Move private functions to the internal C API (pycore_dict.h):
* _PyDictView_Intersect()
* _PyDictView_New()
* _PyDict_ContainsId()
* _PyDict_DelItemId()
* _PyDict_DelItem_KnownHash()
* _PyDict_GetItemIdWithError()
* _PyDict_GetItem_KnownHash()
* _PyDict_HasSplitTable()
* _PyDict_NewPresized()
* _PyDict_Next()
* _PyDict_Pop()
* _PyDict_SetItemId()
* _PyDict_SetItem_KnownHash()
* _PyDict_SizeOf()
No longer export most of these functions.
Move also the _PyDictViewObject structure to the internal C API.
Move dict_getitem_knownhash() function from _testcapi to the
_testinternalcapi extension. Update test_capi.test_dict for this
change.
Replace _PyDict_GetItemStringWithError() calls with
PyDict_GetItemStringRef() which returns a strong reference to the
item.
Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>
Remove private _PyDict_GetItemStringWithError() function of the
public C API: the new PyDict_GetItemStringRef() can be used instead.
* Move private _PyDict_GetItemStringWithError() to the internal C API.
* _testcapi get_code_extra_index() uses PyDict_GetItemStringRef().
Avoid using private functions in _testcapi which tests the public C
API.
gh-107184 introduced a refleak in test_import.SubinterpImportTests (specifically test_singlephase_check_with_setting_and_override and test_single_init_extension_compat). We fix it here by making sure _testsinglephase is removed from sys.modules whenever we clear the runtime's internal state for the module.
The underlying problem is strictly contained in the internal function _PyImport_ClearExtension() (AKA _testinternalcapi.clear_extension()), which is only used in tests.
(This also fixes an intermittent segfault introduced in the same place, in test_disallowed_reimport.)
This fixes a crasher due to a race condition, triggered infrequently when two isolated (own GIL) subinterpreters simultaneously initialize their sys or builtins modules. The crash happened due the combination of the "detached" thread state we were using and the "last holder" logic we use for the GIL. It turns out it's tricky to use the same thread state for different threads. Who could have guessed?
We solve the problem by eliminating the one object we were still sharing between interpreters. We replace it with a low-level hashtable, using the "raw" allocator to avoid tying it to the main interpreter.
We also remove the accommodations for "detached" thread states, which were a dubious idea to start with.
Remove old aliases which were kept backwards compatibility with
Python 3.8:
* _PyObject_CallMethodNoArgs()
* _PyObject_CallMethodOneArg()
* _PyObject_CallOneArg()
* _PyObject_FastCallDict()
* _PyObject_Vectorcall()
* _PyObject_VectorcallMethod()
* _PyVectorcall_Function()
Update code which used these aliases to use new names.
It now raises an exception if sys.modules doesn't hold a strong
reference to the module.
Elaborate the comment explaining why a weak reference is used to
create a borrowed reference.
* Add tests on PyImport_AddModuleRef(), PyImport_AddModule() and
PyImport_AddModuleObject().
* pythonrun.c: Replace Py_XNewRef(PyImport_AddModule(name)) with
PyImport_AddModuleRef(name).
This fixes a race during import. The existing _PyRuntimeState.imports.pkgcontext is shared between interpreters, and occasionally this would cause a crash when multiple interpreters were importing extensions modules at the same time. To solve this we add a thread-local variable for the value. We also leave the existing state (and infrequent race) in place for platforms that do not support thread-local variables.
Here we are doing no more than adding the value for Py_mod_multiple_interpreters and using it for stdlib modules. We will start checking for it in gh-104206 (once PyInterpreterState.ceval.own_gil is added in gh-104204).
Deep-frozen code objects are cannot be shared (currently) by
interpreters, due to how adaptive specialization can modify the
bytecodes. We work around this by only using the deep-frozen objects in
the main interpreter. This does incur a performance penalty for
subinterpreters, which we may be able to resolve later.
In gh-102744 we added is_core_module() (in Python/import.c), which relies on get_core_module_dict() (also added in that PR). The problem is that_PyImport_FixupBuiltin(), which ultimately calls is_core_module(), is called on the builtins module before interp->builtins_copyis set. Consequently, the builtins module isn't considered a "core" module while it is getting "fixed up" and its module def m_copy erroneously gets set. Under isolated interpreters this causes problems since sys and builtins are allowed even though they are still single-phase init modules. (This was discovered while working on gh-101660.)
