bpo-36763: Add _PyPreConfig._config_init (GH-13481)

* _PyPreConfig_GetGlobalConfig() and  _PyCoreConfig_GetGlobalConfig()
  now do nothing if the configuration was not initialized with
  _PyPreConfig_InitCompatConfig() and _PyCoreConfig_InitCompatConfig()
* Remove utf8_mode=-2 special case: use utf8_mode=-1 instead.
* Fix _PyPreConfig_InitPythonConfig():

  * isolated = 0 instead of -1
  * use_environment = 1 instead of -1

* Rename _PyConfig_INIT to  _PyConfig_INIT_COMPAT
* Rename _PyPreConfig_Init() to _PyPreConfig_InitCompatConfig()
* Rename _PyCoreConfig_Init() to _PyCoreConfig_InitCompatConfig()
* PyInterpreterState_New() now uses _PyCoreConfig_InitPythonConfig()
  as default configuration, but it's very quickly overriden anyway.
* _freeze_importlib.c uses _PyCoreConfig_SetString() to set
  program_name.
* Cleanup preconfig_init_utf8_mode(): cmdline is always non-NULL.

Include/cpython/coreconfig.h

@@ -40,9 +40,18 @@ typedef struct {
 
 #define _Py_CONFIG_VERSION 1
 
+typedef enum {
+    /* Py_Initialize() API: backward compatibility with Python 3.6 and 3.7 */
+    _PyConfig_INIT_COMPAT = 1,
+    _PyConfig_INIT_PYTHON = 2,
+    _PyConfig_INIT_ISOLATED = 3
+} _PyConfigInitEnum;
+
+
 typedef struct {
     int _config_version;  /* Internal configuration version,
                              used for ABI compatibility */
+    int _config_init;     /* _PyConfigInitEnum value */
 
     /* Parse _Py_PreInitializeFromArgs() arguments?
        See _PyCoreConfig.parse_argv */
@@ -107,10 +116,7 @@ typedef struct {
        Set to 0 by "-X utf8=0" and PYTHONUTF8=0.
 
        If equals to -1, it is set to 1 if the LC_CTYPE locale is "C" or
-       "POSIX", otherwise it is set to 0.
-
-       If equals to -2, inherit Py_UTF8Mode value value (which is equal to 0
-       by default). */
+       "POSIX", otherwise it is set to 0. Inherit Py_UTF8Mode value value. */
     int utf8_mode;
 
     int dev_mode;           /* Development mode. PYTHONDEVMODE, -X dev */
@@ -126,16 +132,10 @@ PyAPI_FUNC(void) _PyPreConfig_InitIsolatedConfig(_PyPreConfig *config);
 
 /* --- _PyCoreConfig ---------------------------------------------- */
 
-typedef enum {
-    _PyCoreConfig_INIT = 0,
-    _PyCoreConfig_INIT_PYTHON = 1,
-    _PyCoreConfig_INIT_ISOLATED = 2
-} _PyCoreConfigInitEnum;
-
 typedef struct {
     int _config_version;  /* Internal configuration version,
                              used for ABI compatibility */
-    int _config_init;     /* _PyCoreConfigInitEnum value */
+    int _config_init;     /* _PyConfigInitEnum value */
 
     int isolated;         /* Isolated mode? see _PyPreConfig.isolated */
     int use_environment;  /* Use environment variables? see _PyPreConfig.use_environment */

Include/internal/pycore_coreconfig.h

@@ -120,7 +120,7 @@ PyAPI_FUNC(_PyInitError) _PyPreCmdline_Read(_PyPreCmdline *cmdline,
 
 /* --- _PyPreConfig ----------------------------------------------- */
 
-PyAPI_FUNC(void) _PyPreConfig_Init(_PyPreConfig *config);
+PyAPI_FUNC(void) _PyPreConfig_InitCompatConfig(_PyPreConfig *config);
 PyAPI_FUNC(void) _PyPreConfig_InitFromCoreConfig(
     _PyPreConfig *config,
     const _PyCoreConfig *coreconfig);
@@ -139,7 +139,7 @@ PyAPI_FUNC(_PyInitError) _PyPreConfig_Write(const _PyPreConfig *config);
 
 /* --- _PyCoreConfig ---------------------------------------------- */
 
-PyAPI_FUNC(void) _PyCoreConfig_Init(_PyCoreConfig *config);
+PyAPI_FUNC(void) _PyCoreConfig_InitCompatConfig(_PyCoreConfig *config);
 PyAPI_FUNC(_PyInitError) _PyCoreConfig_Copy(
     _PyCoreConfig *config,
     const _PyCoreConfig *config2);

