command: add escape-ass

This adds a command to escape ASS tags to remove code duplication
between sub/osd_libass.c, console.lua, osc.lua, stats.lua and any user
script that calls mp.create_osd_overlay().

A command is used instead of scripting functions so that all clients can
use this and not just use Lua and JS ones.

osd_mangle_ass() also interprets osd-sym-cc and osd-ass-cc/{0,1}, but
since they use invalid UTF-8 characters there is no risk of escape-ass
users using them by accident, like with any OSD message.

Always replacing \n with \\N in mangle_ass() even when it is not called
by escape-ass doesn't seem to cause any issue, but I made it conditional
anyway to avoid changing how all OSD messages are treated unnecessarily.

7 files changed, 41 insertions, 28 deletions

diff --git a/DOCS/interface-changes/escape-ass.txt b/DOCS/interface-changes/escape-ass.txt
new file mode 100644
index 0000000000..93ec1ca5e5
--- /dev/null
+++ b/DOCS/interface-changes/escape-ass.txt
@@ -0,0 +1 @@
+add the `escape-ass` command
diff --git a/DOCS/man/input.rst b/DOCS/man/input.rst
index cf4a43636b..bb353cd117 100644
--- a/DOCS/man/input.rst
+++ b/DOCS/man/input.rst
@@ -1259,6 +1259,18 @@ Input Commands that are Possibly Subject to Change
         use the ``mp.create_osd_overlay()`` helper instead of invoking this
         command directly.
 
+``escape-ass <text>``
+    Modify ``text`` so that commands and functions that interpret ASS tags,
+    such as ``osd-overlay`` and ``mp.create_osd_overlay``, will display it
+    verbatim, and return it. This can only be used through the client API or
+    from a script using ``mp.command_native``.
+
+    .. admonition:: Example
+
+        ``mp.osd_message(mp.command_native({"escape-ass", "foo {bar}"}))``
+
+        This line of Lua prints "foo \\{bar}" on the OSD.
+
 ``script-message [<arg1> [<arg2> [...]]]``
     Send a message to all clients, and pass it the following list of arguments.
     What this message means, how many arguments it takes, and what the arguments
diff --git a/player/command.c b/player/command.c
index 27fadf91b1..ce7d6d5541 100644
--- a/player/command.c
+++ b/player/command.c
@@ -5582,6 +5582,19 @@ static void cmd_expand_path(void *p)
     };
 }
 
+static void cmd_escape_ass(void *p)
+{
+    struct mp_cmd_ctx *cmd = p;
+    bstr dst = {0};
+
+    osd_mangle_ass(&dst, cmd->args[0].v.s, true);
+
+    cmd->result = (mpv_node){
+        .format = MPV_FORMAT_STRING,
+        .u.string = dst.len ? (char *)dst.start : talloc_strdup(NULL, ""),
+    };
+}
+
 static struct load_action get_load_action(struct MPContext *mpctx, int action_flag)
 {
     switch (action_flag) {
@@ -6679,6 +6692,8 @@ const struct mp_cmd_def mp_cmds[] = {
         .is_noisy = true },
     { "expand-path", cmd_expand_path, { {"text", OPT_STRING(v.s)} },
         .is_noisy = true },
+    { "escape-ass", cmd_escape_ass, { {"text", OPT_STRING(v.s)} },
+        .is_noisy = true },
     { "show-progress", cmd_show_progress, .allow_auto_repeat = true,
         .is_noisy = true },
 
diff --git a/player/lua/console.lua b/player/lua/console.lua
index 02cf890791..238216445d 100644
--- a/player/lua/console.lua
+++ b/player/lua/console.lua
@@ -229,19 +229,7 @@ end
 
 -- Escape a string for verbatim display on the OSD
 function ass_escape(str)
-    -- There is no escape for '\' in ASS (I think?) but '\' is used verbatim if
-    -- it isn't followed by a recognised character, so add a zero-width
-    -- non-breaking space
-    str = str:gsub('\\', '\\\239\187\191')
-    str = str:gsub('{', '\\{')
-    str = str:gsub('}', '\\}')
-    -- Precede newlines with a ZWNBSP to prevent ASS's weird collapsing of
-    -- consecutive newlines
-    str = str:gsub('\n', '\239\187\191\\N')
-    -- Turn leading spaces into hard spaces to prevent ASS from stripping them
-    str = str:gsub('\\N ', '\\N\\h')
-    str = str:gsub('^ ', '\\h')
-    return str
+    return mp.command_native({'escape-ass', str})
 end
 
