bpo-46712: share more global strings in deepfreeze (gh-32152)

(for gh-90868)

Author: kumaraditya303

Modules/_io/textio.c

@@ -1959,6 +1959,7 @@ _io_TextIOWrapper_read_impl(textio *self, Py_ssize_t n)
         if (chunks != NULL) {
             if (result != NULL && PyList_Append(chunks, result) < 0)
                 goto fail;
+            _Py_DECLARE_STR(empty, "");
             Py_XSETREF(result, PyUnicode_Join(&_Py_STR(empty), chunks));
             if (result == NULL)
                 goto fail;

Modules/_pickle.c

@@ -1812,7 +1812,7 @@ get_dotted_path(PyObject *obj, PyObject *name)
 {
     PyObject *dotted_path;
     Py_ssize_t i, n;
-
+    _Py_DECLARE_STR(dot, ".");
     dotted_path = PyUnicode_Split(name, &_Py_STR(dot), -1);
     if (dotted_path == NULL)
         return NULL;

Objects/unicodeobject.c

@@ -258,6 +258,7 @@ static int unicode_is_singleton(PyObject *unicode);
 // Return a borrowed reference to the empty string singleton.
 static inline PyObject* unicode_get_empty(void)
 {
+    _Py_DECLARE_STR(empty, "");
     return &_Py_STR(empty);
 }
 

Python/compile.c

@@ -782,6 +782,7 @@ compiler_set_qualname(struct compiler *c)
     }
 
     if (base != NULL) {
+        _Py_DECLARE_STR(dot, ".");
         name = PyUnicode_Concat(base, &_Py_STR(dot));
         Py_DECREF(base);
         if (name == NULL)
@@ -3945,6 +3946,7 @@ compiler_from_import(struct compiler *c, stmt_ty s)
         ADDOP_NAME(c, IMPORT_NAME, s->v.ImportFrom.module, names);
     }
     else {
+        _Py_DECLARE_STR(empty, "");
         ADDOP_NAME(c, IMPORT_NAME, &_Py_STR(empty), names);
     }
     for (i = 0; i < n; i++) {
@@ -4885,6 +4887,7 @@ compiler_joined_str(struct compiler *c, expr_ty e)
 
     Py_ssize_t value_count = asdl_seq_LEN(e->v.JoinedStr.values);
     if (value_count > STACK_USE_GUIDELINE) {
+        _Py_DECLARE_STR(empty, "");
         ADDOP_LOAD_CONST_NEW(c, &_Py_STR(empty));
         ADDOP_NAME(c, LOAD_METHOD, &_Py_ID(join), names);
         ADDOP_I(c, BUILD_LIST, 0);

Tools/scripts/deepfreeze.py

@@ -18,7 +18,7 @@
 from generate_global_objects import get_identifiers_and_strings
 
 verbose = False
-identifiers = get_identifiers_and_strings()[0]
+identifiers, strings = get_identifiers_and_strings()
 
 def isprintable(b: bytes) -> bool:
     return all(0x20 <= c < 0x7f for c in b)
@@ -168,6 +168,8 @@ def generate_bytes(self, name: str, b: bytes) -> str:
         return f"& {name}.ob_base.ob_base"
 
     def generate_unicode(self, name: str, s: str) -> str:
+        if s in strings:
+            return f"&_Py_STR({strings[s]})"
         if s in identifiers:
             return f"&_Py_ID({s})"
         if re.match(r'\A[A-Za-z0-9_]+\Z', s):

Tools/scripts/generate_global_objects.py

@@ -1,20 +1,13 @@
 import contextlib
-import glob
 import io
 import os.path
 import re
-import sys
-
 
 __file__ = os.path.abspath(__file__)
 ROOT = os.path.dirname(os.path.dirname(os.path.dirname(__file__)))
 INTERNAL = os.path.join(ROOT, 'Include', 'internal')
 
 
-STRING_LITERALS = {
-    'empty': '',
-    'dot': '.',
-}
 IGNORED = {
     'ACTION',  # Python/_warnings.c
     'ATTR',  # Python/_warnings.c and Objects/funcobject.c
@@ -211,7 +204,7 @@ def generate_global_strings(identifiers, strings):
         printer.write(START)
         with printer.block('struct _Py_global_strings', ';'):
             with printer.block('struct', ' literals;'):
-                for name, literal in sorted(strings.items()):
+                for literal, name in sorted(strings.items(), key=lambda x: x[1]):
                     printer.write(f'STRUCT_FOR_STR({name}, "{literal}")')
             outfile.write('\n')
             with printer.block('struct', ' identifiers;'):
@@ -276,7 +269,7 @@ def generate_runtime_init(identifiers, strings):
                 # Global strings.
                 with printer.block('.strings =', ','):
                     with printer.block('.literals =', ','):
-                        for name, literal in sorted(strings.items()):
+                        for literal, name in sorted(strings.items(), key=lambda x: x[1]):
                             printer.write(f'INIT_STR({name}, "{literal}"),')
                     with printer.block('.identifiers =', ','):
                         for name in sorted(identifiers):
@@ -297,15 +290,15 @@ def generate_runtime_init(identifiers, strings):
 
 def get_identifiers_and_strings() -> 'tuple[set[str], dict[str, str]]':
     identifiers = set(IDENTIFIERS)
-    strings = dict(STRING_LITERALS)
+    strings = {}
     for name, string, *_ in iter_global_strings():
         if string is None:
             if name not in IGNORED:
                 identifiers.add(name)
         else:
-            if name not in strings:
-                strings[name] = string
-            elif string != strings[name]:
+            if string not in strings:
+                strings[string] = name
+            elif name != strings[string]:
                 raise ValueError(f'string mismatch for {name!r} ({string!r} != {strings[name]!r}')
     return identifiers, strings