/* Python interpreter top-level routines, including init/exit */ #include "Python.h" #include "Python-ast.h" #undef Yield /* undefine macro conflicting with winbase.h */ #include "grammar.h" #include "node.h" #include "token.h" #include "parsetok.h" #include "errcode.h" #include "code.h" #include "compile.h" #include "symtable.h" #include "pyarena.h" #include "ast.h" #include "eval.h" #include "marshal.h" #include "osdefs.h" #ifdef HAVE_SIGNAL_H #include #endif #ifdef MS_WINDOWS #include "malloc.h" /* for alloca */ #endif #ifdef HAVE_LANGINFO_H #include #include #endif #ifdef MS_WINDOWS #undef BYTE #include "windows.h" #define PATH_MAX MAXPATHLEN #endif #ifndef Py_REF_DEBUG #define PRINT_TOTAL_REFS() #else /* Py_REF_DEBUG */ #define PRINT_TOTAL_REFS() fprintf(stderr, \ "[%" PY_FORMAT_SIZE_T "d refs]\n", \ _Py_GetRefTotal()) #endif #ifdef __cplusplus extern "C" { #endif extern wchar_t *Py_GetPath(void); extern grammar _PyParser_Grammar; /* From graminit.c */ /* Forward */ static void initmain(void); static void initsite(void); static int initstdio(void); static void flush_io(void); static PyObject *run_mod(mod_ty, const char *, PyObject *, PyObject *, PyCompilerFlags *, PyArena *); static PyObject *run_pyc_file(FILE *, const char *, PyObject *, PyObject *, PyCompilerFlags *); static void err_input(perrdetail *); static void initsigs(void); static void call_py_exitfuncs(void); static void call_ll_exitfuncs(void); extern void _PyUnicode_Init(void); extern void _PyUnicode_Fini(void); extern int _PyLong_Init(void); extern void PyLong_Fini(void); #ifdef WITH_THREAD extern void _PyGILState_Init(PyInterpreterState *, PyThreadState *); extern void _PyGILState_Fini(void); #endif /* WITH_THREAD */ int Py_DebugFlag; /* Needed by parser.c */ int Py_VerboseFlag; /* Needed by import.c */ int Py_InteractiveFlag; /* Needed by Py_FdIsInteractive() below */ int Py_InspectFlag; /* Needed to determine whether to exit at SystemError */ int Py_NoSiteFlag; /* Suppress 'import site' */ int Py_BytesWarningFlag; /* Warn on str(bytes) and str(buffer) */ int Py_DontWriteBytecodeFlag; /* Suppress writing bytecode files (*.py[co]) */ int Py_UseClassExceptionsFlag = 1; /* Needed by bltinmodule.c: deprecated */ int Py_FrozenFlag; /* Needed by getpath.c */ int Py_IgnoreEnvironmentFlag; /* e.g. PYTHONPATH, PYTHONHOME */ int Py_NoUserSiteDirectory = 0; /* for -s and site.py */ int Py_UnbufferedStdioFlag = 0; /* Unbuffered binary std{in,out,err} */ /* PyModule_GetWarningsModule is no longer necessary as of 2.6 since _warnings is builtin. This API should not be used. */ PyObject * PyModule_GetWarningsModule(void) { return PyImport_ImportModule("warnings"); } static int initialized = 0; /* API to access the initialized flag -- useful for esoteric use */ int Py_IsInitialized(void) { return initialized; } /* Global initializations. Can be undone by Py_Finalize(). Don't call this twice without an intervening Py_Finalize() call. When initializations fail, a fatal error is issued and the function does not return. On return, the first thread and interpreter state have been created. Locking: you must hold the interpreter lock while calling this. (If the lock has not yet been initialized, that's equivalent to having the lock, but you cannot use multiple threads.) */ static int add_flag(int flag, const char *envs) { int env = atoi(envs); if (flag < env) flag = env; if (flag < 1) flag = 1; return flag; } #if defined(HAVE_LANGINFO_H) && defined(CODESET) static char* get_codeset(void) { char* codeset; PyObject *codec, *name; codeset = nl_langinfo(CODESET); if (!codeset || codeset[0] == '\0') return NULL; codec = _PyCodec_Lookup(codeset); if (!codec) goto error; name = PyObject_GetAttrString(codec, "name"); Py_CLEAR(codec); if (!name) goto error; codeset = strdup(_PyUnicode_AsString(name)); Py_DECREF(name); return codeset; error: Py_XDECREF(codec); PyErr_Clear(); return NULL; } #endif void Py_InitializeEx(int install_sigs) { PyInterpreterState *interp; PyThreadState *tstate; PyObject *bimod, *sysmod, *pstderr; char *p; #if defined(HAVE_LANGINFO_H) && defined(CODESET) char *codeset; #endif extern void _Py_ReadyTypes(void); if (initialized) return; initialized = 1; #ifdef HAVE_SETLOCALE /* Set up the LC_CTYPE locale, so we can obtain the locale's charset without having to switch locales. */ setlocale(LC_CTYPE, ""); #endif if ((p = Py_GETENV("PYTHONDEBUG")) && *p != '\0') Py_DebugFlag = add_flag(Py_DebugFlag, p); if ((p = Py_GETENV("PYTHONVERBOSE")) && *p != '\0') Py_VerboseFlag = add_flag(Py_VerboseFlag, p); if ((p = Py_GETENV("PYTHONOPTIMIZE")) && *p != '\0') Py_OptimizeFlag = add_flag(Py_OptimizeFlag, p); if ((p = Py_GETENV("PYTHONDONTWRITEBYTECODE")) && *p != '\0') Py_DontWriteBytecodeFlag = add_flag(Py_DontWriteBytecodeFlag, p); interp = PyInterpreterState_New(); if (interp == NULL) Py_FatalError("Py_Initialize: can't make first interpreter"); tstate = PyThreadState_New(interp); if (tstate == NULL) Py_FatalError("Py_Initialize: can't make first thread"); (void) PyThreadState_Swap(tstate); _Py_ReadyTypes(); if (!_PyFrame_Init()) Py_FatalError("Py_Initialize: can't init frames"); if (!_PyLong_Init()) Py_FatalError("Py_Initialize: can't init longs"); if (!PyByteArray_Init()) Py_FatalError("Py_Initialize: can't init bytes"); _PyFloat_Init(); interp->modules = PyDict_New(); if (interp->modules == NULL) Py_FatalError("Py_Initialize: can't make modules dictionary"); interp->modules_reloading = PyDict_New(); if (interp->modules_reloading == NULL) Py_FatalError("Py_Initialize: can't make modules_reloading dictionary"); /* Init Unicode implementation; relies on the codec registry */ _PyUnicode_Init(); bimod = _PyBuiltin_Init(); if (bimod == NULL) Py_FatalError("Py_Initialize: can't initialize builtins modules"); _PyImport_FixupExtension(bimod, "builtins", "builtins"); interp->builtins = PyModule_GetDict(bimod); if (interp->builtins == NULL) Py_FatalError("Py_Initialize: can't initialize builtins dict"); Py_INCREF(interp->builtins); /* initialize builtin exceptions */ _PyExc_Init(); sysmod = _PySys_Init(); if (sysmod == NULL) Py_FatalError("Py_Initialize: can't initialize sys"); interp->sysdict = PyModule_GetDict(sysmod); if (interp->sysdict == NULL) Py_FatalError("Py_Initialize: can't initialize sys dict"); Py_INCREF(interp->sysdict); _PyImport_FixupExtension(sysmod, "sys", "sys"); PySys_SetPath(Py_GetPath()); PyDict_SetItemString(interp->sysdict, "modules", interp->modules); /* Set up a preliminary stderr printer until we have enough infrastructure for the io module in place. */ pstderr = PyFile_NewStdPrinter(fileno(stderr)); if (pstderr == NULL) Py_FatalError("Py_Initialize: can't set preliminary stderr"); PySys_SetObject("stderr", pstderr); PySys_SetObject("__stderr__", pstderr); _PyImport_Init(); _PyImportHooks_Init(); if (install_sigs) initsigs(); /* Signal handling stuff, including initintr() */ /* Initialize warnings. */ _PyWarnings_Init(); if (PySys_HasWarnOptions()) { PyObject *warnings_module = PyImport_ImportModule("warnings"); if (!warnings_module) PyErr_Clear(); Py_XDECREF(warnings_module); } initmain(); /* Module __main__ */ if (initstdio() < 0) Py_FatalError( "Py_Initialize: can't initialize sys standard streams"); if (!Py_NoSiteFlag) initsite(); /* Module site */ /* auto-thread-state API, if available */ #ifdef WITH_THREAD _PyGILState_Init(interp, tstate); #endif /* WITH_THREAD */ #if defined(HAVE_LANGINFO_H) && defined(CODESET) /* On Unix, set the file system encoding according to the user's preference, if the CODESET names a well-known Python codec, and Py_FileSystemDefaultEncoding isn't initialized by other means. Also set the encoding of stdin and stdout if these are terminals. */ codeset = get_codeset(); if (codeset) { if (!Py_FileSystemDefaultEncoding) Py_FileSystemDefaultEncoding = codeset; else free(codeset); } #endif } void Py_Initialize(void) { Py_InitializeEx(1); } #ifdef COUNT_ALLOCS extern void dump_counts(FILE*); #endif /* Flush stdout and stderr */ static void flush_std_files(void) { PyObject *fout = PySys_GetObject("stdout"); PyObject *ferr = PySys_GetObject("stderr"); PyObject *tmp; if (fout != NULL && fout != Py_None) { tmp = PyObject_CallMethod(fout, "flush", ""); if (tmp == NULL) PyErr_Clear(); else Py_DECREF(tmp); } if (ferr != NULL || ferr != Py_None) { tmp = PyObject_CallMethod(ferr, "flush", ""); if (tmp == NULL) PyErr_Clear(); else Py_DECREF(tmp); } } /* Undo the effect of Py_Initialize(). Beware: if multiple interpreter and/or thread states exist, these are not wiped out; only the current thread and interpreter state are deleted. But since everything else is deleted, those other interpreter and thread states should no longer be used. (XXX We should do better, e.g. wipe out all interpreters and threads.) Locking: as above. */ void Py_Finalize(void) { PyInterpreterState *interp; PyThreadState *tstate; if (!initialized) return; /* The interpreter is still entirely intact at this point, and the * exit funcs may be relying on that. In particular, if some thread * or exit func is still waiting to do an import, the import machinery * expects Py_IsInitialized() to return true. So don't say the * interpreter is uninitialized until after the exit funcs have run. * Note that Threading.py uses an exit func to do a join on all the * threads created thru it, so this also protects pending imports in * the threads created via Threading. */ call_py_exitfuncs(); initialized = 0; /* Flush stdout+stderr */ flush_std_files(); /* Get current thread state and interpreter pointer */ tstate = PyThreadState_GET(); interp = tstate->interp; /* Disable signal handling */ PyOS_FiniInterrupts(); /* Clear type lookup cache */ PyType_ClearCache(); /* Collect garbage. This may call finalizers; it's nice to call these * before all modules are destroyed. * XXX If a __del__ or weakref callback is triggered here, and tries to * XXX import a module, bad things can happen, because Python no * XXX longer believes it's initialized. * XXX Fatal Python error: Interpreter not initialized (version mismatch?) * XXX is easy to provoke that way. I've also seen, e.g., * XXX Exception exceptions.ImportError: 'No module named sha' * XXX in ignored * XXX but I'm unclear on exactly how that one happens. In any case, * XXX I haven't seen a real-life report of either of these. */ PyGC_Collect(); #ifdef COUNT_ALLOCS /* With COUNT_ALLOCS, it helps to run GC multiple times: each collection might release some types from the type list, so they become garbage. */ while (PyGC_Collect() > 0) /* nothing */; #endif /* Destroy all modules */ PyImport_Cleanup(); /* Flush stdout+stderr (again, in case more was printed) */ flush_std_files(); /* Collect final garbage. This disposes of cycles created by * new-style class definitions, for example. * XXX This is disabled because it caused too many problems. If * XXX a __del__ or weakref callback triggers here, Python code has * XXX a hard time running, because even the sys module has been * XXX cleared out (sys.stdout is gone, sys.excepthook is gone, etc). * XXX One symptom is a sequence of information-free messages * XXX coming from threads (if a __del__ or callback is invoked, * XXX other threads can execute too, and any exception they encounter * XXX triggers a comedy of errors as subsystem after subsystem * XXX fails to find what it *expects* to find in sys to help report * XXX the exception and consequent unexpected failures). I've also * XXX seen segfaults then, after adding print statements to the * XXX Python code getting called. */ #if 0 PyGC_Collect(); #endif /* Destroy the database used by _PyImport_{Fixup,Find}Extension */ _PyImport_Fini(); /* Debugging stuff */ #ifdef COUNT_ALLOCS dump_counts(stdout); #endif PRINT_TOTAL_REFS(); #ifdef Py_TRACE_REFS /* Display all objects still alive -- this can invoke arbitrary * __repr__ overrides, so requires a mostly-intact interpreter. * Alas, a lot of stuff may still be alive now that will be cleaned * up later. */ if (Py_GETENV("PYTHONDUMPREFS")) _Py_PrintReferences(stderr); #endif /* Py_TRACE_REFS */ /* Clear interpreter state */ PyInterpreterState_Clear(interp); /* Now we decref the exception classes. After this point nothing can raise an exception. That's okay, because each Fini() method below has been checked to make sure no exceptions are ever raised. */ _PyExc_Fini(); /* Cleanup auto-thread-state */ #ifdef WITH_THREAD _PyGILState_Fini(); #endif /* WITH_THREAD */ /* Delete current thread */ PyThreadState_Swap(NULL); PyInterpreterState_Delete(interp); /* Sundry finalizers */ PyMethod_Fini(); PyFrame_Fini(); PyCFunction_Fini(); PyTuple_Fini(); PyList_Fini(); PySet_Fini(); PyBytes_Fini(); PyByteArray_Fini(); PyLong_Fini(); PyFloat_Fini(); PyDict_Fini(); /* Cleanup Unicode implementation */ _PyUnicode_Fini(); /* reset file system default encoding */ if (!Py_HasFileSystemDefaultEncoding) { free((char*)Py_FileSystemDefaultEncoding); Py_FileSystemDefaultEncoding = NULL; } /* XXX Still allocated: - various static ad-hoc pointers to interned strings - int and