gtk.cfg: fix the g_ascii_formatd() configuration

g_ascii_formatd(buffer, buf_len, format, d) formats exactly one double
into buffer, but every call was reported as
  g_ascii_formatd format string requires 1 parameter but only 0 are given.
  Uninitialized variable: buffer
- The double is a fixed argument: with it configured as arg 4 the format
  check did not count it. Leave it to the format string check, which now
  also checks its type.
- buffer is written, not read: it need not be initialized or a string.
  Its size must be at least buf_len.

Author: pduchi

cfg/gtk.cfg

@@ -3285,8 +3285,7 @@
     <returnValue type="gchar*"/>
     <arg nr="1" direction="out">
       <not-null/>
-      <not-uninit/>
-      <strz/>
+      <minsize type="argvalue" arg="2"/>
     </arg>
     <arg nr="2" direction="in">
       <not-bool/>
@@ -3299,10 +3298,6 @@
       <strz/>
       <formatstr/>
     </arg>
-    <arg nr="4" direction="in">
-      <not-bool/>
-      <not-uninit/>
-    </arg>
   </function>
   <!-- gboolean g_ascii_isalnum(gchar c); -->
   <function name="g_ascii_isalnum">

test/cfg/gtk.c

@@ -655,3 +655,20 @@ int g_return_val_if_reached_test(const char *s) {
 
     return 0;
 }
+
+void g_ascii_formatd_test(double d)
+{
+    // g_ascii_formatd() writes the buffer and formats exactly one double
+    gchar buf[39];
+    g_ascii_formatd(buf, sizeof(buf), "%.2f", d);
+    g_print("%s", buf);
+
+    // cppcheck-suppress invalidPrintfArgType_sint
+    g_ascii_formatd(buf, sizeof(buf), "%d", d);
+    g_print("%s", buf);
+
+    gchar small[4];
+    // cppcheck-suppress bufferAccessOutOfBounds
+    g_ascii_formatd(small, sizeof(buf), "%.2f", d);
+    g_print("%s", small);
+}