The solution is to stop relying on get_core_module_dict() in is_core_module().
Sharing mutable (or non-immortal) objects between interpreters is generally not safe. We can work around that but not easily.
There are two restrictions that are critical for objects that break interpreter isolation.
The first is that the object's state be guarded by a global lock. For now the GIL meets this requirement, but a granular global lock is needed once we have a per-interpreter GIL.
The second restriction is that the object (and, for a container, its items) be deallocated/resized only when the interpreter in which it was allocated is the current one. This is because every interpreter has (or will have, see gh-101660) its own object allocator. Deallocating an object with a different allocator can cause crashes.
The dict for the cache of module defs is completely internal, which simplifies what we have to do to meet those requirements. To do so, we do the following:
* add a mechanism for re-using a temporary thread state tied to the main interpreter in an arbitrary thread
* add _PyRuntime.imports.extensions.main_tstate`
* add _PyThreadState_InitDetached() and _PyThreadState_ClearDetached() (pystate.c)
* add _PyThreadState_BindDetached() and _PyThreadState_UnbindDetached() (pystate.c)
* make sure the cache dict (_PyRuntime.imports.extensions.dict) and its items are all owned by the main interpreter)
* add a placeholder using for a granular global lock
Note that the cache is only used for legacy extension modules and not for multi-phase init modules.
https://github.com/python/cpython/issues/100227
It doesn't make sense to use multi-phase init for these modules. Using a per-interpreter "m_copy" (instead of PyModuleDef.m_base.m_copy) makes this work okay. (This came up while working on gh-101660.)
Note that we might instead end up disallowing re-load for sys/builtins since they are so special.
https://github.com/python/cpython/issues/102660
Specific changes:
* move the import lock to PyInterpreterState
* move the "find_and_load" diagnostic state to PyInterpreterState
Note that the import lock exists to keep multiple imports of the same module in the same interpreter (but in different threads) from stomping on each other. Independently, we use a distinct global lock to protect globally shared import state, especially related to loaded extension modules. For now we can rely on the GIL as that lock but with a per-interpreter GIL we'll need a new global lock.
The remaining state in _PyRuntimeState.imports will (probably) continue being global.
https://github.com/python/cpython/issues/100227
This is related to fixing the refleaks introduced by commit 096d009. I haven't been able to find the leak yet, but these changes are a consequence of that effort. This includes some cleanup, some tweaks to the existing tests, and a bunch of new test cases. The only change here that might have impact outside the tests in question is in imp.py, where I update imp.load_dynamic() to use spec_from_file_location() instead of creating a ModuleSpec directly.
Also note that I've updated the tests to only skip if we're checking for refleaks (regrtest's --huntrleaks), whereas in gh-101969 I had skipped the tests entirely. The tests will be useful for some upcoming work and I'd rather the refleaks not hold that up. (It isn't clear how quickly we'll be able to fix the leaking code, though it will certainly be done in the short term.)
https://github.com/python/cpython/issues/102251
We're adding the function back, only for the stable ABI symbol and not as any form of API. I had removed it yesterday.
This undocumented "private" function was added with the implementation for PEP 3121 (3.0, 2007) for internal use and later moved out of the limited API (3.6, 2016) and then into the internal API (3.9, 2019). I removed it completely yesterday, including from the stable ABI manifest (where it was added because the symbol happened to be exported). It's unlikely that anyone is using _PyState_AddModule(), especially any stable ABI extensions built against 3.2-3.5, but we're playing it safe.
https://github.com/python/cpython/issues/101758
Enforcing (optionally) the restriction set by PEP 489 makes sense. Furthermore, this sets the stage for a potential restriction related to a per-interpreter GIL.
This change includes the following:
* add tests for extension module subinterpreter compatibility
* add _PyInterpreterConfig.check_multi_interp_extensions
* add Py_RTFLAGS_MULTI_INTERP_EXTENSIONS
* add _PyImport_CheckSubinterpIncompatibleExtensionAllowed()
* fail iff the module does not implement multi-phase init and the current interpreter is configured to check
https://github.com/python/cpython/issues/98627
This change is almost entirely moving code around and hiding import state behind internal API. We introduce no changes to behavior, nor to non-internal API. (Since there was already going to be a lot of churn, I took this as an opportunity to re-organize import.c into topically-grouped sections of code.) The motivation is to simplify a number of upcoming changes.
Specific changes:
* move existing import-related code to import.c, wherever possible
* add internal API for interacting with import state (both global and per-interpreter)
* use only API outside of import.c (to limit churn there when changing the location, etc.)