float free list blocks - whatever various modules and libraries allocate */ PyGrammar_RemoveAccelerators(&_PyParser_Grammar); #ifdef Py_TRACE_REFS /* Display addresses (& refcnts) of all objects still alive. * An address can be used to find the repr of the object, printed * above by _Py_PrintReferences. */ if (Py_GETENV("PYTHONDUMPREFS")) _Py_PrintReferenceAddresses(stderr); #endif /* Py_TRACE_REFS */ #ifdef PYMALLOC_DEBUG if (Py_GETENV("PYTHONMALLOCSTATS")) _PyObject_DebugMallocStats(); #endif call_ll_exitfuncs(); } /* Create and initialize a new interpreter and thread, and return the new thread. This requires that Py_Initialize() has been called first. Unsuccessful initialization yields a NULL pointer. Note that *no* exception information is available even in this case -- the exception information is held in the thread, and there is no thread. Locking: as above. */ PyThreadState * Py_NewInterpreter(void) { PyInterpreterState *interp; PyThreadState *tstate, *save_tstate; PyObject *bimod, *sysmod; if (!initialized) Py_FatalError("Py_NewInterpreter: call Py_Initialize first"); interp = PyInterpreterState_New(); if (interp == NULL) return NULL; tstate = PyThreadState_New(interp); if (tstate == NULL) { PyInterpreterState_Delete(interp); return NULL; } save_tstate = PyThreadState_Swap(tstate); /* XXX The following is lax in error checking */ interp->modules = PyDict_New(); interp->modules_reloading = PyDict_New(); bimod = _PyImport_FindExtension("builtins", "builtins"); if (bimod != NULL) { interp->builtins = PyModule_GetDict(bimod); if (interp->builtins == NULL) goto handle_error; Py_INCREF(interp->builtins); } /* initialize builtin exceptions */ _PyExc_Init(); sysmod = _PyImport_FindExtension("sys", "sys"); if (bimod != NULL && sysmod != NULL) { PyObject *pstderr; interp->sysdict = PyModule_GetDict(sysmod); if (interp->sysdict == NULL) goto handle_error; Py_INCREF(interp->sysdict); PySys_SetPath(Py_GetPath()); PyDict_SetItemString(interp->sysdict, "modules", interp->modules); /* Set up a preliminary stderr printer until we have enough infrastructure for the io module in place. */ pstderr = PyFile_NewStdPrinter(fileno(stderr)); if (pstderr == NULL) Py_FatalError("Py_Initialize: can't set preliminary stderr"); PySys_SetObject("stderr", pstderr); PySys_SetObject("__stderr__", pstderr); _PyImportHooks_Init(); if (initstdio() < 0) Py_FatalError( "Py_Initialize: can't initialize sys standard streams"); initmain(); if (!Py_NoSiteFlag) initsite(); } if (!PyErr_Occurred()) return tstate; handle_error: /* Oops, it didn't work. Undo it all. */ PyErr_Print(); PyThreadState_Clear(tstate); PyThreadState_Swap(save_tstate); PyThreadState_Delete(tstate); PyInterpreterState_Delete(interp); return NULL; } /* Delete an interpreter and its last thread. This requires that the given thread state is current, that the thread has no remaining frames, and that it is its interpreter's only remaining thread. It is a fatal error to violate these constraints. (Py_Finalize() doesn't have these constraints -- it zaps everything, regardless.) Locking: as above. */ void Py_EndInterpreter(PyThreadState *tstate) { PyInterpreterState *interp = tstate->interp; if (tstate != PyThreadState_GET()) Py_FatalError("Py_EndInterpreter: thread is not current"); if (tstate->frame != NULL) Py_FatalError("Py_EndInterpreter: thread still has a frame"); if (tstate != interp->tstate_head || tstate->next != NULL) Py_FatalError("Py_EndInterpreter: not the last thread"); PyImport_Cleanup(); PyInterpreterState_Clear(interp); PyThreadState_Swap(NULL); PyInterpreterState_Delete(interp); } static wchar_t *progname = L"python"; void Py_SetProgramName(wchar_t *pn) { if (pn && *pn) progname = pn; } wchar_t * Py_GetProgramName(void) { return progname; } static wchar_t *default_home = NULL; static wchar_t env_home[PATH_MAX+1]; void Py_SetPythonHome(wchar_t *home) { default_home = home; } wchar_t * Py_GetPythonHome(void) { wchar_t *home = default_home; if (home == NULL && !Py_IgnoreEnvironmentFlag) { char* chome = Py_GETENV("PYTHONHOME"); if (chome) { size_t r = mbstowcs(env_home, chome, PATH_MAX+1); if (r != (size_t)-1 && r <= PATH_MAX) home = env_home; } } return home; } /* Create __main__ module */ static void initmain(void) { PyObject *m, *d; m = PyImport_AddModule("__main__"); if (m == NULL) Py_FatalError("can't create __main__ module"); d = PyModule_GetDict(m); if (PyDict_GetItemString(d, "__builtins__") == NULL) { PyObject *bimod = PyImport_ImportModule("builtins"); if (bimod == NULL || PyDict_SetItemString(d, "__builtins__", bimod) != 0) Py_FatalError("can't add __builtins__ to __main__"); Py_DECREF(bimod); } } /* Import the site module (not into __main__ though) */ static void initsite(void) { PyObject *m, *f; m = PyImport_ImportModule("site"); if (m == NULL) { f = PySys_GetObject("stderr"); if (f == NULL || f == Py_None) return; if (Py_VerboseFlag) { PyFile_WriteString( "'import site' failed; traceback:\n", f); PyErr_Print(); } else { PyFile_WriteString( "'import site' failed; use -v for traceback\n", f); PyErr_Clear(); } } else { Py_DECREF(m); } } static PyObject* create_stdio(PyObject* io, int fd, int write_mode, char* name, char* encoding, char* errors) { PyObject *buf = NULL, *stream = NULL, *text = NULL, *raw = NULL, *res; const char* mode; PyObject *line_buffering; int buffering, isatty; /* stdin is always opened in buffered mode, first because it shouldn't make a difference in common use cases, second because TextIOWrapper depends on the presence of a read1() method which only exists on buffered streams. */ if (Py_UnbufferedStdioFlag && write_mode) buffering = 0; else buffering = -1; if (write_mode) mode = "wb"; else mode = "rb"; buf = PyObject_CallMethod(io, "open", "isiOOOi", fd, mode, buffering, Py_None, Py_None, Py_None, 0); if (buf == NULL) goto error; if (buffering) { raw = PyObject_GetAttrString(buf, "raw"); if (raw == NULL) goto error; } else { raw = buf; Py_INCREF(raw); } text = PyUnicode_FromString(name); if (text == NULL || PyObject_SetAttrString(raw, "_name", text) < 0) goto error; res = PyObject_CallMethod(raw, "isatty", ""); if (res == NULL) goto error; isatty = PyObject_IsTrue(res); Py_DECREF(res); if (isatty == -1) goto error; if (isatty || Py_UnbufferedStdioFlag) line_buffering = Py_True; else line_buffering = Py_False; Py_CLEAR(raw); Py_CLEAR(text); stream = PyObject_CallMethod(io, "TextIOWrapper", "OsssO", buf, encoding, errors, "\n", line_buffering); Py_CLEAR(buf); if (stream == NULL) goto error; if (write_mode) mode = "w"; else mode = "r"; text = PyUnicode_FromString(mode); if (!text || PyObject_SetAttrString(stream, "mode", text) < 0) goto error; Py_CLEAR(text); return stream; error: Py_XDECREF(buf); Py_XDECREF(stream); Py_XDECREF(text); Py_XDECREF(raw); return NULL; } /* Initialize sys.stdin, stdout, stderr and builtins.open */ static int initstdio(void) { PyObject *iomod = NULL, *wrapper; PyObject *bimod = NULL; PyObject *m; PyObject *std = NULL; int status = 0, fd; PyObject * encoding_attr; char *encoding = NULL, *errors; /* Hack to avoid a nasty recursion issue when Python is invoked in verbose mode: pre-import the Latin-1 and UTF-8 codecs */ if ((m = PyImport_ImportModule("encodings.utf_8")) == NULL) { goto error; } Py_DECREF(m); if (!(m = PyImport_ImportModule("encodings.latin_1"))) { goto error; } Py_DECREF(m); if (!(bimod = PyImport_ImportModule("builtins"))) { goto error; } if (!(iomod = PyImport_ImportModule("io"))) { goto error; } if (!(wrapper = PyObject_GetAttrString(iomod, "OpenWrapper"))) { goto error; } /* Set builtins.open */ if (PyObject_SetAttrString(bimod, "open", wrapper) == -1) { goto error; } encoding = Py_GETENV("PYTHONIOENCODING"); errors = NULL; if (encoding) { encoding = strdup(encoding); errors = strchr(encoding, ':'); if (errors) { *errors = '\0'; errors++; } } /* Set sys.stdin */ fd = fileno(stdin); /* Under some conditions stdin, stdout and stderr may not be connected * and fileno() may point to an invalid file descriptor. For example * GUI apps don't have valid standard streams by default. */ if (fd < 0) { #ifdef MS_WINDOWS std = Py_None; Py_INCREF(std); #else goto error; #endif } else { std = create_stdio(iomod, fd, 0, "", encoding, errors); if (std == NULL) goto error; } /* if (fd < 0) */ PySys_SetObject("__stdin__", std); PySys_SetObject("stdin", std); Py_DECREF(std); /* Set sys.stdout */ fd = fileno(stdout); if (fd < 0) { #ifdef MS_WINDOWS std = Py_None; Py_INCREF(std); #else goto error; #endif } else { std = create_stdio(iomod, fd, 1, "", encoding, errors); if (std == NULL) goto error; } /* if (fd < 0) */ PySys_SetObject("__stdout__", std); PySys_SetObject("stdout", std); Py_DECREF(std); #if 1 /* Disable this if you have trouble debugging bootstrap stuff */ /* Set sys.stderr, replaces the preliminary stderr */ fd = fileno(stderr); if (fd < 0) { #ifdef MS_WINDOWS std = Py_None; Py_INCREF(std); #else goto error; #endif } else { std = create_stdio(iomod, fd, 1, "", encoding, "backslashreplace"); if (std == NULL) goto error; } /* if (fd < 0) */ /* Same as hack above, pre-import stderr's codec to avoid recursion when import.c tries to write to stderr in verbose mode. */ encoding_attr = PyObject_GetAttrString(std, "encoding"); if (encoding_attr != NULL) { const char * encoding; encoding = _PyUnicode_AsString(encoding_attr); if (encoding != NULL) { _PyCodec_Lookup(encoding); } } PyErr_Clear(); /* Not a fatal error if codec isn't available */ PySys_SetObject("__stderr__", std); PySys_SetObject("stderr", std); Py_DECREF(std); #endif if (0) { error: status = -1; } if (encoding) free(encoding); Py_XDECREF(bimod); Py_XDECREF(iomod); return status; } /* Parse input from a file and execute it */ int PyRun_AnyFileExFlags(FILE *fp, const char *filename, int closeit, PyCompilerFlags *flags) { if (filename == NULL) filename = "???"; if (Py_FdIsInteractive(fp, filename)) { int err = PyRun_InteractiveLoopFlags(fp, filename, flags); if (closeit) fclose(fp); return err; } else return PyRun_SimpleFileExFlags(fp, filename, closeit, flags); } int PyRun_InteractiveLoopFlags(FILE *fp, const char *filename, PyCompilerFlags *flags) { PyObject *v; int ret; PyCompilerFlags local_flags; if (flags == NULL) { flags = &local_flags; local_flags.cf_flags = 0; } v = PySys_GetObject("ps1"); if (v == NULL) { PySys_SetObject("ps1", v = PyUnicode_FromString(">>> ")); Py_XDECREF(v); } v = PySys_GetObject("ps2"); if (v == NULL) { PySys_SetObject("ps2", v = PyUnicode_FromString("... ")); Py_XDECREF(v); } for (;;) { ret = PyRun_InteractiveOneFlags(fp, filename, flags); PRINT_TOTAL_REFS(); if (ret == E_EOF) return 0; /* if (ret == E_NOMEM) return -1; */ } } /* compute parser flags based on compiler flags */ #define PARSER_FLAGS(flags) \ ((flags) ? ((((flags)->cf_flags & PyCF_DONT_IMPLY_DEDENT) ? \ PyPARSE_DONT_IMPLY_DEDENT : 0)) : 0) #if 0 /* Keep an example of flags with future keyword support. */ #define PARSER_FLAGS(flags) \ ((flags) ? ((((flags)->cf_flags & PyCF_DONT_IMPLY_DEDENT) ? \ PyPARSE_DONT_IMPLY_DEDENT : 0) \ | ((flags)->cf_flags & CO_FUTURE_WITH_STATEMENT ? \ PyPARSE_WITH_IS_KEYWORD : 0)) : 0) #endif int PyRun_InteractiveOneFlags(FILE *fp, const char *filename, PyCompilerFlags *flags) { PyObject *m, *d, *v, *w, *oenc = NULL; mod_ty mod; PyArena *arena; char *ps1 = "", *ps2 = "", *enc = NULL; int errcode = 0; if (fp == stdin) { /* Fetch encoding from sys.stdin */ v = PySys_GetObject("stdin"); if (v == NULL || v == Py_None) return -1; oenc = PyObject_GetAttrString(v, "encoding"); if (!oenc) return -1; enc = _PyUnicode_AsString(oenc); } v = PySys_GetObject("ps1"); if (v != NULL) { v = PyObject_Str(v); if (v == NULL) PyErr_Clear(); else if (PyUnicode_Check(v)) ps1 = _PyUnicode_AsString(v); } w = PySys_GetObject("ps2"); if (w != NULL) { w = PyObject_Str(w); if (w == NULL) PyErr_Clear(); else if (PyUnicode_Check(w)) ps2 = _PyUnicode_AsString(w); } arena = PyArena_New(); if (arena == NULL) { Py_XDECREF(v); Py_XDECREF(w); Py_XDECREF(oenc); return -1; } mod = PyParser_ASTFromFile(fp, filename, enc, Py_single_input, ps1, ps2, flags, &errcode, arena); Py_XDECREF(v); Py_XDECREF(w); Py_XDECREF(oenc); if (mod == NULL) { PyArena_Free(arena); if (errcode == E_EOF) { PyErr_Clear(); return E_EOF; } PyErr_Print(); return -1; } m = PyImport_AddModule("__main__"); if (m == NULL) { PyArena_Free(arena); return -1; } d = PyModule_GetDict(m); v = run_mod(mod, filename, d, d, flags, arena); PyArena_Free(arena); flush_io(); if (v == NULL) { PyErr_Print(); return -1; } Py_DECREF(v); return 0; } /* Check whether a file maybe a pyc file: Look at the extension, the file type, and, if we may close it, at the first few bytes. */ static int maybe_pyc_file(FILE *fp, const char* filename, const char* ext, int closeit) { if (strcmp(ext, ".pyc") == 0 || strcmp(ext, ".pyo") == 0) return 1; /* Only look into the file if we are allowed to close it, since it then should also be seekable. */ if (closeit) { /* Read only two bytes of the magic. If the file was opened in text mode, the bytes 3 and 4 of the magic (\r\n) might not be read as