+    Tks = emqx_placeholder:preproc_tmpl(
+        <<
+            "a:${中},a:${中-1},b:${-_+=<>,/?:;\"'\\[]|},"
+            "b:${-_+=<>,./?:;\"'\\[]|},c:${!@#$%^&*()},d:${d.$ff},d1:${d.${f}}"/utf8
+        >>
+    ),
+    ?assertEqual(
+        <<"a:1,a:1-1,b:1,b:2,c:1.0,d:oo,d1:hi}">>,
+        emqx_placeholder:proc_tmpl(Tks, Selected)
+    ).
+
+t_proc_tmpl_arbitrary_var_name_double_quote(_) ->
+    Selected = #{
+        <<"中"/utf8>> => <<"1">>,
+        <<"中-1"/utf8>> => <<"1-1">>,
+        <<"-_+=<>,/?:;\"'\\[]|">> => 1,
+        <<"-_+=<>,">> => #{<<"/?:;\"'\\[]|">> => 2},
+        <<"!@#$%^&*()">> => 1.0,
+        <<"d">> => #{
+            <<"$ff">> => <<"oo">>,
+            <<"${f">> => <<"hi">>,
+            <<"${f}">> => <<"qq">>
+        }
+    },
+    Tks = emqx_placeholder:preproc_tmpl(
+        <<
+            "a:\"${中}\",a:\"${中-1}\",b:\"${-_+=<>,/?:;\"'\\[]|}\","
+            "b:\"${-_+=<>,./?:;\"'\\[]|}\",c:\"${!@#$%^&*()}\",d:\"${d.$ff}\",d1:\"${d.${f}\"}"/utf8
+        >>,
+        #{strip_double_quote => true}
+    ),
+    ct:print("TKs:~p~n", [Tks]),
+    ?assertEqual(
+        <<"a:1,a:1-1,b:1,b:2,c:1.0,d:oo,d1:hi}">>,
+        emqx_placeholder:proc_tmpl(Tks, Selected)
+    ).