* consolidate the import-related state of PyInterpreterState into a single struct field (this changes layout slightly)
* add macros for import state in import.c (to simplify changing the location)
* group code in import.c into sections
*remove _PyState_AddModule()
https://github.com/python/cpython/issues/101758
The new test exercises the most important variants for single-phase init extension modules. We also add some explanation about those variants to import.c.
https://github.com/python/cpython/issues/101758
* Make sure that the current exception is always normalized.
* Remove redundant type and traceback fields for the current exception.
* Add new API functions: PyErr_GetRaisedException, PyErr_SetRaisedException
* Add new API functions: PyException_GetArgs, PyException_SetArgs
We actually don't move PyImport_Inittab. Instead, we make a copy that we keep on _PyRuntimeState and use only that after Py_Initialize(). We also prevent folks from modifying PyImport_Inittab (the best we can) after that point.
https://github.com/python/cpython/issues/81057
Replace Py_INCREF() and Py_XINCREF() with Py_NewRef() and
Py_XNewRef() in C files of the Python/ directory.
Update Parser/asdl_c.py to regenerate Python/Python-ast.c.
It combines PyImport_ImportModule() and PyObject_GetAttrString()
and saves 4-6 lines of code on every use.
Add also _PyImport_GetModuleAttr() which takes Python strings as arguments.
We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code. It is still used in a number of non-builtin stdlib modules.
The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime. A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings).
https://bugs.python.org/issue46541#msg411799 explains the rationale for this change.
The core of the change is in:
* (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros
* Include/internal/pycore_runtime_init.h - added the static initializers for the global strings
* Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState
* Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers
I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings. That check is added to the PR CI config.
The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()). This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *.
The following are not changed (yet):
* stop using _Py_IDENTIFIER() in the stdlib modules
* (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API
* (maybe) intern the strings during runtime init
https://bugs.python.org/issue46541
This reduces the size of the data segment by **300 KB** of the executable because if the modules are deep-frozen then the marshalled frozen data just wastes space. This was inspired by comment by @gvanrossum in https://github.com/python/cpython/pull/29118#issuecomment-958521863. Note: There is a new option `--deepfreeze-only` in `freeze_modules.py` to change this behavior, it is on be default to save disk space.
```console
# du -s ./python before
27892 ./python
# du -s ./python after
27524 ./python
```
Automerge-Triggered-By: GH:ericsnowcurrently
This gains 10% or more in startup time for `python -c pass` on UNIX-ish systems.
The Makefile.pre.in generating code builds on Eric's work for bpo-45020, but the .c file generator is new.
Windows version TBD.
Currently custom modules (the array set on PyImport_FrozenModules) replace all the frozen stdlib modules. That can be problematic and is unlikely to be what the user wants. This change treats the custom frozen modules as additions instead. They take precedence over all other frozen modules except for those needed to bootstrap the import system. If the "code" field of an entry in the custom array is NULL then that frozen module is treated as disabled, which allows a custom entry to disable a frozen stdlib module.
This change allows us to get rid of is_essential_frozen_module() and simplifies the logic for which frozen modules should be ignored.
https://bugs.python.org/issue45395
Rename Include/namespaceobject.h to
Include/internal/pycore_namespace.h.
The _testmultiphase extension is now built with the
Py_BUILD_CORE_MODULE macro defined to access _PyNamespace_Type.
object.c: remove unused "pycore_context.h" include.
Currently frozen modules do not have __file__ set. In their spec, origin is set to "frozen" and they are marked as not having a location. (Similarly, for frozen packages __path__ is set to an empty list.) However, for frozen stdlib modules we are able to extrapolate __file__ as long as we can determine the stdlib directory at runtime. (We now do so since gh-28586.) Having __file__ set is helpful for a number of reasons. Likewise, having a non-empty __path__ means we can import submodules of a frozen package from the filesystem (e.g. we could partially freeze the encodings module).
This change sets __file__ (and adds to __path__) for frozen stdlib modules. It uses sys._stdlibdir (from gh-28586) and the frozen module alias information (from gh-28655). All that work is done in FrozenImporter (in Lib/importlib/_bootstrap.py).
Also, if a frozen module is imported before importlib is bootstrapped (during interpreter initialization) then we fix up that module and its spec during the importlib bootstrapping step (i.e. imporlib._bootstrap._setup()) to match what gets set by FrozenImporter, including setting the file info (if the stdlib dir is known). To facilitate this, modules imported using PyImport_ImportFrozenModule() have __origname__ set using the frozen module alias info. __origname__ is popped off during importlib bootstrap.