they are on disk. */ unsigned int halfmagic = PyImport_GetMagicNumber() & 0xFFFF; unsigned char buf[2]; /* Mess: In case of -x, the stream is NOT at its start now, and ungetc() was used to push back the first newline, which makes the current stream position formally undefined, and a x-platform nightmare. Unfortunately, we have no direct way to know whether -x was specified. So we use a terrible hack: if the current stream position is not 0, we assume -x was specified, and give up. Bug 132850 on SourceForge spells out the hopelessness of trying anything else (fseek and ftell don't work predictably x-platform for text-mode files). */ int ispyc = 0; if (ftell(fp) == 0) { if (fread(buf, 1, 2, fp) == 2 && ((unsigned int)buf[1]<<8 | buf[0]) == halfmagic) ispyc = 1; rewind(fp); } return ispyc; } return 0; } int PyRun_SimpleFileExFlags(FILE *fp, const char *filename, int closeit, PyCompilerFlags *flags) { PyObject *m, *d, *v; const char *ext; int set_file_name = 0, ret; m = PyImport_AddModule("__main__"); if (m == NULL) return -1; d = PyModule_GetDict(m); if (PyDict_GetItemString(d, "__file__") == NULL) { PyObject *f; f = PyUnicode_DecodeFSDefault(filename); if (f == NULL) return -1; if (PyDict_SetItemString(d, "__file__", f) < 0) { Py_DECREF(f); return -1; } set_file_name = 1; Py_DECREF(f); } ext = filename + strlen(filename) - 4; if (maybe_pyc_file(fp, filename, ext, closeit)) { /* Try to run a pyc file. First, re-open in binary */ if (closeit) fclose(fp); if ((fp = fopen(filename, "rb")) == NULL) { fprintf(stderr, "python: Can't reopen .pyc file\n"); ret = -1; goto done; } /* Turn on optimization if a .pyo file is given */ if (strcmp(ext, ".pyo") == 0) Py_OptimizeFlag = 1; v = run_pyc_file(fp, filename, d, d, flags); } else { v = PyRun_FileExFlags(fp, filename, Py_file_input, d, d, closeit, flags); } flush_io(); if (v == NULL) { PyErr_Print(); ret = -1; goto done; } Py_DECREF(v); ret = 0; done: if (set_file_name && PyDict_DelItemString(d, "__file__")) PyErr_Clear(); return ret; } int PyRun_SimpleStringFlags(const char *command, PyCompilerFlags *flags) { PyObject *m, *d, *v; m = PyImport_AddModule("__main__"); if (m == NULL) return -1; d = PyModule_GetDict(m); v = PyRun_StringFlags(command, Py_file_input, d, d, flags); if (v == NULL) { PyErr_Print(); return -1; } Py_DECREF(v); return 0; } static int parse_syntax_error(PyObject *err, PyObject **message, const char **filename, int *lineno, int *offset, const char **text) { long hold; PyObject *v; /* old style errors */ if (PyTuple_Check(err)) return PyArg_ParseTuple(err, "O(ziiz)", message, filename, lineno, offset, text); /* new style errors. `err' is an instance */ if (! (v = PyObject_GetAttrString(err, "msg"))) goto finally; *message = v; if (!(v = PyObject_GetAttrString(err, "filename"))) goto finally; if (v == Py_None) *filename = NULL; else if (! (*filename = _PyUnicode_AsString(v))) goto finally; Py_DECREF(v); if (!(v = PyObject_GetAttrString(err, "lineno"))) goto finally; hold = PyLong_AsLong(v); Py_DECREF(v); v = NULL; if (hold < 0 && PyErr_Occurred()) goto finally; *lineno = (int)hold; if (!(v = PyObject_GetAttrString(err, "offset"))) goto finally; if (v == Py_None) { *offset = -1; Py_DECREF(v); v = NULL; } else { hold = PyLong_AsLong(v); Py_DECREF(v); v = NULL; if (hold < 0 && PyErr_Occurred()) goto finally; *offset = (int)hold; } if (!(v = PyObject_GetAttrString(err, "text"))) goto finally; if (v == Py_None) *text = NULL; else if (!PyUnicode_Check(v) || !(*text = _PyUnicode_AsString(v))) goto finally; Py_DECREF(v); return 1; finally: Py_XDECREF(v); return 0; } void PyErr_Print(void) { PyErr_PrintEx(1); } static void print_error_text(PyObject *f, int offset, const char *text) { char *nl; if (offset >= 0) { if (offset > 0 && offset == (int)strlen(text)) offset--; for (;;) { nl = strchr(text, '\n'); if (nl == NULL || nl-text >= offset) break; offset -= (int)(nl+1-text); text = nl+1; } while (*text == ' ' || *text == '\t') { text++; offset--; } } PyFile_WriteString(" ", f); PyFile_WriteString(text, f); if (*text == '\0' || text[strlen(text)-1] != '\n') PyFile_WriteString("\n", f); if (offset == -1) return; PyFile_WriteString(" ", f); offset--; while (offset > 0) { PyFile_WriteString(" ", f); offset--; } PyFile_WriteString("^\n", f); } static void handle_system_exit(void) { PyObject *exception, *value, *tb; int exitcode = 0; if (Py_InspectFlag) /* Don't exit if -i flag was given. This flag is set to 0 * when entering interactive mode for inspecting. */ return; PyErr_Fetch(&exception, &value, &tb); fflush(stdout); if (value == NULL || value == Py_None) goto done; if (PyExceptionInstance_Check(value)) { /* The error code should be in the `code' attribute. */ PyObject *code = PyObject_GetAttrString(value, "code"); if (code) { Py_DECREF(value); value = code; if (value == Py_None) goto done; } /* If we failed to dig out the 'code' attribute, just let the else clause below print the error. */ } if (PyLong_Check(value)) exitcode = (int)PyLong_AsLong(value); else { PyObject_Print(value, stderr, Py_PRINT_RAW); PySys_WriteStderr("\n"); exitcode = 1; } done: /* Restore and clear the exception info, in order to properly decref * the exception, value, and traceback. If we just exit instead, * these leak, which confuses PYTHONDUMPREFS output, and may prevent * some finalizers from running. */ PyErr_Restore(exception, value, tb); PyErr_Clear(); Py_Exit(exitcode); /* NOTREACHED */ } void PyErr_PrintEx(int set_sys_last_vars) { PyObject *exception, *v, *tb, *hook; if (PyErr_ExceptionMatches(PyExc_SystemExit)) { handle_system_exit(); } PyErr_Fetch(&exception, &v, &tb); if (exception == NULL) return; PyErr_NormalizeException(&exception, &v, &tb); if (tb == NULL) { tb = Py_None; Py_INCREF(tb); } PyException_SetTraceback(v, tb); if (exception == NULL) return; /* Now we know v != NULL too */ if (set_sys_last_vars) { PySys_SetObject("last_type", exception); PySys_SetObject("last_value", v); PySys_SetObject("last_traceback", tb); } hook = PySys_GetObject("excepthook"); if (hook) { PyObject *args = PyTuple_Pack(3, exception, v, tb); PyObject *result = PyEval_CallObject(hook, args); if (result == NULL) { PyObject *exception2, *v2, *tb2; if (PyErr_ExceptionMatches(PyExc_SystemExit)) { handle_system_exit(); } PyErr_Fetch(&exception2, &v2, &tb2); PyErr_NormalizeException(&exception2, &v2, &tb2); /* It should not be possible for exception2 or v2 to be NULL. However PyErr_Display() can't tolerate NULLs, so just be safe. */ if (exception2 == NULL) { exception2 = Py_None; Py_INCREF(exception2); } if (v2 == NULL) { v2 = Py_None; Py_INCREF(v2); } fflush(stdout); PySys_WriteStderr("Error in sys.excepthook:\n"); PyErr_Display(exception2, v2, tb2); PySys_WriteStderr("\nOriginal exception was:\n"); PyErr_Display(exception, v, tb); Py_DECREF(exception2); Py_DECREF(v2); Py_XDECREF(tb2); } Py_XDECREF(result); Py_XDECREF(args); } else { PySys_WriteStderr("sys.excepthook is missing\n"); PyErr_Display(exception, v, tb); } Py_XDECREF(exception); Py_XDECREF(v); Py_XDECREF(tb); } static void print_exception(PyObject *f, PyObject *value) { int err = 0; PyObject *type, *tb; if (!PyExceptionInstance_Check(value)) { PyFile_WriteString("TypeError: print_exception(): Exception expected for value, ", f); PyFile_WriteString(Py_TYPE(value)->tp_name, f); PyFile_WriteString(" found\n", f); return; } Py_INCREF(value); fflush(stdout); type = (PyObject *) Py_TYPE(value); tb = PyException_GetTraceback(value); if (tb && tb != Py_None) err = PyTraceBack_Print(tb, f); if (err == 0 && PyObject_HasAttrString(value, "print_file_and_line")) { PyObject *message; const char *filename, *text; int lineno, offset; if (!parse_syntax_error(value, &message, &filename, &lineno, &offset, &text)) PyErr_Clear(); else { char buf[10]; PyFile_WriteString(" File \"", f); if (filename == NULL) PyFile_WriteString("", f); else PyFile_WriteString(filename, f); PyFile_WriteString("\", line ", f); PyOS_snprintf(buf, sizeof(buf), "%d", lineno); PyFile_WriteString(buf, f); PyFile_WriteString("\n", f); if (text != NULL) print_error_text(f, offset, text); Py_DECREF(value); value = message; /* Can't be bothered to check all those PyFile_WriteString() calls */ if (PyErr_Occurred()) err = -1; } } if (err) { /* Don't do anything else */ } else { PyObject* moduleName; char* className; assert(PyExceptionClass_Check(type)); className = PyExceptionClass_Name(type); if (className != NULL) { char *dot = strrchr(className, '.'); if (dot != NULL) className = dot+1; } moduleName = PyObject_GetAttrString(type, "__module__"); if (moduleName == NULL || !PyUnicode_Check(moduleName)) { Py_DECREF(moduleName); err = PyFile_WriteString("", f); } else { char* modstr = _PyUnicode_AsString(moduleName); if (modstr && strcmp(modstr, "builtins")) { err = PyFile_WriteString(modstr, f); err += PyFile_WriteString(".", f); } Py_DECREF(moduleName); } if (err == 0) { if (className == NULL) err = PyFile_WriteString("", f); else err = PyFile_WriteString(className, f); } } if (err == 0 && (value != Py_None)) { PyObject *s = PyObject_Str(value); /* only print colon if the str() of the object is not the empty string */ if (s == NULL) err = -1; else if (!PyUnicode_Check(s) || PyUnicode_GetSize(s) != 0) err = PyFile_WriteString(": ", f); if (err == 0) err = PyFile_WriteObject(s, f, Py_PRINT_RAW); Py_XDECREF(s); } /* try to write a newline in any case */ err += PyFile_WriteString("\n", f); Py_XDECREF(tb); Py_DECREF(value); /* If an error happened here, don't show it. XXX This is wrong, but too many callers rely on this behavior. */ if (err != 0) PyErr_Clear(); } static const char *cause_message = "\nThe above exception was the direct cause " "of the following exception:\n\n"; static const char *context_message = "\nDuring handling of the above exception, " "another exception occurred:\n\n"; static void print_exception_recursive(PyObject *f, PyObject *value, PyObject *seen) { int err = 0, res; PyObject *cause, *context; if (seen != NULL) { /* Exception chaining */ if (PySet_Add(seen, value) == -1) PyErr_Clear(); else if (PyExceptionInstance_Check(value)) { cause = PyException_GetCause(value); context = PyException_GetContext(value); if (cause) { res = PySet_Contains(seen, cause); if (res == -1) PyErr_Clear(); if (res == 0) { print_exception_recursive( f, cause, seen); err |= PyFile_WriteString( cause_message, f); } } if (context) { res = PySet_Contains(seen, context); if (res == -1) PyErr_Clear(); if (res == 0) { print_exception_recursive( f, context, seen); err |= PyFile_WriteString( context_message, f); } } Py_XDECREF(context); Py_XDECREF(cause); } } print_exception(f, value); if (err != 0) PyErr_Clear(); } void PyErr_Display(PyObject *exception, PyObject *value, PyObject *tb) { PyObject *seen; PyObject *f = PySys_GetObject("stderr"); if (f == Py_None) { /* pass */ } else if (f == NULL) { _PyObject_Dump(value); fprintf(stderr, "lost sys.stderr\n"); } else { /* We choose to ignore seen being possibly NULL, and report at least the main exception (it could be a MemoryError). */ seen = PySet_New(NULL); if (seen == NULL) PyErr_Clear(); print_exception_recursive(f, value, seen); Py_XDECREF(seen); } } PyObject * PyRun_StringFlags(const char *str, int start, PyObject *globals, PyObject *locals, PyCompilerFlags *flags) { PyObject *ret = NULL; mod_ty mod; PyArena *arena = PyArena_New(); if (arena == NULL) return NULL; mod = PyParser_ASTFromString(str, "", start, flags, arena); if (mod != NULL) ret = run_mod(mod, "", globals, locals, flags, arena); PyArena_Free(arena); return ret; } PyObject * PyRun_FileExFlags(FILE *fp, const char *filename, int start, PyObject *globals, PyObject *locals, int closeit, PyCompilerFlags *flags) { PyObject *ret; mod_ty mod; PyArena *arena = PyArena_New(); if (arena == NULL) return NULL; mod = PyParser_ASTFromFile(fp, filename, NULL, start, 0, 0, flags, NULL, arena); if (closeit) fclose(fp); if (mod == NULL) { PyArena_Free(arena); return NULL; } ret = run_mod(mod, filename, globals, locals, flags, arena); PyArena_Free(arena); return ret; } static void flush_io(void) { PyObject *f, *r; PyObject *type, *value, *traceback; /* Save the current exception */ PyErr_Fetch(&type, &value, &traceback); f = PySys_GetObject("stderr"); if (f != NULL) { r = PyObject_CallMethod(f, "flush", ""); if (r) Py_DECREF(r); else PyErr_Clear(); } f = PySys_GetObject("stdout"); if (f != NULL) { r = PyObject_CallMethod(f, "flush", ""); if (r) Py_DECREF(r); else PyErr_Clear(); } PyErr_Restore(type, value, traceback); } static PyObject * run_mod(mod_ty mod, const char *filename, PyObject *globals, PyObject *locals, PyCompilerFlags *flags, PyArena *arena) { PyCodeObject *co; PyObject *v; co = PyAST_Compile(mod, filename, flags, arena); if (co == NULL) return NULL; v = PyEval_EvalCode(co, globals, locals); Py_DECREF(co); return v; } static PyObject * run_pyc_file(FILE *fp, const char *filename, PyObject *globals, PyObject *locals, PyCompilerFlags *flags) { PyCodeObject *co; PyObject *v; long magic; long PyImport_GetMagicNumber(void); magic = PyMarshal_ReadLongFromFile(fp); if (magic != PyImport_GetMagicNumber()) { PyErr_SetString(PyExc_RuntimeError, "Bad magic number in .pyc file"); return NULL; } (void) PyMarshal_ReadLongFromFile(fp); v = PyMarshal_ReadLastObjectFromFile(fp); fclose(fp); if (v == NULL || !PyCode_Check(v)) { Py_XDECREF(v); PyErr_SetString(PyExc_RuntimeError, "Bad code object in .pyc file"); return NULL; } co = (PyCodeObject *)v; v = PyEval_EvalCode(co, globals, locals); if (v && flags) flags->cf_flags |= (co->co_flags & PyCF_MASK); Py_DECREF(co); return v; } PyObject * Py_CompileStringFlags(const char *str, const char *filename, int start, PyCompilerFlags *flags) { PyCodeObject *co; mod_ty mod; PyArena *arena = PyArena_New(); if (arena == NULL) return NULL; mod = PyParser_ASTFromString(str, filename, start, flags, arena); if (mod == NULL) { PyArena_Free(arena); return NULL; } if (flags && (flags->cf_flags & PyCF_ONLY_AST)) { PyObject *result = PyAST_mod2obj(mod); PyArena_Free(arena); return result; } co = PyAST_Compile(mod, filename, flags, arena); PyArena_Free(arena); return (PyObject *)co; } struct symtable * Py_SymtableString(const char *str, const char *filename, int start) { struct symtable *st; mod_ty mod; PyCompilerFlags flags; PyArena *arena = PyArena_New(); if (arena == NULL) return NULL; flags.cf_flags = 0; mod = PyParser_ASTFromString(str, filename, start, &flags, arena); if (mod == NULL) { PyArena_Free(arena); return NULL; } st = PySymtable_Build(mod, filename, 0); PyArena_Free(arena); return st; } /* Preferred access to parser is through AST. */ mod_ty PyParser_ASTFromString(const char *s, const char *filename, int start, PyCompilerFlags *flags, PyArena *arena) { mod_ty mod; PyCompilerFlags localflags; perrdetail err; int iflags = PARSER_FLAGS(flags); node *n = PyParser_ParseStringFlagsFilenameEx(s, filename, &_PyParser_Grammar, start, &err, &iflags); if (flags == NULL) { localflags.cf_flags = 0; flags = &localflags; } if (n) { flags->cf_flags |= iflags & PyCF_MASK; mod = PyAST_FromNode(n, flags, filename, arena); PyNode_Free(n); return mod; } else { err_input(&err); return NULL; } } mod_ty PyParser_ASTFromFile(FILE *fp, const char *filename, const char* enc, int start, char *ps1, char *ps2, PyCompilerFlags *flags, int *errcode, PyArena *arena) { mod_ty mod; PyCompilerFlags localflags; perrdetail err; int iflags = PARSER_FLAGS(flags); node *n = PyParser_ParseFileFlagsEx(fp, filename, enc, &_PyParser_Grammar, start, ps1, ps2, &err, &iflags); if (flags == NULL) { localflags.cf_flags = 0; flags = &localflags; } if (n) { flags->cf_flags |= iflags & PyCF_MASK; mod = PyAST_FromNode(n, flags, filename, arena); PyNode_Free(n); return mod; } else { err_input(&err); if (errcode) *errcode = err.error; return NULL; } } /* Simplified interface to parsefile -- return node or set exception */ node * PyParser_SimpleParseFileFlags(FILE *fp, const char *filename, int start, int flags) { perrdetail err; node *n = PyParser_ParseFileFlags(fp, filename, NULL, &_PyParser_Grammar, start, NULL, NULL, &err, flags); if (n == NULL) err_input(&err); return n; } /* Simplified interface to parsestring -- return node or set exception */ node * PyParser_SimpleParseStringFlags(const char *str, int start, int flags) { perrdetail err; node *n = PyParser_ParseStringFlags(str, &_PyParser_Grammar, start, &err, flags); if (n == NULL) err_input(&err); return n; } node * PyParser_SimpleParseStringFlagsFilename(const char *str, const char *filename, int start, int flags) { perrdetail err; node *n = PyParser_ParseStringFlagsFilename(str, filename, &_PyParser_Grammar, start, &err, flags); if (n == NULL) err_input(&err); return n; } node * PyParser_SimpleParseStringFilename(const char *str, const char *filename, int start) { return PyParser_SimpleParseStringFlagsFilename(str, filename, start, 0); } /* May want to move a more generalized form of this to parsetok.c or even parser modules. */ void PyParser_SetError(perrdetail *err) { err_input(err); } /* Set the error appropriate to the given input error code (see errcode.h) */ static void err_input(perrdetail *err) { PyObject *v, *w, *errtype, *errtext; PyObject* u = NULL; char *msg = NULL; errtype = PyExc_SyntaxError; switch (err->error) { case E_SYNTAX: errtype = PyExc_IndentationError; if (err->expected == INDENT) msg = "expected an indented block"; else if (err->token == INDENT) msg = "unexpected indent"; else if (err->token == DEDENT) msg = "unexpected unindent"; else { errtype = PyExc_SyntaxError; msg = "invalid syntax"; } break; case E_TOKEN: msg = "invalid token"; break; case E_EOFS: msg = "EOF while scanning triple-quoted string literal"; break; case E_EOLS: msg = "EOL while scanning string literal"; break; case E_INTR: if (!PyErr_Occurred()) PyErr_SetNone(PyExc_KeyboardInterrupt); goto cleanup; case E_NOMEM: PyErr_NoMemory(); goto cleanup; case E_EOF: msg = "unexpected EOF while parsing"; break; case E_TABSPACE: errtype = PyExc_TabError; msg = "inconsistent use of tabs and spaces in indentation"; break; case E_OVERFLOW: msg = "expression too long"; break; case E_DEDENT: errtype = PyExc_IndentationError; msg = "unindent does not match any outer indentation level"; break; case E_TOODEEP: errtype = PyExc_IndentationError; msg = "too many levels of indentation"; break; case E_DECODE: { PyObject *type, *value, *tb; PyErr_Fetch(&type, &value, &tb); if (value != NULL) { u = PyObject_Str(value); if (u != NULL) { msg = _PyUnicode_AsString(u); } } if (msg == NULL) msg = "unknown decode error"; Py_XDECREF(type); Py_XDECREF(value); Py_XDECREF(tb); break; } case E_LINECONT: msg = "unexpected character after line continuation character"; break; case E_IDENTIFIER: msg = "invalid character in identifier"; break; default: fprintf(stderr, "error=%d\n", err->error); msg = "unknown parsing error"; break; } /* err->text may not be UTF-8 in case of decoding errors. Explicitly convert to an object. */ if (!err->text) { errtext = Py_None; Py_INCREF(Py_None); } else { errtext = PyUnicode_DecodeUTF8(err->text, strlen(err->text), "replace"); } v = Py_BuildValue("(ziiN)", err->filename, err->lineno, err->offset, errtext); w = NULL; if (v != NULL) w = Py_BuildValue("(sO)", msg, v); Py_XDECREF(u); Py_XDECREF(v); PyErr_SetObject(errtype, w); Py_XDECREF(w); cleanup: if (err->text != NULL) { PyObject_FREE(err->text); err->text = NULL; } } /* Print fatal error message and abort */ void Py_FatalError(const char *msg) { fprintf(stderr, "Fatal Python error: %s\n", msg); if (PyErr_Occurred()) { PyErr_Print(); } #ifdef MS_WINDOWS { size_t len = strlen(msg); WCHAR* buffer; size_t i; /* Convert the message to wchar_t. This uses a simple one-to-one conversion, assuming that the this error message actually uses ASCII only. If this ceases to be true, we will have to convert. */ buffer = alloca( (len+1) * (sizeof *buffer)); for( i=0; i<=len; ++i) buffer[i] = msg[i]; OutputDebugStringW(L"Fatal Python error: "); OutputDebugStringW(buffer); OutputDebugStringW(L"\n"); } #ifdef _DEBUG DebugBreak(); #endif #endif /* MS_WINDOWS */ abort(); } /* Clean up and exit */ #ifdef WITH_THREAD #include "pythread.h" #endif static void (*pyexitfunc)(void) = NULL; /* For the atexit module. */ void _Py_PyAtExit(void (*func)(void)) { pyexitfunc = func; } static void call_py_exitfuncs(void) { if (pyexitfunc == NULL) return; (*pyexitfunc)(); PyErr_Clear(); } #define NEXITFUNCS 32 static void (*exitfuncs[NEXITFUNCS])(void); static int nexitfuncs = 0; int Py_AtExit(void (*func)(void)) { if (nexitfuncs >= NEXITFUNCS) return -1; exitfuncs[nexitfuncs++] = func; return 0; } static void call_ll_exitfuncs(void) { while (nexitfuncs > 0) (*exitfuncs[--nexitfuncs])(); fflush(stdout); fflush(stderr); } void Py_Exit(int sts) { Py_Finalize(); exit(sts); } static void initsigs(void) { #ifdef SIGPIPE PyOS_setsig(SIGPIPE, SIG_IGN); #endif #ifdef SIGXFZ PyOS_setsig(SIGXFZ, SIG_IGN); #endif #ifdef SIGXFSZ PyOS_setsig(SIGXFSZ, SIG_IGN); #endif PyOS_InitInterrupts(); /* May imply initsignal() */ } /* * The file descriptor fd is considered ``interactive'' if either * a) isatty(fd) is TRUE, or * b) the -i flag was given, and the filename associated with * the descriptor is NULL or "" or "???". */ int Py_FdIsInteractive(FILE *fp, const char *filename) { if (isatty((int)fileno(fp))) return 1; if (!Py_InteractiveFlag) return 0; return (filename == NULL) || (strcmp(filename, "") == 0) || (strcmp(filename, "???") == 0); } #if defined(USE_STACKCHECK) #if defined(WIN32) && defined(_MSC_VER) /* Stack checking for Microsoft C */ #include #include /* * Return non-zero when we run out of memory on the stack; zero otherwise. */ int PyOS_CheckStack(void) { __try { /* alloca throws a stack overflow exception if there's not enough space left on the stack */ alloca(PYOS_STACK_MARGIN * sizeof(void*)); return 0; } __except (GetExceptionCode() == STATUS_STACK_OVERFLOW ? EXCEPTION_EXECUTE_HANDLER : EXCEPTION_CONTINUE_SEARCH) { int errcode = _resetstkoflw(); if (errcode == 0) { Py_FatalError("Could not reset the stack!"); } } return 1; } #endif /* WIN32 && _MSC_VER */ /* Alternate implementations can be added here... */ #endif /* USE_STACKCHECK */ /* Wrappers around sigaction() or signal(). */ PyOS_sighandler_t PyOS_getsig(int sig) { #ifdef HAVE_SIGACTION struct sigaction context; if (sigaction(sig, NULL, &context) == -1) return SIG_ERR; return context.sa_handler; #else PyOS_sighandler_t handler; /* Special signal handling for the secure CRT in Visual Studio 2005 */ #if defined(_MSC_VER) && _MSC_VER >= 1400 switch (sig) { /* Only these signals are valid */ case SIGINT: case SIGILL: case SIGFPE: case SIGSEGV: case SIGTERM: case SIGBREAK: case SIGABRT: break; /* Don't call signal() with other values or it will assert */ default: return SIG_ERR; } #endif /* _MSC_VER && _MSC_VER >= 1400 */ handler = signal(sig, SIG_IGN); if (handler != SIG_ERR) signal(sig, handler); return handler; #endif } PyOS_sighandler_t PyOS_setsig(int sig, PyOS_sighandler_t handler) { #ifdef HAVE_SIGACTION struct sigaction context, ocontext; context.sa_handler = handler; sigemptyset(&context.sa_mask); context.sa_flags = 0; if (sigaction(sig, &context, &ocontext) == -1) return SIG_ERR; return ocontext.sa_handler; #else PyOS_sighandler_t oldhandler; oldhandler = signal(sig, handler); #ifdef HAVE_SIGINTERRUPT siginterrupt(sig, 1); #endif return oldhandler; #endif } /* Deprecated C API functions still provided for binary compatiblity */ #undef PyParser_SimpleParseFile PyAPI_FUNC(node *) PyParser_SimpleParseFile(FILE *fp, const char *filename, int start) { return PyParser_SimpleParseFileFlags(fp, filename, start, 0); } #undef PyParser_SimpleParseString PyAPI_FUNC(node *) PyParser_SimpleParseString(const char *str, int start) { return PyParser_SimpleParseStringFlags(str, start, 0); } #undef PyRun_AnyFile PyAPI_FUNC(int) PyRun_AnyFile(FILE *fp, const char *name) { return PyRun_AnyFileExFlags(fp, name, 0, NULL); } #undef PyRun_AnyFileEx PyAPI_FUNC(int) PyRun_AnyFileEx(FILE *fp, const char *name, int closeit) { return PyRun_AnyFileExFlags(fp, name, closeit, NULL); } #undef PyRun_AnyFileFlags PyAPI_FUNC(int) PyRun_AnyFileFlags(FILE *fp, const char *name, PyCompilerFlags *flags) { return PyRun_AnyFileExFlags(fp, name, 0, flags); } #undef PyRun_File PyAPI_FUNC(PyObject *) PyRun_File(FILE *fp, const char *p, int s, PyObject *g, PyObject *l) { return PyRun_FileExFlags(fp, p, s, g, l, 0, NULL); } #undef PyRun_FileEx PyAPI_FUNC(PyObject *) PyRun_FileEx(FILE *fp, const char *p, int s, PyObject *g, PyObject *l, int c) { return PyRun_FileExFlags(fp, p, s, g, l, c, NULL); } #undef PyRun_FileFlags PyAPI_FUNC(PyObject *) PyRun_FileFlags(FILE *fp, const char *p, int s, PyObject *g, PyObject *l, PyCompilerFlags *flags) { return PyRun_FileExFlags(fp, p, s, g, l, 0, flags); } #undef PyRun_SimpleFile PyAPI_FUNC(int) PyRun_SimpleFile(FILE *f, const char *p) { return PyRun_SimpleFileExFlags(f, p, 0, NULL); } #undef PyRun_SimpleFileEx PyAPI_FUNC(int) PyRun_SimpleFileEx(FILE *f, const char *p, int c) { return PyRun_SimpleFileExFlags(f, p, c, NULL); } #undef PyRun_String PyAPI_FUNC(PyObject *) PyRun_String(const char *str, int s, PyObject *g, PyObject *l) { return PyRun_StringFlags(str, s, g, l, NULL); } #undef PyRun_SimpleString PyAPI_FUNC(int) PyRun_SimpleString(const char *s) { return PyRun_SimpleStringFlags(s, NULL); } #undef Py_CompileString PyAPI_FUNC(PyObject *) Py_CompileString(const char *str, const char *p, int s) { return Py_CompileStringFlags(str, p, s, NULL); } #undef PyRun_InteractiveOne PyAPI_FUNC(int) PyRun_InteractiveOne(FILE *f, const char *p) { return PyRun_InteractiveOneFlags(f, p, NULL); } #undef PyRun_InteractiveLoop PyAPI_FUNC(int) PyRun_InteractiveLoop(FILE *f, const char *p) { return PyRun_InteractiveLoopFlags(f, p, NULL); } #ifdef __cplusplus } #endif 94'>1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 
#include "Python.h"
#include "pycore_fileutils.h"     // fileutils definitions
#include "pycore_runtime.h"       // _PyRuntime
#include "pycore_pystate.h"       // _Py_AssertHoldsTstate()
#include "osdefs.h"               // SEP

