Fix <target>_TESTS variable generated by catch_discover_tests

There were 2 separate bugs, one recently added and one ancient
(since roughly the first version of the registration script).

1) The recent refactoring of how the JSON output from Catch2 is
   parsed caused shadowing between the variable containing the
   per-test JSON fragments and the accumulator of test names for
   <target>_TESTS variable. This led to the variable containing
   both the names and the JSON fragments, and thus being completely
   wrong.

2) The test name accumulation has never accounted for characters
   that need escaping to be present in a CMake list. This means
   that e.g. test names with semicolon in them would end up with
   two elements in the test list.

Both of these are now fixed, at the cost of extra complexity and
my sanity as I had to learn more about CMake escaping rules.

(CMake escaping rules are dumb)

extras/CatchAddTests.cmake

@@ -12,14 +12,28 @@
 # exist in JSON unescaped.
 #
 #   0x01  <=> ';'   (CMake list separator)
-#   0x02  ==  element boundary marker used while splitting the tests array
-#   0x03  <-> '['   (opens a CMake bracket-argument context)
-#   0x04  <-> ']'   (closes a CMake bracket-argument context)
+#   0x02  <-> '['   (opens a CMake bracket-argument context)
+#   0x03  <-> ']'   (closes a CMake bracket-argument context)
+#   0x04  ==  element boundary marker used while splitting the tests array
 #
 string(ASCII 1 _SemicolonEscape)
-string(ASCII 2 _BoundaryEscape)
-string(ASCII 3 _OpenBracketEscape)
-string(ASCII 4 _CloseBracketEscape)
+string(ASCII 2 _OpenBracketEscape)
+string(ASCII 3 _CloseBracketEscape)
+string(ASCII 4 _BoundaryEscape)
+
+
+# As far as I can tell, the only way of having unclosed '[' or ']' in the
+# list of test names we define for CTest, is to escape them into completely
+# different characters. Otherwise, they will break parsing of unrelated
+# semicolons (inserted by CMake after parsing the list from the user-facing
+# space separated format). This means that the consumers of the list have
+# to unescape them back, but that's CMake :v
+#
+#   0x02  <-> '['
+#   0x03  <-> ']'
+#
+string(ASCII 2 _OpenBracketListingEscape)
+string(ASCII 3 _CloseBracketListingEscape)
 
 
 # Placeholder bytes in the listing would break our parsing hack, so
@@ -221,6 +235,10 @@ function(catch_discover_tests_impl)
   set(script)
   set(suite)
   set(tests)
+  # Holds the list of all registered test names **as a string**, not list.
+  # This avoids issue of CMake removing semicolon escapes even inside
+  # bracket-quoted strings.
+  set(_test_names)
 
   if(WIN32)
     set(dl_paths_variable_name PATH)
@@ -431,13 +449,25 @@ function(catch_discover_tests_impl)
       string(APPEND script "${_Command}")
     endif()
 
-    list(APPEND tests "${prefix}${plain_name}${suffix}")
+    # The test name has to be escaped using the same rules as prepare_command
+    # uses for the arguments, so that it keeps being single element in list
+    # even with weird characters and semicolons.
+    set(full_list_name "${prefix}${plain_name}${suffix}")
+    string(REPLACE ";" "\\;" full_list_name "${full_list_name}")
+    string(REPLACE "[" "${_OpenBracketListingEscape}" full_list_name "${full_list_name}")
+    string(REPLACE "]" "${_CloseBracketListingEscape}" full_list_name "${full_list_name}")
+    if(full_list_name MATCHES "[^-./:a-zA-Z0-9_]")
+      # The space before the start of quote is important, so that we get
+      # space-separated list in the final file.
+      string(APPEND _test_names " [==[${full_list_name}]==]")
+    else()
+      string(APPEND _test_names " ${full_list_name}")
+    endif()
   endforeach()
 
   # Create a list of all discovered tests, which users may use to e.g. set
   # properties on the tests
-  prepare_command(set ${_TEST_LIST} ${tests})
-  string(APPEND script "${_Command}")
+  string(APPEND script "set(${_TEST_LIST}${_test_names})\n")
 
   # Write any script leftovers we have
   file(APPEND "${_CTEST_FILE}" "${script}")