Programs/_freeze_importlib.c

@@ -76,18 +76,30 @@ main(int argc, char *argv[])
     }
     text[text_size] = '\0';
 
+    _PyInitError err;
     _PyCoreConfig config;
-    _PyCoreConfig_InitIsolatedConfig(&config);
+
+    err = _PyCoreConfig_InitIsolatedConfig(&config);
+    if (_PyInitError_Failed(err)) {
+        _PyCoreConfig_Clear(&config);
+        _Py_ExitInitError(err);
+    }
 
     config.site_import = 0;
-    config.program_name = L"./_freeze_importlib";
+
+    err = _PyCoreConfig_SetString(&config, &config.program_name,
+                                  L"./_freeze_importlib");
+    if (_PyInitError_Failed(err)) {
+        _PyCoreConfig_Clear(&config);
+        _Py_ExitInitError(err);
+    }
+
     /* Don't install importlib, since it could execute outdated bytecode. */
     config._install_importlib = 0;
     config._init_main = 0;
 
-    _PyInitError err = _Py_InitializeFromConfig(&config);
-    /* No need to call _PyCoreConfig_Clear() since we didn't allocate any
-       memory: program_name is a constant string. */
+    err = _Py_InitializeFromConfig(&config);
+    _PyCoreConfig_Clear(&config);
     if (_PyInitError_Failed(err)) {
         _Py_ExitInitError(err);
     }

Programs/_testembed.c

@@ -317,7 +317,7 @@ dump_config(void)
 }
 
 
-static int test_init_default_config(void)
+static int test_init_initialize_config(void)
 {
     _testembed_Py_Initialize();
     dump_config();
@@ -326,6 +326,47 @@ static int test_init_default_config(void)
 }
 
 
+static int check_init_compat_config(int preinit)
+{
+    _PyInitError err;
+
+    if (preinit) {
+        _PyPreConfig preconfig;
+        _PyPreConfig_InitCompatConfig(&preconfig);
+
+        err = _Py_PreInitialize(&preconfig);
+        if (_PyInitError_Failed(err)) {
+            _Py_ExitInitError(err);
+        }
+    }
+
+    _PyCoreConfig config;
+    _PyCoreConfig_InitCompatConfig(&config);
+    config.program_name = L"./_testembed";
+
+    err = _Py_InitializeFromConfig(&config);
+    if (_PyInitError_Failed(err)) {
+        _Py_ExitInitError(err);
+    }
+
+    dump_config();
+    Py_Finalize();
+    return 0;
+}
+
+
+static int test_preinit_compat_config(void)
+{
+    return check_init_compat_config(1);
+}
+
+
+static int test_init_compat_config(void)
+{
+    return check_init_compat_config(0);
+}
+
+
 static int test_init_global_config(void)
 {
     /* FIXME: test Py_IgnoreEnvironmentFlag */
@@ -380,7 +421,7 @@ static int test_init_from_config(void)
     _PyInitError err;
 
     _PyPreConfig preconfig;
-    _PyPreConfig_Init(&preconfig);
+    _PyPreConfig_InitCompatConfig(&preconfig);
 
     putenv("PYTHONMALLOC=malloc_debug");
     preconfig.allocator = PYMEM_ALLOCATOR_MALLOC;
@@ -396,7 +437,7 @@ static int test_init_from_config(void)
 
     /* Test _Py_InitializeFromConfig() */
     _PyCoreConfig config;
-    _PyCoreConfig_Init(&config);
+    _PyCoreConfig_InitCompatConfig(&config);
     config.install_signal_handlers = 0;
 
     /* FIXME: test use_environment */
@@ -676,7 +717,7 @@ static int test_preinit_isolated1(void)
     _PyInitError err;
 
     _PyPreConfig preconfig;
-    _PyPreConfig_Init(&preconfig);
+    _PyPreConfig_InitCompatConfig(&preconfig);
     preconfig.isolated = 1;
 
     err = _Py_PreInitialize(&preconfig);
@@ -685,7 +726,7 @@ static int test_preinit_isolated1(void)
     }
 
     _PyCoreConfig config;
-    _PyCoreConfig_Init(&config);
+    _PyCoreConfig_InitCompatConfig(&config);
     config.program_name = L"./_testembed";
 
     set_all_env_vars();
@@ -705,7 +746,7 @@ static int test_preinit_isolated2(void)
     _PyInitError err;
 