#include <stdlib.h>               // mbstowcs()
#ifdef HAVE_UNISTD_H
#  include <unistd.h>             // getcwd()
#endif

#ifdef MS_WINDOWS
#  include <malloc.h>
#  include <windows.h>
#  include <winioctl.h>             // FILE_DEVICE_* constants
#  include "pycore_fileutils_windows.h" // FILE_STAT_BASIC_INFORMATION
#  if defined(MS_WINDOWS_GAMES) && !defined(MS_WINDOWS_DESKTOP)
#    define PATHCCH_ALLOW_LONG_PATHS 0x01
#  else
#    include <pathcch.h>            // PathCchCombineEx
#  endif
extern int winerror_to_errno(int);
#endif

#ifdef HAVE_LANGINFO_H
#  include <langinfo.h>           // nl_langinfo(CODESET)
#endif

#ifdef HAVE_SYS_IOCTL_H
#include <sys/ioctl.h>
#endif

#ifdef HAVE_NON_UNICODE_WCHAR_T_REPRESENTATION
#  include <iconv.h>              // iconv_open()
#endif

#ifdef HAVE_FCNTL_H
#  include <fcntl.h>              // fcntl(F_GETFD)
#endif

#ifdef O_CLOEXEC
/* Does open() support the O_CLOEXEC flag? Possible values:

   -1: unknown
    0: open() ignores O_CLOEXEC flag, ex: Linux kernel older than 2.6.23
    1: open() supports O_CLOEXEC flag, close-on-exec is set

   The flag is used by _Py_open(), _Py_open_noraise(), io.FileIO
   and os.open(). */
int _Py_open_cloexec_works = -1;
#endif

// The value must be the same in unicodeobject.c.
#define MAX_UNICODE 0x10ffff

// mbstowcs() and mbrtowc() errors
static const size_t DECODE_ERROR = ((size_t)-1);
static const size_t INCOMPLETE_CHARACTER = (size_t)-2;


static int
get_surrogateescape(_Py_error_handler errors, int *surrogateescape)
{
    switch (errors)
    {
    case _Py_ERROR_STRICT:
        *surrogateescape = 0;
        return 0;
    case _Py_ERROR_SURROGATEESCAPE:
        *surrogateescape = 1;
        return 0;
    default:
        return -1;
    }
}


PyObject *
_Py_device_encoding(int fd)
{
    int valid;
    Py_BEGIN_ALLOW_THREADS
    _Py_BEGIN_SUPPRESS_IPH
    valid = isatty(fd);
    _Py_END_SUPPRESS_IPH
    Py_END_ALLOW_THREADS
    if (!valid)
        Py_RETURN_NONE;

#ifdef MS_WINDOWS
#ifdef HAVE_WINDOWS_CONSOLE_IO
    UINT cp;
    if (fd == 0)
        cp = GetConsoleCP();
    else if (fd == 1 || fd == 2)
        cp = GetConsoleOutputCP();
    else
        cp = 0;
    /* GetConsoleCP() and GetConsoleOutputCP() return 0 if the application
       has no console */
    if (cp == 0) {
        Py_RETURN_NONE;
    }

    return PyUnicode_FromFormat("cp%u", (unsigned int)cp);
#else
    Py_RETURN_NONE;
#endif /* HAVE_WINDOWS_CONSOLE_IO */
#else
    if (_PyRuntime.preconfig.utf8_mode) {
        _Py_DECLARE_STR(utf_8, "utf-8");
        return &_Py_STR(utf_8);
    }
    return _Py_GetLocaleEncodingObject();
#endif
}


static int
is_valid_wide_char(wchar_t ch)
{
#ifdef HAVE_NON_UNICODE_WCHAR_T_REPRESENTATION
    /* Oracle Solaris doesn't use Unicode code points as wchar_t encoding
       for non-Unicode locales, which makes values higher than MAX_UNICODE
       possibly valid. */
    return 1;
#endif
    if (Py_UNICODE_IS_SURROGATE(ch)) {
        // Reject lone surrogate characters
        return 0;
    }
    if (ch > MAX_UNICODE) {
        // bpo-35883: Reject characters outside [U+0000; U+10ffff] range.
        // The glibc mbstowcs() UTF-8 decoder does not respect the RFC 3629,
        // it creates characters outside the [U+0000; U+10ffff] range:
        // https://sourceware.org/bugzilla/show_bug.cgi?id=2373
        return 0;
    }
    return 1;
}


static size_t
_Py_mbstowcs(wchar_t *dest, const char *src, size_t n)
{
    size_t count = mbstowcs(dest, src, n);
    if (dest != NULL && count != DECODE_ERROR) {
        for (size_t i=0; i < count; i++) {
            wchar_t ch = dest[i];
            if (!is_valid_wide_char(ch)) {
                return DECODE_ERROR;
            }
        }
    }
    return count;
}


#ifdef HAVE_MBRTOWC
static size_t
_Py_mbrtowc(wchar_t *pwc, const char *str, size_t len, mbstate_t *pmbs)
{
    assert(pwc != NULL);
    size_t count = mbrtowc(pwc, str, len, pmbs);
    if (count != 0 && count != DECODE_ERROR && count != INCOMPLETE_CHARACTER) {
        if (!is_valid_wide_char(*pwc)) {
            return DECODE_ERROR;
        }
    }
    return count;
}
#endif


#if !defined(_Py_FORCE_UTF8_FS_ENCODING) && !defined(MS_WINDOWS)

#define USE_FORCE_ASCII

extern int _Py_normalize_encoding(const char *, char *, size_t);

/* Workaround FreeBSD and OpenIndiana locale encoding issue with the C locale
   and POSIX locale. nl_langinfo(CODESET) announces an alias of the
   ASCII encoding, whereas mbstowcs() and wcstombs() functions use the
   ISO-8859-1 encoding. The problem is that os.fsencode() and os.fsdecode() use
   locale.getpreferredencoding() codec. For example, if command line arguments
   are decoded by mbstowcs() and encoded back by os.fsencode(), we get a
   UnicodeEncodeError instead of retrieving the original byte string.

   The workaround is enabled if setlocale(LC_CTYPE, NULL) returns "C",
   nl_langinfo(CODESET) announces "ascii" (or an alias to ASCII), and at least
   one byte in range 0x80-0xff can be decoded from the locale encoding. The
   workaround is also enabled on error, for example if getting the locale
   failed.

   On HP-UX with the C locale or the POSIX locale, nl_langinfo(CODESET)
   announces "roman8" but mbstowcs() uses Latin1 in practice. Force also the
   ASCII encoding in this case.

   Values of force_ascii:

       1: the workaround is used: Py_EncodeLocale() uses
          encode_ascii_surrogateescape() and Py_DecodeLocale() uses
          decode_ascii()
       0: the workaround is not used: Py_EncodeLocale() uses wcstombs() and
          Py_DecodeLocale() uses mbstowcs()
      -1: unknown, need to call check_force_ascii() to get the value
*/
#define force_ascii (_PyRuntime.fileutils.force_ascii)

static int
check_force_ascii(void)
{
    char *loc = setlocale(LC_CTYPE, NULL);
    if (loc == NULL) {
        goto error;
    }
    if (strcmp(loc, "C") != 0 && strcmp(loc, "POSIX") != 0) {
        /* the LC_CTYPE locale is different than C and POSIX */
        return 0;
    }

#if defined(HAVE_LANGINFO_H) && defined(CODESET)
    const char *codeset = nl_langinfo(CODESET);
    if (!codeset || codeset[0] == '\0') {
        /* CODESET is not set or empty */
        goto error;
    }

    char encoding[20];   /* longest name: "iso_646.irv_1991\0" */
    if (!_Py_normalize_encoding(codeset, encoding, sizeof(encoding))) {
        goto error;
    }

#ifdef __hpux
    if (strcmp(encoding, "roman8") == 0) {
        unsigned char ch;
        wchar_t wch;
        size_t res;

        ch = (unsigned char)0xA7;
        res = _Py_mbstowcs(&wch, (char*)&ch, 1);
        if (res != DECODE_ERROR && wch == L'\xA7') {
            /* On HP-UX with C locale or the POSIX locale,
               nl_langinfo(CODESET) announces "roman8", whereas mbstowcs() uses
               Latin1 encoding in practice. Force ASCII in this case.

               Roman8 decodes 0xA7 to U+00CF. Latin1 decodes 0xA7 to U+00A7. */
            return 1;
        }
    }
#else
    const char* ascii_aliases[] = {
        "ascii",
        /* Aliases from Lib/encodings/aliases.py */
        "646",
        "ansi_x3.4_1968",
        "ansi_x3.4_1986",
        "ansi_x3_4_1968",
        "cp367",
        "csascii",
        "ibm367",
        "iso646_us",
        "iso_646.irv_1991",
        "iso_ir_6",
        "us",
        "us_ascii",
        NULL
    };

    int is_ascii = 0;
    for (const char **alias=ascii_aliases; *alias != NULL; alias++) {
        if (strcmp(encoding, *alias) == 0) {
            is_ascii = 1;
            break;
        }
    }
    if (!is_ascii) {
        /* nl_langinfo(CODESET) is not "ascii" or an alias of ASCII */
        return 0;
    }

    for (unsigned int i=0x80; i<=0xff; i++) {
        char ch[1];
        wchar_t wch[1];
        size_t res;

        unsigned uch = (unsigned char)i;
        ch[0] = (char)uch;
        res = _Py_mbstowcs(wch, ch, 1);
        if (res != DECODE_ERROR) {
            /* decoding a non-ASCII character from the locale encoding succeed:
               the locale encoding is not ASCII, force ASCII */
            return 1;
        }
    }
    /* None of the bytes in the range 0x80-0xff can be decoded from the locale
       encoding: the locale encoding is really ASCII */
#endif   /* !defined(__hpux) */
    return 0;
#else
    /* nl_langinfo(CODESET) is not available: always force ASCII */
    return 1;
#endif   /* defined(HAVE_LANGINFO_H) && defined(CODESET) */

error:
    /* if an error occurred, force the ASCII encoding */
    return 1;
}


int
_Py_GetForceASCII(void)
{
    if (force_ascii == -1) {
        force_ascii = check_force_ascii();
    }
    return force_ascii;
}


void
_Py_ResetForceASCII(void)
{
    force_ascii = -1;
}


static int
encode_ascii(const wchar_t *text, char **str,
             size_t *error_pos, const char **reason,
             int raw_malloc, _Py_error_handler errors)
{
    char *result = NULL, *out;
    size_t len, i;
    wchar_t ch;

    int surrogateescape;
    if (get_surrogateescape(errors, &surrogateescape) < 0) {
        return -3;
    }

    len = wcslen(text);