(To be clear, even with this change the new code to set __file__ during fixups in imporlib._bootstrap._setup() doesn't actually get triggered yet. This is because sys._stdlibdir hasn't been set yet in interpreter initialization at the point importlib is bootstrapped. However, we do fix up such modules at that point to otherwise match the result of importing through FrozenImporter, just not the __file__ and __path__ parts. Doing so will require changes in the order in which things happen during interpreter initialization. That can be addressed separately. Once it is, the file-related fixup code from this PR will kick in.)
Here are things this change does not do:
* set __file__ for non-stdlib modules (no way of knowing the parent dir)
* set __file__ if the stdlib dir is not known (nor assume the expense of finding it)
* relatedly, set __file__ if the stdlib is in a zip file
* verify that the filename set to __file__ actually exists (too expensive)
* update __path__ for frozen packages that alias a non-package (since there is no package dir)
Other things this change skips, but we may do later:
* set __file__ on modules imported using PyImport_ImportFrozenModule()
* set co_filename when we unmarshal the frozen code object while importing the module (e.g. in FrozenImporter.exec_module()) -- this would allow tracebacks to show source lines
* implement FrozenImporter.get_filename() and FrozenImporter.get_source()
https://bugs.python.org/issue21736
In the list of generated frozen modules at the top of Tools/scripts/freeze_modules.py, you will find that some of the modules have a different name than the module (or .py file) that is actually frozen. Let's call each case an "alias". Aliases do not come into play until we get to the (generated) list of modules in Python/frozen.c. (The tool for freezing modules, Programs/_freeze_module, is only concerned with the source file, not the module it will be used for.)
Knowledge of which frozen modules are aliases (and the identity of the original module) normally isn't important. However, this information is valuable when we go to set __file__ on frozen stdlib modules. This change updates Tools/scripts/freeze_modules.py to map aliases to the original module name (or None if not a stdlib module) in Python/frozen.c. We also add a helper function in Python/import.c to look up a frozen module's alias and add the result of that function to the frozen info returned from find_frozen().
https://bugs.python.org/issue45020
Before this change we end up duplicating effort and throwing away data in FrozenImporter.find_spec(). Now we do the work once in find_spec() and the only thing we do in FrozenImporter.exec_module() is turn the raw frozen data into a code object and then exec it.
We've added _imp.find_frozen(), add an arg to _imp.get_frozen_object(), and updated FrozenImporter. We've also moved some code around to reduce duplication, get a little more consistency in outcomes, and be more efficient.
Note that this change is mostly necessary if we want to set __file__ on frozen stdlib modules. (See https://bugs.python.org/issue21736.)
https://bugs.python.org/issue45324
Currently we freeze several modules into the runtime. For each of these modules it is essential to bootstrapping the runtime that they be frozen. Any other stdlib module that we later freeze into the runtime is not essential. We can just as well import from the .py file. This PR lets users explicitly choose which should be used, with the new "-X frozen_modules=[on|off]" CLI flag. The default is "off" for now.
https://bugs.python.org/issue45020
There are a few things I missed in gh-27980. This is a follow-up that will make subsequent PRs cleaner. It includes fixes to tests and tools that reference the frozen modules.
https://bugs.python.org/issue45019
Py_RunMain() now resets PyImport_Inittab to its initial value at
exit. It must be possible to call PyImport_AppendInittab() or
PyImport_ExtendInittab() at each Python initialization.
* pycore_ast.h no longer defines the Yield macro.
* Fix a compiler warning on Windows: "warning C4005: 'Yield': macro
redefinition".
* Python-ast.c now defines directly functions with their real
_Py_xxx() name, rather than xxx().
* Remove "#undef Yield" in C files including pycore_ast.h.
These functions were undocumented and excluded from the limited C
API.
Most names defined by these header files were not prefixed by "Py"
and so could create names conflicts. For example, Python-ast.h
defined a "Yield" macro which was conflict with the "Yield" name used
by the Windows <winbase.h> header.
Use the Python ast module instead.
* Move Include/asdl.h to Include/internal/pycore_asdl.h.
* Move Include/Python-ast.h to Include/internal/pycore_ast.h.
* Remove ast.h header file.
* pycore_symtable.h no longer includes Python-ast.h.