 -- Takes a list of strings, a max width in characters and
diff --git a/player/lua/stats.lua b/player/lua/stats.lua
index 4d25a6c6bf..2feded2808 100644
--- a/player/lua/stats.lua
+++ b/player/lua/stats.lua
@@ -129,9 +129,6 @@ local function graph_add_value(graph, value)
     graph.max = max(graph.max, value)
 end
 
--- "\\<U+2060>" in UTF-8 (U+2060 is WORD-JOINER)
-local ESC_BACKSLASH = "\\" .. string.char(0xE2, 0x81, 0xA0)
-
 local function no_ASS(t)
     if not o.use_ass then
         return t
@@ -139,16 +136,7 @@ local function no_ASS(t)
         -- mp.osd_message supports ass-escape using osd-ass-cc/{0|1}
         return ass_stop .. t .. ass_start
     else
-        -- mp.set_osd_ass doesn't support ass-escape. roll our own.
-        -- similar to mpv's sub/osd_libass.c:mangle_ass(...), excluding
-        -- space after newlines because no_ASS is not used with multi-line.
-        -- space at the beginning is replaced with "\\h" because it matters
-        -- at the beginning of a line, and we can't know where our output
-        -- ends up. no issue if it ends up at the middle of a line.
-        return tostring(t)
-               :gsub("\\", ESC_BACKSLASH)
-               :gsub("{", "\\{")
-               :gsub("^ ", "\\h")
+        return mp.command_native({"escape-ass", tostring(t)})
     end
 end
 
diff --git a/sub/osd.h b/sub/osd.h
index 5fc268b5a6..2677847da5 100644
--- a/sub/osd.h
+++ b/sub/osd.h
@@ -245,5 +245,6 @@ void osd_set_external(struct osd_state *osd, struct osd_external_ass *ov);
 void osd_set_external_remove_owner(struct osd_state *osd, void *owner);
 void osd_get_text_size(struct osd_state *osd, int *out_screen_h, int *out_font_h);
 void osd_get_function_sym(char *buffer, size_t buffer_size, int osd_function);
+void osd_mangle_ass(bstr *dst, const char *in, bool replace_newlines);
 
 #endif /* MPLAYER_SUB_H */
diff --git a/sub/osd_libass.c b/sub/osd_libass.c
index 16d94b3cbf..f2de27bd64 100644
--- a/sub/osd_libass.c
+++ b/sub/osd_libass.c
@@ -193,7 +193,7 @@ void osd_get_function_sym(char *buffer, size_t buffer_size, int osd_function)
     snprintf(buffer, buffer_size, "\xFF%c", osd_function);
 }
 
-static void mangle_ass(bstr *dst, const char *in)
+void osd_mangle_ass(bstr *dst, const char *in, bool replace_newlines)
 {
     const char *start = in;
     bool escape_ass = true;
@@ -213,6 +213,14 @@ static void mangle_ass(bstr *dst, const char *in)
         }
         if (escape_ass && *in == '{')
             bstr_xappend(NULL, dst, bstr0("\\"));
+        // Replace newlines with \N for escape-ass. This is necessary to apply
+        // ASS tags past newlines and to preserve consecutive newlines with
+        // osd-overlay because update_external() adds a ASS event per line.
+        if (replace_newlines && *in == '\n') {
+            bstr_xappend(NULL, dst, bstr0("\\N"));
+            in += 1;
+            continue;
+        }
         // Libass will strip leading whitespace
         if (in[0] == ' ' && (in == start || in[-1] == '\n')) {
             bstr_xappend(NULL, dst, bstr0("\\h"));
@@ -231,7 +239,7 @@ static ASS_Event *add_osd_ass_event_escaped(ASS_Track *track, const char *style,
                                             const char *text)
 {
     bstr buf = {0};
-    mangle_ass(&buf, text);
+    osd_mangle_ass(&buf, text, false);
     ASS_Event *e = add_osd_ass_event(track, style, buf.start);
     talloc_free(buf.start);
     return e;