    /* +1 for NULL byte */
    if (raw_malloc) {
        result = PyMem_RawMalloc(len + 1);
    }
    else {
        result = PyMem_Malloc(len + 1);
    }
    if (result == NULL) {
        return -1;
    }

    out = result;
    for (i=0; i<len; i++) {
        ch = text[i];

        if (ch <= 0x7f) {
            /* ASCII character */
            *out++ = (char)ch;
        }
        else if (surrogateescape && 0xdc80 <= ch && ch <= 0xdcff) {
            /* UTF-8b surrogate */
            *out++ = (char)(ch - 0xdc00);
        }
        else {
            if (raw_malloc) {
                PyMem_RawFree(result);
            }
            else {
                PyMem_Free(result);
            }
            if (error_pos != NULL) {
                *error_pos = i;
            }
            if (reason) {
                *reason = "encoding error";
            }
            return -2;
        }
    }
    *out = '\0';
    *str = result;
    return 0;
}
#else
int
_Py_GetForceASCII(void)
{
    return 0;
}

void
_Py_ResetForceASCII(void)
{
    /* nothing to do */
}
#endif   /* !defined(_Py_FORCE_UTF8_FS_ENCODING) && !defined(MS_WINDOWS) */


#if !defined(HAVE_MBRTOWC) || defined(USE_FORCE_ASCII)
static int
decode_ascii(const char *arg, wchar_t **wstr, size_t *wlen,
             const char **reason, _Py_error_handler errors)
{
    wchar_t *res;
    unsigned char *in;
    wchar_t *out;
    size_t argsize = strlen(arg) + 1;

    int surrogateescape;
    if (get_surrogateescape(errors, &surrogateescape) < 0) {
        return -3;
    }

    if (argsize > PY_SSIZE_T_MAX / sizeof(wchar_t)) {
        return -1;
    }
    res = PyMem_RawMalloc(argsize * sizeof(wchar_t));
    if (!res) {
        return -1;
    }

    out = res;
    for (in = (unsigned char*)arg; *in; in++) {
        unsigned char ch = *in;
        if (ch < 128) {
            *out++ = ch;
        }
        else {
            if (!surrogateescape) {
                PyMem_RawFree(res);
                if (wlen) {
                    *wlen = in - (unsigned char*)arg;
                }
                if (reason) {
                    *reason = "decoding error";
                }
                return -2;
            }
            *out++ = 0xdc00 + ch;
        }
    }
    *out = 0;

    if (wlen != NULL) {
        *wlen = out - res;
    }
    *wstr = res;
    return 0;
}
#endif   /* !HAVE_MBRTOWC */

static int
decode_current_locale(const char* arg, wchar_t **wstr, size_t *wlen,
                      const char **reason, _Py_error_handler errors)
{
    wchar_t *res;
    size_t argsize;
    size_t count;
#ifdef HAVE_MBRTOWC
    unsigned char *in;
    wchar_t *out;
    mbstate_t mbs;
#endif

    int surrogateescape;
    if (get_surrogateescape(errors, &surrogateescape) < 0) {
        return -3;
    }

#ifdef HAVE_BROKEN_MBSTOWCS
    /* Some platforms have a broken implementation of
     * mbstowcs which does not count the characters that
     * would result from conversion.  Use an upper bound.
     */
    argsize = strlen(arg);
#else
    argsize = _Py_mbstowcs(NULL, arg, 0);
#endif
    if (argsize != DECODE_ERROR) {
        if (argsize > PY_SSIZE_T_MAX / sizeof(wchar_t) - 1) {
            return -1;
        }
        res = (wchar_t *)PyMem_RawMalloc((argsize + 1) * sizeof(wchar_t));
        if (!res) {
            return -1;
        }

        count = _Py_mbstowcs(res, arg, argsize + 1);
        if (count != DECODE_ERROR) {
            *wstr = res;
            if (wlen != NULL) {
                *wlen = count;
            }
            return 0;
        }
        PyMem_RawFree(res);
    }

    /* Conversion failed. Fall back to escaping with surrogateescape. */
#ifdef HAVE_MBRTOWC
    /* Try conversion with mbrtwoc (C99), and escape non-decodable bytes. */

    /* Overallocate; as multi-byte characters are in the argument, the
       actual output could use less memory. */
    argsize = strlen(arg) + 1;
    if (argsize > PY_SSIZE_T_MAX / sizeof(wchar_t)) {
        return -1;
    }
    res = (wchar_t*)PyMem_RawMalloc(argsize * sizeof(wchar_t));
    if (!res) {
        return -1;
    }

    in = (unsigned char*)arg;
    out = res;
    memset(&mbs, 0, sizeof mbs);
    while (argsize) {
        size_t converted = _Py_mbrtowc(out, (char*)in, argsize, &mbs);
        if (converted == 0) {
            /* Reached end of string; null char stored. */
            break;
        }

        if (converted == INCOMPLETE_CHARACTER) {
            /* Incomplete character. This should never happen,
               since we provide everything that we have -
               unless there is a bug in the C library, or I
               misunderstood how mbrtowc works. */
            goto decode_error;
        }

        if (converted == DECODE_ERROR) {
            if (!surrogateescape) {
                goto decode_error;
            }

            /* Decoding error. Escape as UTF-8b, and start over in the initial
               shift state. */
            *out++ = 0xdc00 + *in++;
            argsize--;
            memset(&mbs, 0, sizeof mbs);
            continue;
        }

        // _Py_mbrtowc() reject lone surrogate characters
        assert(!Py_UNICODE_IS_SURROGATE(*out));

        /* successfully converted some bytes */
        in += converted;
        argsize -= converted;
        out++;
    }
    if (wlen != NULL) {
        *wlen = out - res;
    }
    *wstr = res;
    return 0;

decode_error:
    PyMem_RawFree(res);
    if (wlen) {
        *wlen = in - (unsigned char*)arg;
    }
    if (reason) {
        *reason = "decoding error";
    }
    return -2;
#else   /* HAVE_MBRTOWC */
    /* Cannot use C locale for escaping; manually escape as if charset
       is ASCII (i.e. escape all bytes > 128. This will still roundtrip
       correctly in the locale's charset, which must be an ASCII superset. */
    return decode_ascii(arg, wstr, wlen, reason, errors);
#endif   /* HAVE_MBRTOWC */
}


/* Decode a byte string from the locale encoding.

   Use the strict error handler if 'surrogateescape' is zero.  Use the
   surrogateescape error handler if 'surrogateescape' is non-zero: undecodable
   bytes are decoded as characters in range U+DC80..U+DCFF. If a byte sequence
   can be decoded as a surrogate character, escape the bytes using the
   surrogateescape error handler instead of decoding them.

   On success, return 0 and write the newly allocated wide character string into
   *wstr (use PyMem_RawFree() to free the memory). If wlen is not NULL, write
   the number of wide characters excluding the null character into *wlen.

   On memory allocation failure, return -1.

   On decoding error, return -2. If wlen is not NULL, write the start of
   invalid byte sequence in the input string into *wlen. If reason is not NULL,
   write the decoding error message into *reason.

   Return -3 if the error handler 'errors' is not supported.

   Use the Py_EncodeLocaleEx() function to encode the character string back to
   a byte string. */
int
_Py_DecodeLocaleEx(const char* arg, wchar_t **wstr, size_t *wlen,
                   const char **reason,
                   int current_locale, _Py_error_handler errors)
{
    if (current_locale) {
#ifdef _Py_FORCE_UTF8_LOCALE
        return _Py_DecodeUTF8Ex(arg, strlen(arg), wstr, wlen, reason,
                                errors);
#else
        return decode_current_locale(arg, wstr, wlen, reason, errors);
#endif
    }

#ifdef _Py_FORCE_UTF8_FS_ENCODING
    return _Py_DecodeUTF8Ex(arg, strlen(arg), wstr, wlen, reason,
                            errors);
#else
    int use_utf8 = (_PyRuntime.preconfig.utf8_mode >= 1);
#ifdef MS_WINDOWS
    use_utf8 |= (_PyRuntime.preconfig.legacy_windows_fs_encoding == 0);
#endif
    if (use_utf8) {
        return _Py_DecodeUTF8Ex(arg, strlen(arg), wstr, wlen, reason,
                                errors);
    }

#ifdef USE_FORCE_ASCII
    if (force_ascii == -1) {
        force_ascii = check_force_ascii();
    }

    if (force_ascii) {
        /* force ASCII encoding to workaround mbstowcs() issue */
        return decode_ascii(arg, wstr, wlen, reason, errors);
    }
#endif

    return decode_current_locale(arg, wstr, wlen, reason, errors);
#endif   /* !_Py_FORCE_UTF8_FS_ENCODING */
}


/* Decode a byte string from the locale encoding with the
   surrogateescape error handler: undecodable bytes are decoded as characters
   in range U+DC80..U+DCFF. If a byte sequence can be decoded as a surrogate
   character, escape the bytes using the surrogateescape error handler instead
   of decoding them.

   Return a pointer to a newly allocated wide character string, use
   PyMem_RawFree() to free the memory. If size is not NULL, write the number of
   wide characters excluding the null character into *size

   Return NULL on decoding error or memory allocation error. If *size* is not
   NULL, *size is set to (size_t)-1 on memory error or set to (size_t)-2 on
   decoding error.

   Decoding errors should never happen, unless there is a bug in the C
   library.

   Use the Py_EncodeLocale() function to encode the character string back to a
   byte string. */
wchar_t*
Py_DecodeLocale(const char* arg, size_t *wlen)
{
    wchar_t *wstr;
    int res = _Py_DecodeLocaleEx(arg, &wstr, wlen,
                                 NULL, 0,
                                 _Py_ERROR_SURROGATEESCAPE);
    if (res != 0) {
        assert(res != -3);
        if (wlen != NULL) {
            *wlen = (size_t)res;
        }
        return NULL;
    }
    return wstr;
}


static int
encode_current_locale(const wchar_t *text, char **str,
                      size_t *error_pos, const char **reason,
                      int raw_malloc, _Py_error_handler errors)
{
    const size_t len = wcslen(text);
    char *result = NULL, *bytes = NULL;
    size_t i, size, converted;
    wchar_t c, buf[2];

    int surrogateescape;
    if (get_surrogateescape(errors, &surrogateescape) < 0) {
        return -3;
    }

    /* The function works in two steps:
       1. compute the length of the output buffer in bytes (size)
       2. outputs the bytes */
    size = 0;
    buf[1] = 0;
    while (1) {
        for (i=0; i < len; i++) {
            c = text[i];
            if (c >= 0xdc80 && c <= 0xdcff) {
                if (!surrogateescape) {
                    goto encode_error;
                }
                /* UTF-8b surrogate */
                if (bytes != NULL) {
                    *bytes++ = c - 0xdc00;
                    size--;
                }
                else {
                    size++;
                }
                continue;
            }
            else {
                buf[0] = c;
                if (bytes != NULL) {
                    converted = wcstombs(bytes, buf, size);
                }
                else {
                    converted = wcstombs(NULL, buf, 0);
                }
                if (converted == DECODE_ERROR) {
                    goto encode_error;
                }
                if (bytes != NULL) {
                    bytes += converted;
                    size -= converted;
                }
                else {
                    size += converted;
                }
            }
        }
        if (result != NULL) {
            *bytes = '\0';
            break;
        }

        size += 1; /* nul byte at the end */
        if (raw_malloc) {
            result = PyMem_RawMalloc(size);
        }
        else {
            result = PyMem_Malloc(size);
        }
        if (result == NULL) {
            return -1;
        }
        bytes = result;
    }
    *str = result;
    return 0;

encode_error:
    if (raw_malloc) {
        PyMem_RawFree(result);
    }
    else {
        PyMem_Free(result);
    }
    if (error_pos != NULL) {
        *error_pos = i;
    }
    if (reason) {
        *reason = "encoding error";
    }
    return -2;
}


/* Encode a string to the locale encoding.

   Parameters:

   * raw_malloc: if non-zero, allocate memory using PyMem_RawMalloc() instead
     of PyMem_Malloc().
   * current_locale: if non-zero, use the current LC_CTYPE, otherwise use
     Python filesystem encoding.
   * errors: error handler like "strict" or "surrogateescape".

   Return value:

    0: success, *str is set to a newly allocated decoded string.
   -1: memory allocation failure
   -2: encoding error, set *error_pos and *reason (if set).
   -3: the error handler 'errors' is not supported.
 */
static int
encode_locale_ex(const wchar_t *text, char **str, size_t *error_pos,
                 const char **reason,
                 int raw_malloc, int current_locale, _Py_error_handler errors)
{
    if (current_locale) {
#ifdef _Py_FORCE_UTF8_LOCALE
        return _Py_EncodeUTF8Ex(text, str, error_pos, reason,
                                raw_malloc, errors);
#else
        return encode_current_locale(text, str, error_pos, reason,
                                     raw_malloc, errors);
#endif
    }

#ifdef _Py_FORCE_UTF8_FS_ENCODING
    return _Py_EncodeUTF8Ex(text, str, error_pos, reason,
                            raw_malloc, errors);
#else
    int use_utf8 = (_PyRuntime.preconfig.utf8_mode >= 1);
#ifdef MS_WINDOWS
    use_utf8 |= (_PyRuntime.preconfig.legacy_windows_fs_encoding == 0);
#endif
    if (use_utf8) {
        return _Py_EncodeUTF8Ex(text, str, error_pos, reason,
                                raw_malloc, errors);
    }

#ifdef USE_FORCE_ASCII
    if (force_ascii == -1) {
        force_ascii = check_force_ascii();
    }

    if (force_ascii) {
        return encode_ascii(text, str, error_pos, reason,
                            raw_malloc, errors);
    }
#endif

    return encode_current_locale(text, str, error_pos, reason,
                                 raw_malloc, errors);
#endif   /* _Py_FORCE_UTF8_FS_ENCODING */
}

static char*
encode_locale(const wchar_t *text, size_t *error_pos,
              int raw_malloc, int current_locale)
{
    char *str;
    int res = encode_locale_ex(text, &str, error_pos, NULL,
                               raw_malloc, current_locale,
                               _Py_ERROR_SURROGATEESCAPE);
    if (res != -2 && error_pos) {
        *error_pos = (size_t)-1;
    }
    if (res != 0) {
        return NULL;
    }
    return str;
}

/* Encode a wide character string to the locale encoding with the
   surrogateescape error handler: surrogate characters in the range
   U+DC80..U+DCFF are converted to bytes 0x80..0xFF.

   Return a pointer to a newly allocated byte string, use PyMem_Free() to free
   the memory. Return NULL on encoding or memory allocation error.

   If error_pos is not NULL, *error_pos is set to (size_t)-1 on success, or set
   to the index of the invalid character on encoding error.

