Ticket #286: split-encoding-charset-labeling-from-init.patch

include/parrot/encoding.h

@@ -152 +152 @@
         __attribute__nonnull__(3);
 
 void parrot_deinit_encodings(void);
-void parrot_init_encodings_2(void);
+void Parrot_str_internal_register_encoding_names(PARROT_INTERP)
+        __attribute__nonnull__(1);
+
 #define ASSERT_ARGS_Parrot_default_encoding __attribute__unused__ int _ASSERT_ARGS_CHECK = 0
 #define ASSERT_ARGS_Parrot_encoding_c_name __attribute__unused__ int _ASSERT_ARGS_CHECK = 0
 #define ASSERT_ARGS_Parrot_encoding_name __attribute__unused__ int _ASSERT_ARGS_CHECK = 0
@@ -180 +182 @@
     || PARROT_ASSERT_ARG(encodingname) \
     || PARROT_ASSERT_ARG(encoding)
 #define ASSERT_ARGS_parrot_deinit_encodings __attribute__unused__ int _ASSERT_ARGS_CHECK = 0
-#define ASSERT_ARGS_parrot_init_encodings_2 __attribute__unused__ int _ASSERT_ARGS_CHECK = 0
+#define ASSERT_ARGS_Parrot_str_internal_register_encoding_names \
+     __attribute__unused__ int _ASSERT_ARGS_CHECK = \
+       PARROT_ASSERT_ARG(interp)
 /* Don't modify between HEADERIZER BEGIN / HEADERIZER END.  Your changes will be lost. */
 /* HEADERIZER END: src/string/encoding.c */
 

include/parrot/string_funcs.h

@@ -298 +298 @@
     ARGIN_NULLOK(const char *buffer),
     UINTVAL len,
     ARGIN(const ENCODING *encoding),
-    ARGIN_NULLOK(const CHARSET *charset),
+    ARGIN(const CHARSET *charset),
     UINTVAL flags)
         __attribute__nonnull__(1)
-        __attribute__nonnull__(4);
+        __attribute__nonnull__(4)
+        __attribute__nonnull__(5);
 
 PARROT_EXPORT
 PARROT_CANNOT_RETURN_NULL
@@ -638 +639 @@
     || PARROT_ASSERT_ARG(s)
 #define ASSERT_ARGS_Parrot_str_new_init __attribute__unused__ int _ASSERT_ARGS_CHECK = \
        PARROT_ASSERT_ARG(interp) \
-    || PARROT_ASSERT_ARG(encoding)
+    || PARROT_ASSERT_ARG(encoding) \
+    || PARROT_ASSERT_ARG(charset)
 #define ASSERT_ARGS_Parrot_str_new_noinit __attribute__unused__ int _ASSERT_ARGS_CHECK = \
        PARROT_ASSERT_ARG(interp)
 #define ASSERT_ARGS_Parrot_str_not_equal __attribute__unused__ int _ASSERT_ARGS_CHECK = \

src/string/api.c

@@ -762 +762 @@
 PARROT_CANNOT_RETURN_NULL
 STRING *
 Parrot_str_new_init(PARROT_INTERP, ARGIN_NULLOK(const char *buffer), UINTVAL len,
-        ARGIN(const ENCODING *encoding), ARGIN_NULLOK(const CHARSET *charset), UINTVAL flags)
+        ARGIN(const ENCODING *encoding), ARGIN(const CHARSET *charset), UINTVAL flags)
 {
     ASSERT_ARGS(Parrot_str_new_init)
     DECL_CONST_CAST;

src/string/charset.c

@@ -63 +63 @@
 /* HEADERIZER BEGIN: static */
 /* Don't modify between HEADERIZER BEGIN / HEADERIZER END.  Your changes will be lost. */
 
+static void Parrot_str_internal_register_charset_names(PARROT_INTERP)
+        __attribute__nonnull__(1);
+
 static INTVAL register_charset(PARROT_INTERP,
     ARGIN(const char *charsetname),
     ARGIN(CHARSET *charset))
@@ -73 +76 @@
 static void register_static_converters(PARROT_INTERP)
         __attribute__nonnull__(1);
 