     _PyPreConfig preconfig;
-    _PyPreConfig_Init(&preconfig);
+    _PyPreConfig_InitCompatConfig(&preconfig);
     preconfig.isolated = 0;
 
     err = _Py_PreInitialize(&preconfig);
@@ -715,7 +756,7 @@ static int test_preinit_isolated2(void)
 
     /* Test _PyCoreConfig.isolated=1 */
     _PyCoreConfig config;
-    _PyCoreConfig_Init(&config);
+    _PyCoreConfig_InitCompatConfig(&config);
 
     Py_IsolatedFlag = 0;
     config.isolated = 1;
@@ -885,12 +926,14 @@ static int check_preinit_isolated_config(int preinit)
     _PyCoreConfig config;
     err = _PyCoreConfig_InitIsolatedConfig(&config);
     if (_PyInitError_Failed(err)) {
+        _PyCoreConfig_Clear(&config);
         _Py_ExitInitError(err);
     }
     config.program_name = L"./_testembed";
 
     err = _Py_InitializeFromConfig(&config);
     if (_PyInitError_Failed(err)) {
+        _PyCoreConfig_Clear(&config);
         _Py_ExitInitError(err);
     }
 
@@ -1207,7 +1250,9 @@ static struct TestCase TestCases[] = {
     { "bpo20891", test_bpo20891 },
     { "initialize_twice", test_initialize_twice },
     { "initialize_pymain", test_initialize_pymain },
-    { "init_default_config", test_init_default_config },
+    { "init_initialize_config", test_init_initialize_config },
+    { "preinit_compat_config", test_preinit_compat_config },
+    { "init_compat_config", test_init_compat_config },
     { "init_global_config", test_init_global_config },
     { "init_from_config", test_init_from_config },
     { "init_parse_argv", test_init_parse_argv },

Python/coreconfig.c

@@ -109,7 +109,7 @@ static const char usage_6[] =
 /* UTF-8 mode (PEP 540): if equals to 1, use the UTF-8 encoding, and change
    stdin and stdout error handler to "surrogateescape". It is equal to
    -1 by default: unknown, will be set by Py_Main() */
-int Py_UTF8Mode = 0;
+int Py_UTF8Mode = -1;
 int Py_DebugFlag = 0; /* Needed by parser.c */
 int Py_VerboseFlag = 0; /* Needed by import.c */
 int Py_QuietFlag = 0; /* Needed by sysmodule.c */
@@ -546,12 +546,12 @@ _PyCoreConfig_Clear(_PyCoreConfig *config)
 
 
 void
-_PyCoreConfig_Init(_PyCoreConfig *config)
+_PyCoreConfig_InitCompatConfig(_PyCoreConfig *config)
 {
     memset(config, 0, sizeof(*config));
 
     config->_config_version = _Py_CONFIG_VERSION;
-    config->_config_init = (int)_PyCoreConfig_INIT;
+    config->_config_init = (int)_PyConfig_INIT_COMPAT;
     config->isolated = -1;
     config->use_environment = -1;
     config->dev_mode = -1;
@@ -586,7 +586,7 @@ _PyCoreConfig_Init(_PyCoreConfig *config)
 static void
 _PyCoreConfig_InitDefaults(_PyCoreConfig *config)
 {
-    _PyCoreConfig_Init(config);
+    _PyCoreConfig_InitCompatConfig(config);
 
     config->isolated = 0;
     config->use_environment = 1;
@@ -613,7 +613,7 @@ _PyCoreConfig_InitPythonConfig(_PyCoreConfig *config)
 {
     _PyCoreConfig_InitDefaults(config);
 
-    config->_config_init = (int)_PyCoreConfig_INIT_PYTHON;
+    config->_config_init = (int)_PyConfig_INIT_PYTHON;
     config->configure_c_stdio = 1;
     config->parse_argv = 1;
 
@@ -626,7 +626,7 @@ _PyCoreConfig_InitIsolatedConfig(_PyCoreConfig *config)
 {
     _PyCoreConfig_InitDefaults(config);
 
-    config->_config_init = (int)_PyCoreConfig_INIT_ISOLATED;
+    config->_config_init = (int)_PyConfig_INIT_ISOLATED;
     config->isolated = 1;
     config->use_environment = 0;
     config->user_site_directory = 0;
@@ -962,6 +962,11 @@ _PyCoreConfig_GetEnvDup(_PyCoreConfig *config,
 static void
 _PyCoreConfig_GetGlobalConfig(_PyCoreConfig *config)
 {
+    if (config->_config_init != _PyConfig_INIT_COMPAT) {
+        /* Python and Isolated configuration ignore global variables */
+        return;
+    }
+
 #define COPY_FLAG(ATTR, VALUE) \
         if (config->ATTR == -1) { \
             config->ATTR = VALUE; \