   Use the Py_DecodeLocale() function to decode the bytes string back to a wide
   character string. */
char*
Py_EncodeLocale(const wchar_t *text, size_t *error_pos)
{
    return encode_locale(text, error_pos, 0, 0);
}


/* Similar to Py_EncodeLocale(), but result must be freed by PyMem_RawFree()
   instead of PyMem_Free(). */
char*
_Py_EncodeLocaleRaw(const wchar_t *text, size_t *error_pos)
{
    return encode_locale(text, error_pos, 1, 0);
}


int
_Py_EncodeLocaleEx(const wchar_t *text, char **str,
                   size_t *error_pos, const char **reason,
                   int current_locale, _Py_error_handler errors)
{
    return encode_locale_ex(text, str, error_pos, reason, 1,
                            current_locale, errors);
}


// Get the current locale encoding name:
//
// - Return "utf-8" if _Py_FORCE_UTF8_LOCALE macro is defined (ex: on Android)
// - Return "utf-8" if the UTF-8 Mode is enabled
// - On Windows, return the ANSI code page (ex: "cp1250")
// - Return "utf-8" if nl_langinfo(CODESET) returns an empty string.
// - Otherwise, return nl_langinfo(CODESET).
//
// Return NULL on memory allocation failure.
//
// See also config_get_locale_encoding()
wchar_t*
_Py_GetLocaleEncoding(void)
{
#ifdef _Py_FORCE_UTF8_LOCALE
    // On Android langinfo.h and CODESET are missing,
    // and UTF-8 is always used in mbstowcs() and wcstombs().
    return _PyMem_RawWcsdup(L"utf-8");
#else

#ifdef MS_WINDOWS
    wchar_t encoding[23];
    unsigned int ansi_codepage = GetACP();
    swprintf(encoding, Py_ARRAY_LENGTH(encoding), L"cp%u", ansi_codepage);
    encoding[Py_ARRAY_LENGTH(encoding) - 1] = 0;
    return _PyMem_RawWcsdup(encoding);
#else
    const char *encoding = nl_langinfo(CODESET);
    if (!encoding || encoding[0] == '\0') {
        // Use UTF-8 if nl_langinfo() returns an empty string. It can happen on
        // macOS if the LC_CTYPE locale is not supported.
        return _PyMem_RawWcsdup(L"utf-8");
    }

    wchar_t *wstr;
    int res = decode_current_locale(encoding, &wstr, NULL,
                                    NULL, _Py_ERROR_SURROGATEESCAPE);
    if (res < 0) {
        return NULL;
    }
    return wstr;
#endif  // !MS_WINDOWS

#endif  // !_Py_FORCE_UTF8_LOCALE
}


PyObject *
_Py_GetLocaleEncodingObject(void)
{
    wchar_t *encoding = _Py_GetLocaleEncoding();
    if (encoding == NULL) {
        PyErr_NoMemory();
        return NULL;
    }

    PyObject *str = PyUnicode_FromWideChar(encoding, -1);
    PyMem_RawFree(encoding);
    return str;
}

#ifdef HAVE_NON_UNICODE_WCHAR_T_REPRESENTATION

/* Check whether current locale uses Unicode as internal wchar_t form. */
int
_Py_LocaleUsesNonUnicodeWchar(void)
{
    /* Oracle Solaris uses non-Unicode internal wchar_t form for
       non-Unicode locales and hence needs conversion to UTF first. */
    char* codeset = nl_langinfo(CODESET);
    if (!codeset) {
        return 0;
    }
    /* 646 refers to ISO/IEC 646 standard that corresponds to ASCII encoding */
    return (strcmp(codeset, "UTF-8") != 0 && strcmp(codeset, "646") != 0);
}

static wchar_t *
_Py_ConvertWCharForm(const wchar_t *source, Py_ssize_t size,
                     const char *tocode, const char *fromcode)
{
    static_assert(sizeof(wchar_t) == 4, "wchar_t must be 32-bit");

    /* Ensure we won't overflow the size. */
    if (size > (PY_SSIZE_T_MAX / (Py_ssize_t)sizeof(wchar_t))) {
        PyErr_NoMemory();
        return NULL;
    }

    /* the string doesn't have to be NULL terminated */
    wchar_t* target = PyMem_Malloc(size * sizeof(wchar_t));
    if (target == NULL) {
        PyErr_NoMemory();
        return NULL;
    }

    iconv_t cd = iconv_open(tocode, fromcode);
    if (cd == (iconv_t)-1) {
        PyErr_Format(PyExc_ValueError, "iconv_open() failed");
        PyMem_Free(target);
        return NULL;
    }

    char *inbuf = (char *) source;
    char *outbuf = (char *) target;
    size_t inbytesleft = sizeof(wchar_t) * size;
    size_t outbytesleft = inbytesleft;

    size_t ret = iconv(cd, &inbuf, &inbytesleft, &outbuf, &outbytesleft);
    if (ret == DECODE_ERROR) {
        PyErr_Format(PyExc_ValueError, "iconv() failed");
        PyMem_Free(target);
        iconv_close(cd);
        return NULL;
    }

    iconv_close(cd);
    return target;
}

/* Convert a wide character string to the UCS-4 encoded string. This
   is necessary on systems where internal form of wchar_t are not Unicode
   code points (e.g. Oracle Solaris).

   Return a pointer to a newly allocated string, use PyMem_Free() to free
   the memory. Return NULL and raise exception on conversion or memory
   allocation error. */
wchar_t *
_Py_DecodeNonUnicodeWchar(const wchar_t *native, Py_ssize_t size)
{
    return _Py_ConvertWCharForm(native, size, "UCS-4-INTERNAL", "wchar_t");
}

/* Convert a UCS-4 encoded string to native wide character string. This
   is necessary on systems where internal form of wchar_t are not Unicode
   code points (e.g. Oracle Solaris).

   The conversion is done in place. This can be done because both wchar_t
   and UCS-4 use 4-byte encoding, and one wchar_t symbol always correspond
   to a single UCS-4 symbol and vice versa. (This is true for Oracle Solaris,
   which is currently the only system using these functions; it doesn't have
   to be for other systems).

   Return 0 on success. Return -1 and raise exception on conversion
   or memory allocation error. */
int
_Py_EncodeNonUnicodeWchar_InPlace(wchar_t *unicode, Py_ssize_t size)
{
    wchar_t* result = _Py_ConvertWCharForm(unicode, size, "wchar_t", "UCS-4-INTERNAL");
    if (!result) {
        return -1;
    }
    memcpy(unicode, result, size * sizeof(wchar_t));
    PyMem_Free(result);
    return 0;
}
#endif /* HAVE_NON_UNICODE_WCHAR_T_REPRESENTATION */

#ifdef MS_WINDOWS
static __int64 secs_between_epochs = 11644473600; /* Seconds between 1.1.1601 and 1.1.1970 */

static void
FILE_TIME_to_time_t_nsec(FILETIME *in_ptr, time_t *time_out, int* nsec_out)
{
    /* XXX endianness. Shouldn't matter, as all Windows implementations are little-endian */
    /* Cannot simply cast and dereference in_ptr,
       since it might not be aligned properly */
    __int64 in;
    memcpy(&in, in_ptr, sizeof(in));
    *nsec_out = (int)(in % 10000000) * 100; /* FILETIME is in units of 100 nsec. */
    *time_out = Py_SAFE_DOWNCAST((in / 10000000) - secs_between_epochs, __int64, time_t);
}

static void
LARGE_INTEGER_to_time_t_nsec(LARGE_INTEGER *in_ptr, time_t *time_out, int* nsec_out)
{
    *nsec_out = (int)(in_ptr->QuadPart % 10000000) * 100; /* FILETIME is in units of 100 nsec. */
    *time_out = Py_SAFE_DOWNCAST((in_ptr->QuadPart / 10000000) - secs_between_epochs, __int64, time_t);
}

void
_Py_time_t_to_FILE_TIME(time_t time_in, int nsec_in, FILETIME *out_ptr)
{
    /* XXX endianness */
    __int64 out;
    out = time_in + secs_between_epochs;
    out = out * 10000000 + nsec_in / 100;
    memcpy(out_ptr, &out, sizeof(out));
}

/* Below, we *know* that ugo+r is 0444 */
#if _S_IREAD != 0400
#error Unsupported C library
#endif
static int
attributes_to_mode(DWORD attr)
{
    int m = 0;
    if (attr & FILE_ATTRIBUTE_DIRECTORY)
        m |= _S_IFDIR | 0111; /* IFEXEC for user,group,other */
    else
        m |= _S_IFREG;
    if (attr & FILE_ATTRIBUTE_READONLY)
        m |= 0444;
    else
        m |= 0666;
    return m;
}


typedef union {
    FILE_ID_128 id;
    struct {
        uint64_t st_ino;
        uint64_t st_ino_high;
    };
} id_128_to_ino;


void
_Py_attribute_data_to_stat(BY_HANDLE_FILE_INFORMATION *info, ULONG reparse_tag,
                           FILE_BASIC_INFO *basic_info, FILE_ID_INFO *id_info,
                           struct _Py_stat_struct *result)
{
    memset(result, 0, sizeof(*result));
    result->st_mode = attributes_to_mode(info->dwFileAttributes);
    result->st_size = (((__int64)info->nFileSizeHigh)<<32) + info->nFileSizeLow;
    result->st_dev = id_info ? id_info->VolumeSerialNumber : info->dwVolumeSerialNumber;
    result->st_rdev = 0;
    /* st_ctime is deprecated, but we preserve the legacy value in our caller, not here */
    if (basic_info) {
        LARGE_INTEGER_to_time_t_nsec(&basic_info->CreationTime, &result->st_birthtime, &result->st_birthtime_nsec);
        LARGE_INTEGER_to_time_t_nsec(&basic_info->ChangeTime, &result->st_ctime, &result->st_ctime_nsec);
        LARGE_INTEGER_to_time_t_nsec(&basic_info->LastWriteTime, &result->st_mtime, &result->st_mtime_nsec);
        LARGE_INTEGER_to_time_t_nsec(&basic_info->LastAccessTime, &result->st_atime, &result->st_atime_nsec);
    } else {
        FILE_TIME_to_time_t_nsec(&info->ftCreationTime, &result->st_birthtime, &result->st_birthtime_nsec);
        FILE_TIME_to_time_t_nsec(&info->ftLastWriteTime, &result->st_mtime, &result->st_mtime_nsec);
        FILE_TIME_to_time_t_nsec(&info->ftLastAccessTime, &result->st_atime, &result->st_atime_nsec);
    }
    result->st_nlink = info->nNumberOfLinks;

    if (id_info) {
        id_128_to_ino file_id;
        file_id.id = id_info->FileId;
        result->st_ino = file_id.st_ino;
        result->st_ino_high = file_id.st_ino_high;
    }
    if (!result->st_ino && !result->st_ino_high) {
        /* should only occur for DirEntry_from_find_data, in which case the
           index is likely to be zero anyway. */
        result->st_ino = (((uint64_t)info->nFileIndexHigh) << 32) + info->nFileIndexLow;
    }

    /* bpo-37834: Only actual symlinks set the S_IFLNK flag. But lstat() will
       open other name surrogate reparse points without traversing them. To
       detect/handle these, check st_file_attributes and st_reparse_tag. */
    result->st_reparse_tag = reparse_tag;
    if (info->dwFileAttributes & FILE_ATTRIBUTE_REPARSE_POINT &&
        reparse_tag == IO_REPARSE_TAG_SYMLINK) {
        /* set the bits that make this a symlink */
        result->st_mode = (result->st_mode & ~S_IFMT) | S_IFLNK;
    }
    result->st_file_attributes = info->dwFileAttributes;
}

void
_Py_stat_basic_info_to_stat(FILE_STAT_BASIC_INFORMATION *info,
                            struct _Py_stat_struct *result)
{
    memset(result, 0, sizeof(*result));
    result->st_mode = attributes_to_mode(info->FileAttributes);
    result->st_size = info->EndOfFile.QuadPart;
    LARGE_INTEGER_to_time_t_nsec(&info->CreationTime, &result->st_birthtime, &result->st_birthtime_nsec);
    LARGE_INTEGER_to_time_t_nsec(&info->ChangeTime, &result->st_ctime, &result->st_ctime_nsec);
    LARGE_INTEGER_to_time_t_nsec(&info->LastWriteTime, &result->st_mtime, &result->st_mtime_nsec);
    LARGE_INTEGER_to_time_t_nsec(&info->LastAccessTime, &result->st_atime, &result->st_atime_nsec);
    result->st_nlink = info->NumberOfLinks;
    result->st_dev = info->VolumeSerialNumber.QuadPart;
    /* File systems with less than 128-bits zero pad into this field */
    id_128_to_ino file_id;
    file_id.id = info->FileId128;
    result->st_ino = file_id.st_ino;
    result->st_ino_high = file_id.st_ino_high;
    /* bpo-37834: Only actual symlinks set the S_IFLNK flag. But lstat() will
       open other name surrogate reparse points without traversing them. To
       detect/handle these, check st_file_attributes and st_reparse_tag. */
    result->st_reparse_tag = info->ReparseTag;
    if (info->FileAttributes & FILE_ATTRIBUTE_REPARSE_POINT &&
        info->ReparseTag == IO_REPARSE_TAG_SYMLINK) {
        /* set the bits that make this a symlink */
        result->st_mode = (result->st_mode & ~S_IFMT) | S_IFLNK;
    }
    result->st_file_attributes = info->FileAttributes;
    switch (info->DeviceType) {
    case FILE_DEVICE_DISK:
    case FILE_DEVICE_VIRTUAL_DISK:
    case FILE_DEVICE_DFS:
    case FILE_DEVICE_CD_ROM:
    case FILE_DEVICE_CONTROLLER:
    case FILE_DEVICE_DATALINK:
        break;
    case FILE_DEVICE_DISK_FILE_SYSTEM:
    case FILE_DEVICE_CD_ROM_FILE_SYSTEM:
    case FILE_DEVICE_NETWORK_FILE_SYSTEM:
        result->st_mode = (result->st_mode & ~S_IFMT) | 0x6000; /* _S_IFBLK */
        break;
    case FILE_DEVICE_CONSOLE:
    case FILE_DEVICE_NULL:
    case FILE_DEVICE_KEYBOARD:
    case FILE_DEVICE_MODEM:
    case FILE_DEVICE_MOUSE:
    case FILE_DEVICE_PARALLEL_PORT:
    case FILE_DEVICE_PRINTER:
    case FILE_DEVICE_SCREEN:
    case FILE_DEVICE_SERIAL_PORT:
    case FILE_DEVICE_SOUND:
        result->st_mode = (result->st_mode & ~S_IFMT) | _S_IFCHR;
        break;
    case FILE_DEVICE_NAMED_PIPE:
        result->st_mode = (result->st_mode & ~S_IFMT) | _S_IFIFO;
        break;
    default:
        if (info->FileAttributes & FILE_ATTRIBUTE_DIRECTORY) {
            result->st_mode = (result->st_mode & ~S_IFMT) | _S_IFDIR;
        }
        break;
    }
}

#endif

/* Return information about a file.

