Commit Graph

51 Commits

Author SHA1 Message Date
Mark Shannon 7a65439b93
GH-122390: Replace `_Py_GetbaseOpcode` with `_Py_GetBaseCodeUnit` (GH-122942) 2024-08-13 14:22:57 +01:00
Mark Shannon 2e14a52cce
GH-122160: Remove BUILD_CONST_KEY_MAP opcode. (GH-122164) 2024-07-25 16:24:29 +01:00
Mark Shannon 9cefcc0ee7
GH-120507: Lower the `BEFORE_WITH` and `BEFORE_ASYNC_WITH` instructions. (#120640)
* Remove BEFORE_WITH and BEFORE_ASYNC_WITH instructions.

* Add LOAD_SPECIAL instruction

* Reimplement `with` and `async with` statements using LOAD_SPECIAL
2024-06-18 12:17:46 +01:00
Irit Katriel 97b69db167
gh-93691: fix too broad source locations of for statement iterators (#120330) 2024-06-12 12:53:19 +01:00
Jelle Zijlstra 98e855fcc1
gh-119180: Add LOAD_COMMON_CONSTANT opcode (#119321)
The PEP 649 implementation will require a way to load NotImplementedError
from the bytecode. @markshannon suggested implementing this by converting
LOAD_ASSERTION_ERROR into a more general mechanism for loading constants.

This PR adds this new opcode. I will work on the rest of the implementation
of the PEP separately.

Co-authored-by: Irit Katriel <1055913+iritkatriel@users.noreply.github.com>
2024-05-22 00:46:39 +00:00
Sam Gross 2ba2c142a6
gh-118527: Intern code name and filename on default build (#118576)
Interned and non-interned strings are treated differently by `marshal`,
so be consistent between the default and free-threaded build.
2024-05-06 17:24:14 -04:00
Mark Shannon 981d172f7f
GH-112354: `END_FOR` instruction to only pop one value. (GH-114247)
* Compiler emits END_FOR; POP_TOP instead of END_FOR. To support tier 2 side exits in loops.
2024-01-24 15:10:17 +00:00
Irit Katriel d49aba5a7a
gh-111354: Simplify _PyGen_yf by moving some of its work to the compiler and frame state (#111648) 2023-11-03 10:01:36 +00:00
Irit Katriel 52cc4af6ae
gh-111354: simplify detection of RESUME after YIELD_VALUE at except-depth 1 (#111459) 2023-11-02 10:18:43 +00:00
Brandt Bucher 22e65eecaa
GH-105848: Replace KW_NAMES + CALL with LOAD_CONST + CALL_KW (GH-109300) 2023-09-13 10:25:45 -07:00
Irit Katriel 8b55adfa8f
gh-109256: allocate opcode IDs for internal opcodes in their own range (#109269) 2023-09-12 10:36:17 +00:00
Mark Shannon 0858328ca2
GH-108614: Add `RESUME_CHECK` instruction (GH-108630) 2023-09-07 14:39:03 +01:00
Irit Katriel 665a4391e1
gh-105481: generate op IDs from bytecode.c instead of hard coding them in opcode.py (#107971) 2023-08-16 22:25:18 +00:00
Brandt Bucher a9caf9cf90
GH-105848: Simplify the arrangement of CALL's stack (GH-107788) 2023-08-09 18:19:39 +00:00
Mark Shannon 1d857da7f0
GH-77273: Better bytecodes for f-strings (GH-6132) 2023-06-14 16:15:08 +01:00
Mark Shannon 4bfa01b9d9
GH-104584: Plugin optimizer API (GH-105100) 2023-06-02 11:46:18 +01:00
Pablo Galindo Salgado 1ef61cf71a
gh-102856: Initial implementation of PEP 701 (#102855)
Co-authored-by: Lysandros Nikolaou <lisandrosnik@gmail.com>
Co-authored-by: Batuhan Taskaya <isidentical@gmail.com>
Co-authored-by: Marta Gómez Macías <mgmacias@google.com>
Co-authored-by: sunmy2019 <59365878+sunmy2019@users.noreply.github.com>
2023-04-19 11:18:16 -05:00
Brandt Bucher b4978ff872
GH-88691: Shrink the CALL caches (GH-103230) 2023-04-05 14:15:49 -07:00
Brandt Bucher 121057aa36
GH-89987: Shrink the BINARY_SUBSCR caches (GH-103022) 2023-03-29 15:53:30 -07:00
penguin_wwy 753fc8a5d6
gh-101632: Add the new RETURN_CONST opcode (#101633) 2023-02-07 22:32:21 +00:00
Brandt Bucher 3c033a2e6f
GH-99554: Pack location tables more effectively (GH-99556) 2022-12-21 16:41:18 -08:00
Mark Shannon 22863df7ca
GH-96793: Change `FOR_ITER` to not pop the iterator on exhaustion. (GH-96801)
Change FOR_ITER to have the same stack effect regardless of whether it branches or not.
Performance is unchanged as FOR_ITER (and specialized forms jump over the cleanup code).
2022-10-27 11:55:03 +01:00
Irit Katriel 324d01944d
gh-94485: Set line number of module's RESUME instruction to 0, as specified by PEP 626 (GH-94552)
Co-authored-by: Mark Shannon <mark@hotpy.org>
2022-07-05 14:38:44 +02:00
Dennis Sweeney 5fcfdd87c9
GH-91432: Specialize FOR_ITER (GH-91713)
* Adds FOR_ITER_LIST and FOR_ITER_RANGE specializations.