Python/frozenmain.c

@@ -40,7 +40,11 @@ Py_FrozenMain(int argc, char **argv)
     }
 
     _PyCoreConfig config;
-    _PyCoreConfig_InitPythonConfig(&config);
+    err = _PyCoreConfig_InitPythonConfig(&config);
+    if (_PyInitError_Failed(err)) {
+        _PyCoreConfig_Clear(&config);
+        _Py_ExitInitError(err);
+    }
     config.pathconfig_warnings = 0;   /* Suppress errors from getpath.c */
 
     if ((p = Py_GETENV("PYTHONINSPECT")) && *p != '\0')
@@ -82,8 +86,7 @@ Py_FrozenMain(int argc, char **argv)
         Py_SetProgramName(argv_copy[0]);
 
     err = _Py_InitializeFromConfig(&config);
-    /* No need to call _PyCoreConfig_Clear() since we didn't allocate any
-       memory: program_name is a constant string. */
+    _PyCoreConfig_Clear(&config);
     if (_PyInitError_Failed(err)) {
         _Py_ExitInitError(err);
     }

Python/pathconfig.c

@@ -383,7 +383,7 @@ pathconfig_global_init(void)
 
     _PyInitError err;
     _PyCoreConfig config;
-    _PyCoreConfig_Init(&config);
+    _PyCoreConfig_InitCompatConfig(&config);
 
     err = _PyCoreConfig_Read(&config);
     if (_Py_INIT_FAILED(err)) {

Python/preconfig.c

@@ -262,16 +262,17 @@ _PyPreCmdline_Read(_PyPreCmdline *cmdline, const _PyPreConfig *preconfig)
 
 
 void
-_PyPreConfig_Init(_PyPreConfig *config)
+_PyPreConfig_InitCompatConfig(_PyPreConfig *config)
 {
     memset(config, 0, sizeof(*config));
 
     config->_config_version = _Py_CONFIG_VERSION;
+    config->_config_init = (int)_PyConfig_INIT_COMPAT;
     config->parse_argv = 0;
     config->isolated = -1;
     config->use_environment = -1;
     config->configure_locale = 1;
-    config->utf8_mode = -2;
+    config->utf8_mode = -1;
     config->dev_mode = -1;
     config->allocator = PYMEM_ALLOCATOR_NOT_SET;
 #ifdef MS_WINDOWS
@@ -283,23 +284,30 @@ _PyPreConfig_Init(_PyPreConfig *config)
 void
 _PyPreConfig_InitPythonConfig(_PyPreConfig *config)
 {
-    _PyPreConfig_Init(config);
+    _PyPreConfig_InitCompatConfig(config);
 
+    config->_config_init = (int)_PyConfig_INIT_PYTHON;
+    config->isolated = 0;
     config->parse_argv = 1;
+    config->use_environment = 1;
     /* Set to -1 to enable C locale coercion (PEP 538) and UTF-8 Mode (PEP 540)
        depending on the LC_CTYPE locale, PYTHONUTF8 and PYTHONCOERCECLOCALE
        environment variables. */
     config->coerce_c_locale = -1;
     config->coerce_c_locale_warn = -1;
     config->utf8_mode = -1;
+#ifdef MS_WINDOWS
+    config->legacy_windows_fs_encoding = 0;
+#endif
 }
 
 
 void
 _PyPreConfig_InitIsolatedConfig(_PyPreConfig *config)
 {
-    _PyPreConfig_Init(config);
+    _PyPreConfig_InitCompatConfig(config);
 
+    config->_config_init = (int)_PyConfig_INIT_ISOLATED;
     config->configure_locale = 0;
     config->isolated = 1;
     config->use_environment = 0;
@@ -315,7 +323,7 @@ void
 _PyPreConfig_InitFromPreConfig(_PyPreConfig *config,
                                const _PyPreConfig *config2)
 {
-    _PyPreConfig_Init(config);
+    _PyPreConfig_InitCompatConfig(config);
     _PyPreConfig_Copy(config, config2);
 }
 