+#define ASSERT_ARGS_Parrot_str_internal_register_charset_names \
+     __attribute__unused__ int _ASSERT_ARGS_CHECK = \
+       PARROT_ASSERT_ARG(interp)
 #define ASSERT_ARGS_register_charset __attribute__unused__ int _ASSERT_ARGS_CHECK = \
        PARROT_ASSERT_ARG(interp) \
     || PARROT_ASSERT_ARG(charsetname) \
@@ -340 +346 @@
                 (n + 1) * sizeof (One_charset));
     all_charsets->n_charsets++;
     all_charsets->set[n].charset = charset;
-    all_charsets->set[n].name = Parrot_str_new_constant(interp, charsetname);
     all_charsets->set[n].n_converters = 0;
 
     return 1;
@@ -348 +353 @@
 
 /*
 
+=item C<static void Parrot_str_internal_register_charset_names>
+
+Helper function for initializing characterset names. We can't create the
+STRING names until the default encodings and charsets are already initted,
+so the name generation is split into a second init stage.
+
+=cut
+
+*/
+
+static void
+Parrot_str_internal_register_charset_names(PARROT_INTERP)
+{
+    ASSERT_ARGS(Parrot_str_internal_register_charset_names)
+    int n;
+    for (n = 0; n < all_charsets->n_charsets; n++)
+        all_charsets->set[n].name =
+            Parrot_str_new_constant(interp, all_charsets->set[n].charset->name);
+}
+
+/*
+
 =item C<static void register_static_converters>
 
 Registers several standard converters between common charsets, including:
@@ -462 +489 @@
     Parrot_charset_binary_init(interp);
     Parrot_charset_unicode_init(interp);
 
-    /*
-     * now encoding strings don't have a charset yet - set default
+    /* Now that the plugins are registered, we can create STRING
+     * names for them.
      */
-    parrot_init_encodings_2();
+    Parrot_str_internal_register_encoding_names(interp);
+    Parrot_str_internal_register_charset_names(interp);
+
     /*
      * now install charset converters
      */

src/string/encoding.c

@@ -58 +58 @@
 
 /*
 
-=item C<void parrot_init_encodings_2>
-
-Helper function for initializing characterset encodings. Initializes the
-C<all_encodings> array.
-
-=cut
-
-*/
-
-void
-parrot_init_encodings_2(void)
-{
-    ASSERT_ARGS(parrot_init_encodings_2)
-    const int n = all_encodings->n_encodings;
-    int i;
-
-    for (i = 0; i < n; ++i) {
-        all_encodings->enc[i].name->charset = Parrot_default_charset_ptr;
-    }
-}
-
-/*
-
 =item C<void parrot_deinit_encodings>
 
 Deinitialize encodings and free all memory used by them.
@@ -338 +315 @@
                 (n + 1) * sizeof (One_encoding));
     all_encodings->n_encodings++;
     all_encodings->enc[n].encoding = encoding;
-    all_encodings->enc[n].name = Parrot_str_new_constant(interp, encodingname);
 
     return 1;
 }
 
 /*
 
+=item C<void Parrot_str_internal_register_encoding_names>
+
+Helper function for initializing characterset encoding names. We can't create
+the STRING names until the default encodings and charsets are already initted,
+so the name generation is split into a second init stage.
+
+=cut
+
+*/
+
+
+void
+Parrot_str_internal_register_encoding_names(PARROT_INTERP)
+{
+    ASSERT_ARGS(Parrot_str_internal_register_encoding_names)
+    int n;
+    for (n = 0; n < all_encodings->n_encodings; n++)
+        all_encodings->enc[n].name =
+            Parrot_str_new_constant(interp, all_encodings->enc[n].encoding->name);
+}
+
+/*
+
 =item C<INTVAL Parrot_register_encoding>
 
 Registers a character encoding C<encoding> with name C<encodingname>.