Pass the current interpreter (interp) rather than the current Python
thread state (tstate) to internal functions which only use the
interpreter.
Modified functions:
* _PyXXX_Fini() and _PyXXX_ClearFreeList() functions
* _PyEval_SignalAsyncExc(), make_pending_calls()
* _PySys_GetObject(), sys_set_object(), sys_set_object_id(), sys_set_object_str()
* should_audit(), set_flags_from_config(), make_flags()
* _PyAtExit_Call()
* init_stdio_encoding()
* etc.
Convert the _imp extension module to the multi-phase initialization
API (PEP 489).
* Add _PyImport_BootstrapImp() which fix a bootstrap issue: import
the _imp module before importlib is initialized.
* Add create_builtin() sub-function, used by _imp_create_builtin().
* Initialize PyInterpreterState.import_func earlier, in
pycore_init_builtins().
* Remove references to _PyImport_Cleanup(). This function has been
renamed to finalize_modules() and moved to pylifecycle.c.
* Call _PyTime_Init() and _PyWarnings_InitState() earlier during the
Python initialization.
* Inline _PyImportHooks_Init() into _PySys_InitCore().
* The _warnings initialization function no longer call
_PyWarnings_InitState() to prevent resetting filters_version to 0.
* _PyWarnings_InitState() now returns an int and no longer clear the
state in case of error (it's done anyway at Python exit).
* Rework init_importlib(), fix refleaks on errors.
Move private _PyGC_CollectNoFail() to the internal C API.
Remove the private _PyGC_CollectIfEnabled() which was just an alias
to the public PyGC_Collect() function since Python 3.8.
Rename functions:
* collect() => gc_collect_main()
* collect_with_callback() => gc_collect_with_callback()
* collect_generations() => gc_collect_generations()
If PyDict_GetItemWithError is only used to check whether the key is in dict,
it is better to use PyDict_Contains instead.
And if it is used in combination with PyDict_SetItem, PyDict_SetDefault can
replace the combination.
* PyMapping_HasKey() is not safe because it silences all exceptions and can return incorrect result.
* Informative exceptions from PyMapping_DelItem() are overridden with RuntimeError and
the original exception raised before calling remove_module() is lost.
* There is a race condition between PyMapping_HasKey() and PyMapping_DelItem().
Since _PyImport_ReInitLock() now calls _PyThread_at_fork_reinit() on
the import lock, the lock is now in a known state: unlocked. It
became safe to acquire it after fork.
PyOS_AfterFork_Child() helper functions now return a PyStatus:
PyOS_AfterFork_Child() is now responsible to handle errors.
* Move _PySignal_AfterFork() to the internal C API
* Add #ifdef HAVE_FORK on _PyGILState_Reinit(), _PySignal_AfterFork()
and _PyInterpreterState_DeleteExceptMain().
I can add another commit with the new test case I wrote to verify that the warning was being printed before my change, stopped printing after my change, and that the function does not return null after my change.
Automerge-Triggered-By: @brettcannon
Otherwise we leave a dangling pointer to free'd memory. If we
then initialize a new interpreter in the same process and call
PyImport_ExtendInittab, we will (likely) crash when calling
PyMem_RawRealloc(inittab_copy, ...) since the pointer address
is bogus.
Automerge-Triggered-By: @brettcannon
PyFrame_GetCode(frame): return a borrowed reference to the frame
code.
Replace frame->f_code with PyFrame_GetCode(frame) in most code,
except in frameobject.c, genobject.c and ceval.c.
Also add PyFrame_GetLineNumber() to the limited C API.
Rename _PyInterpreterState_GET_UNSAFE() to _PyInterpreterState_GET()
for consistency with _PyThreadState_GET() and to have a shorter name
(help to fit into 80 columns).
Add also "assert(tstate != NULL);" to the function.
Don't access PyInterpreterState.config member directly anymore, but
use new functions:
* _PyInterpreterState_GetConfig()
* _PyInterpreterState_SetConfig()
* _Py_GetConfig()
Mark Shannon
Victor Stinner
Eric Snow
Georg Brandl
Serhiy Storchaka
Hood Chatham
chgnrdv
Barry Warsaw
Irit Katriel
Sergey B Kirpichev
Brett Cannon
Christian Heimes
Kumar Aditya
Guido van Rossum
Filipe Laíns
Christian Clauss
Hai Shi
junyixie
Antoine Pitrou
Robert Rouhani
Gregory Szorc
Dong-hee Na
Andy Lester