@@ -324,17 +332,17 @@ void
 _PyPreConfig_InitFromCoreConfig(_PyPreConfig *config,
                                 const _PyCoreConfig *coreconfig)
 {
-    _PyCoreConfigInitEnum config_init = (_PyCoreConfigInitEnum)coreconfig->_config_init;
+    _PyConfigInitEnum config_init = (_PyConfigInitEnum)coreconfig->_config_init;
     switch (config_init) {
-    case _PyCoreConfig_INIT_PYTHON:
+    case _PyConfig_INIT_PYTHON:
         _PyPreConfig_InitPythonConfig(config);
         break;
-    case _PyCoreConfig_INIT_ISOLATED:
+    case _PyConfig_INIT_ISOLATED:
         _PyPreConfig_InitIsolatedConfig(config);
         break;
-    case _PyCoreConfig_INIT:
+    case _PyConfig_INIT_COMPAT:
     default:
-        _PyPreConfig_Init(config);
+        _PyPreConfig_InitCompatConfig(config);
     }
     _PyPreConfig_GetCoreConfig(config, coreconfig);
 }
@@ -428,6 +436,11 @@ _PyPreConfig_GetCoreConfig(_PyPreConfig *config,
 static void
 _PyPreConfig_GetGlobalConfig(_PyPreConfig *config)
 {
+    if (config->_config_init != _PyConfig_INIT_COMPAT) {
+        /* Python and Isolated configuration ignore global variables */
+        return;
+    }
+
 #define COPY_FLAG(ATTR, VALUE) \
     if (config->ATTR < 0) { \
         config->ATTR = VALUE; \
@@ -439,12 +452,10 @@ _PyPreConfig_GetGlobalConfig(_PyPreConfig *config)
 
     COPY_FLAG(isolated, Py_IsolatedFlag);
     COPY_NOT_FLAG(use_environment, Py_IgnoreEnvironmentFlag);
+    COPY_FLAG(utf8_mode, Py_UTF8Mode);
 #ifdef MS_WINDOWS
     COPY_FLAG(legacy_windows_fs_encoding, Py_LegacyWindowsFSEncodingFlag);
 #endif
-    if (config->utf8_mode == -2) {
-        config->utf8_mode = Py_UTF8Mode;
-    }
 
 #undef COPY_FLAG
 #undef COPY_NOT_FLAG
@@ -565,12 +576,7 @@ preconfig_init_utf8_mode(_PyPreConfig *config, const _PyPreCmdline *cmdline)
     }
 
     const wchar_t *xopt;
-    if (cmdline) {
-        xopt = _Py_get_xoption(&cmdline->xoptions, L"utf8");
-    }
-    else {
-        xopt = NULL;
-    }
+    xopt = _Py_get_xoption(&cmdline->xoptions, L"utf8");
     if (xopt) {
         wchar_t *sep = wcschr(xopt, L'=');
         if (sep) {

Python/pylifecycle.c

@@ -867,7 +867,7 @@ _Py_InitializeCore(_PyRuntimeState *runtime,
     }
 
     _PyCoreConfig local_config;
-    _PyCoreConfig_Init(&local_config);
+    _PyCoreConfig_InitCompatConfig(&local_config);
     err = pyinit_coreconfig(runtime, &local_config, src_config, args, interp_p);
     _PyCoreConfig_Clear(&local_config);
     return err;
@@ -1096,7 +1096,7 @@ Py_InitializeEx(int install_sigs)
     }
 
     _PyCoreConfig config;
-    _PyCoreConfig_Init(&config);
+    _PyCoreConfig_InitCompatConfig(&config);
     config.install_signal_handlers = install_sigs;
 
     err = _Py_InitializeFromConfig(&config);

Python/pystate.c

@@ -49,7 +49,7 @@ _PyRuntimeState_Init_impl(_PyRuntimeState *runtime)
 
     _PyGC_Initialize(&runtime->gc);
     _PyEval_Initialize(&runtime->ceval);
-    _PyPreConfig_Init(&runtime->preconfig);
+    _PyPreConfig_InitPythonConfig(&runtime->preconfig);
 
     runtime->gilstate.check_enabled = 1;
 
@@ -189,7 +189,13 @@ PyInterpreterState_New(void)
     memset(interp, 0, sizeof(*interp));
     interp->id_refcount = -1;
     interp->check_interval = 100;
-    _PyCoreConfig_Init(&interp->core_config);
+
+    _PyInitError err = _PyCoreConfig_InitPythonConfig(&interp->core_config);
+    if (_Py_INIT_FAILED(err)) {
+        PyMem_RawFree(interp);
+        return NULL;
+    }
+
     interp->eval_frame = _PyEval_EvalFrameDefault;
 #ifdef HAVE_DLOPEN
 #if HAVE_DECL_RTLD_NOW