   On POSIX, use fstat().

   On Windows, use GetFileType() and GetFileInformationByHandle() which support
   files larger than 2 GiB.  fstat() may fail with EOVERFLOW on files larger
   than 2 GiB because the file size type is a signed 32-bit integer: see issue
   #23152.

   On Windows, set the last Windows error and return nonzero on error. On
   POSIX, set errno and return nonzero on error. Fill status and return 0 on
   success. */
int
_Py_fstat_noraise(int fd, struct _Py_stat_struct *status)
{
#ifdef MS_WINDOWS
    BY_HANDLE_FILE_INFORMATION info;
    FILE_BASIC_INFO basicInfo;
    FILE_ID_INFO idInfo;
    FILE_ID_INFO *pIdInfo = &idInfo;
    HANDLE h;
    int type;

    h = _Py_get_osfhandle_noraise(fd);

    if (h == INVALID_HANDLE_VALUE) {
        /* errno is already set by _get_osfhandle, but we also set
           the Win32 error for callers who expect that */
        SetLastError(ERROR_INVALID_HANDLE);
        return -1;
    }
    memset(status, 0, sizeof(*status));

    type = GetFileType(h);
    if (type == FILE_TYPE_UNKNOWN) {
        DWORD error = GetLastError();
        if (error != 0) {
            errno = winerror_to_errno(error);
            return -1;
        }
        /* else: valid but unknown file */
    }

    if (type != FILE_TYPE_DISK) {
        if (type == FILE_TYPE_CHAR)
            status->st_mode = _S_IFCHR;
        else if (type == FILE_TYPE_PIPE)
            status->st_mode = _S_IFIFO;
        return 0;
    }

    if (!GetFileInformationByHandle(h, &info) ||
        !GetFileInformationByHandleEx(h, FileBasicInfo, &basicInfo, sizeof(basicInfo))) {
        /* The Win32 error is already set, but we also set errno for
           callers who expect it */
        errno = winerror_to_errno(GetLastError());
        return -1;
    }

    if (!GetFileInformationByHandleEx(h, FileIdInfo, &idInfo, sizeof(idInfo))) {
        /* Failed to get FileIdInfo, so do not pass it along */
        pIdInfo = NULL;
    }

    _Py_attribute_data_to_stat(&info, 0, &basicInfo, pIdInfo, status);
    return 0;
#else
    return fstat(fd, status);
#endif
}

/* Return information about a file.

   On POSIX, use fstat().

   On Windows, use GetFileType() and GetFileInformationByHandle() which support
   files larger than 2 GiB.  fstat() may fail with EOVERFLOW on files larger
   than 2 GiB because the file size type is a signed 32-bit integer: see issue
   #23152.

   Raise an exception and return -1 on error. On Windows, set the last Windows
   error on error. On POSIX, set errno on error. Fill status and return 0 on
   success.

   Release the GIL to call GetFileType() and GetFileInformationByHandle(), or
   to call fstat(). The caller must hold the GIL. */
int
_Py_fstat(int fd, struct _Py_stat_struct *status)
{
    int res;

    _Py_AssertHoldsTstate();

    Py_BEGIN_ALLOW_THREADS
    res = _Py_fstat_noraise(fd, status);
    Py_END_ALLOW_THREADS

    if (res != 0) {
#ifdef MS_WINDOWS
        PyErr_SetFromWindowsErr(0);
#else
        PyErr_SetFromErrno(PyExc_OSError);
#endif
        return -1;
    }
    return 0;
}

/* Like _Py_stat() but with a raw filename. */
int
_Py_wstat(const wchar_t* path, struct stat *buf)
{
    int err;
#ifdef MS_WINDOWS
    struct _stat wstatbuf;
    err = _wstat(path, &wstatbuf);
    if (!err) {
        buf->st_mode = wstatbuf.st_mode;
    }
#else
    char *fname;
    fname = _Py_EncodeLocaleRaw(path, NULL);
    if (fname == NULL) {
        errno = EINVAL;
        return -1;
    }
    err = stat(fname, buf);
    PyMem_RawFree(fname);
#endif
    return err;
}


/* Call _wstat() on Windows, or encode the path to the filesystem encoding and
   call stat() otherwise. Only fill st_mode attribute on Windows.

   Return 0 on success, -1 on _wstat() / stat() error, -2 if an exception was
   raised. */

int
_Py_stat(PyObject *path, struct stat *statbuf)
{
#ifdef MS_WINDOWS
    int err;

    wchar_t *wpath = PyUnicode_AsWideCharString(path, NULL);
    if (wpath == NULL)
        return -2;

    err = _Py_wstat(wpath, statbuf);
    PyMem_Free(wpath);
    return err;
#else
    int ret;
    PyObject *bytes;
    char *cpath;

    bytes = PyUnicode_EncodeFSDefault(path);
    if (bytes == NULL)
        return -2;

    /* check for embedded null bytes */
    if (PyBytes_AsStringAndSize(bytes, &cpath, NULL) == -1) {
        Py_DECREF(bytes);
        return -2;
    }

    ret = stat(cpath, statbuf);
    Py_DECREF(bytes);
    return ret;
#endif
}

#ifdef MS_WINDOWS
// For some Windows API partitions, SetHandleInformation() is declared
// but none of the handle flags are defined.
#ifndef HANDLE_FLAG_INHERIT
#define HANDLE_FLAG_INHERIT 0x00000001
#endif
#endif

/* This function MUST be kept async-signal-safe on POSIX when raise=0. */
static int
get_inheritable(int fd, int raise)
{
#ifdef MS_WINDOWS
    HANDLE handle;
    DWORD flags;

    handle = _Py_get_osfhandle_noraise(fd);
    if (handle == INVALID_HANDLE_VALUE) {
        if (raise)
            PyErr_SetFromErrno(PyExc_OSError);
        return -1;
    }

    if (!GetHandleInformation(handle, &flags)) {
        if (raise)
            PyErr_SetFromWindowsErr(0);
        return -1;
    }

    return (flags & HANDLE_FLAG_INHERIT);
#else
    int flags;

    flags = fcntl(fd, F_GETFD, 0);
    if (flags == -1) {
        if (raise)
            PyErr_SetFromErrno(PyExc_OSError);
        return -1;
    }
    return !(flags & FD_CLOEXEC);
#endif
}

/* Get the inheritable flag of the specified file descriptor.
   Return 1 if the file descriptor can be inherited, 0 if it cannot,
   raise an exception and return -1 on error. */
int
_Py_get_inheritable(int fd)
{
    return get_inheritable(fd, 1);
}


/* This function MUST be kept async-signal-safe on POSIX when raise=0. */
static int
set_inheritable(int fd, int inheritable, int raise, int *atomic_flag_works)
{
#ifdef MS_WINDOWS
    HANDLE handle;
    DWORD flags;
#else
#if defined(HAVE_SYS_IOCTL_H) && defined(FIOCLEX) && defined(FIONCLEX)
    static int ioctl_works = -1;
    int request;
    int err;
#endif
    int flags, new_flags;
    int res;
#endif

    /* atomic_flag_works can only be used to make the file descriptor
       non-inheritable */
    assert(!(atomic_flag_works != NULL && inheritable));

    if (atomic_flag_works != NULL && !inheritable) {
        if (_Py_atomic_load_int_relaxed(atomic_flag_works) == -1) {
            int isInheritable = get_inheritable(fd, raise);
            if (isInheritable == -1)
                return -1;
            _Py_atomic_store_int_relaxed(atomic_flag_works, !isInheritable);
        }

        if (_Py_atomic_load_int_relaxed(atomic_flag_works))
            return 0;
    }

#ifdef MS_WINDOWS
    handle = _Py_get_osfhandle_noraise(fd);
    if (handle == INVALID_HANDLE_VALUE) {
        if (raise)
            PyErr_SetFromErrno(PyExc_OSError);
        return -1;
    }

    if (inheritable)
        flags = HANDLE_FLAG_INHERIT;
    else
        flags = 0;

    if (!SetHandleInformation(handle, HANDLE_FLAG_INHERIT, flags)) {
        if (raise)
            PyErr_SetFromWindowsErr(0);
        return -1;
    }
    return 0;

#else

#if defined(HAVE_SYS_IOCTL_H) && defined(FIOCLEX) && defined(FIONCLEX)
    if (raise != 0 && _Py_atomic_load_int_relaxed(&ioctl_works) != 0) {
        /* fast-path: ioctl() only requires one syscall */
        /* caveat: raise=0 is an indicator that we must be async-signal-safe
         * thus avoid using ioctl() so we skip the fast-path. */
        if (inheritable)
            request = FIONCLEX;
        else
            request = FIOCLEX;
        err = ioctl(fd, request, NULL);
        if (!err) {
            if (_Py_atomic_load_int_relaxed(&ioctl_works) == -1) {
                _Py_atomic_store_int_relaxed(&ioctl_works, 1);
            }
            return 0;
        }

#ifdef O_PATH
        if (errno == EBADF) {
            // bpo-44849: On Linux and FreeBSD, ioctl(FIOCLEX) fails with EBADF
            // on O_PATH file descriptors. Fall through to the fcntl()
            // implementation.
        }
        else
#endif
        if (errno != ENOTTY && errno != EACCES) {
            if (raise)
                PyErr_SetFromErrno(PyExc_OSError);
            return -1;
        }
        else {
            /* Issue #22258: Here, ENOTTY means "Inappropriate ioctl for
               device". The ioctl is declared but not supported by the kernel.
               Remember that ioctl() doesn't work. It is the case on
               Illumos-based OS for example.

               Issue #27057: When SELinux policy disallows ioctl it will fail
               with EACCES. While FIOCLEX is safe operation it may be
               unavailable because ioctl was denied altogether.
               This can be the case on Android. */
            _Py_atomic_store_int_relaxed(&ioctl_works, 0);
        }
        /* fallback to fcntl() if ioctl() does not work */
    }
#endif

    /* slow-path: fcntl() requires two syscalls */
    flags = fcntl(fd, F_GETFD);
    if (flags < 0) {
        if (raise)
            PyErr_SetFromErrno(PyExc_OSError);
        return -1;
    }

    if (inheritable) {
        new_flags = flags & ~FD_CLOEXEC;
    }
    else {
        new_flags = flags | FD_CLOEXEC;
    }

    if (new_flags == flags) {
        /* FD_CLOEXEC flag already set/cleared: nothing to do */
        return 0;
    }

    res = fcntl(fd, F_SETFD, new_flags);
    if (res < 0) {
        if (raise)
            PyErr_SetFromErrno(PyExc_OSError);
        return -1;
    }
    return 0;
#endif
}

/* Make the file descriptor non-inheritable.
   Return 0 on success, set errno and return -1 on error. */
static int
make_non_inheritable(int fd)
{
    return set_inheritable(fd, 0, 0, NULL);
}

/* Set the inheritable flag of the specified file descriptor.
   On success: return 0, on error: raise an exception and return -1.

   If atomic_flag_works is not NULL:

    * if *atomic_flag_works==-1, check if the inheritable is set on the file
      descriptor: if yes, set *atomic_flag_works to 1, otherwise set to 0 and
      set the inheritable flag
    * if *atomic_flag_works==1: do nothing
    * if *atomic_flag_works==0: set inheritable flag to False

   Set atomic_flag_works to NULL if no atomic flag was used to create the
   file descriptor.

   atomic_flag_works can only be used to make a file descriptor
   non-inheritable: atomic_flag_works must be NULL if inheritable=1. */
int
_Py_set_inheritable(int fd, int inheritable, int *atomic_flag_works)
{
    return set_inheritable(fd, inheritable, 1, atomic_flag_works);
}

/* Same as _Py_set_inheritable() but on error, set errno and
   don't raise an exception.
   This function is async-signal-safe. */
int
_Py_set_inheritable_async_safe(int fd, int inheritable, int *atomic_flag_works)
{
    return set_inheritable(fd, inheritable, 0, atomic_flag_works);
}

static int
_Py_open_impl(const char *pathname, int flags, int gil_held)
{
    int fd;
    int async_err = 0;
#ifndef MS_WINDOWS
    int *atomic_flag_works;
#endif

#ifdef MS_WINDOWS
    flags |= O_NOINHERIT;
#elif defined(O_CLOEXEC)
    atomic_flag_works = &_Py_open_cloexec_works;
    flags |= O_CLOEXEC;
#else
    atomic_flag_works = NULL;
#endif

    if (gil_held) {
        PyObject *pathname_obj = PyUnicode_DecodeFSDefault(pathname);
        if (pathname_obj == NULL) {
            return -1;
        }
        if (PySys_Audit("open", "OOi", pathname_obj, Py_None, flags) < 0) {
            Py_DECREF(pathname_obj);
            return -1;
        }

        do {
            Py_BEGIN_ALLOW_THREADS
            fd = open(pathname, flags);
            Py_END_ALLOW_THREADS
        } while (fd < 0
                 && errno == EINTR && !(async_err = PyErr_CheckSignals()));
        if (async_err) {
            Py_DECREF(pathname_obj);
            return -1;
        }
        if (fd < 0) {
            PyErr_SetFromErrnoWithFilenameObjects(PyExc_OSError, pathname_obj, NULL);
            Py_DECREF(pathname_obj);
            return -1;
        }
        Py_DECREF(pathname_obj);
    }
    else {
        fd = open(pathname, flags);
        if (fd < 0)
            return -1;