* Adds _PyLong_AssignValue() internal function to avoid temporary boxing of ints.
2022-06-21 11:19:26 +01:00
Ken Jin b083450f88
GH-93429: Merge `LOAD_METHOD` back into `LOAD_ATTR` (GH-93430) 2022-06-14 11:36:22 +01:00
Ken Jin d52ffc1d1f
gh-93382: Cache result of `PyCode_GetCode` in codeobject (GH-93383)
Co-authored-by: Kumar Aditya <59607654+kumaraditya303@users.noreply.github.com>
Co-authored-by: Dennis Sweeney <36520290+sweeneyde@users.noreply.github.com>
2022-06-04 00:41:18 +08:00
Mark Shannon e48ac9c100
GH-90690: Remove `PRECALL` instruction (GH-92925) 2022-05-19 11:05:26 +01:00
Inada Naoki 6dcfd6c5e3
gh-78214: marshal: Stabilize FLAG_REF usage (GH-8226)
Use FLAG_REF always for interned strings.

Refcounts of interned string is very unstable.
When compiling same source, refcounts of interned string in the output may be 1 or >1.
It makes FLAG_REF usage unstable.

To help reproducible build, use FLAG_REF for interned string even if refcnt(obj)==1.
2022-05-04 10:01:15 +09:00
Mark Shannon 944fffee89
GH-88116: Use a compact format to represent end line and column offsets. (GH-91666)
* Stores all location info in linetable to conform to PEP 626.

* Remove column table from code objects.

* Remove end-line table from code objects.

* Document new location table format
2022-04-21 16:10:37 +01:00
Irit Katriel a00518d9ad
bpo-47120: Replace the JUMP_ABSOLUTE opcode by the relative JUMP_BACKWARD (GH-32115) 2022-03-31 14:14:15 +01:00
Brandt Bucher f193631387
bpo-46841: Use inline caching for calls (GH-31709) 2022-03-07 11:45:00 -08:00
Brandt Bucher 05a8bc1c94
bpo-46841: Use inline caching for attribute accesses (GH-31640) 2022-03-03 15:31:00 -08:00
Mark Shannon 3b0f1c5a71
bpo-46841: Use inline cache for `BINARY_SUBSCR`. (GH-31618) 2022-03-01 16:00:34 +00:00
Mark Shannon 59585d6b2e
bpo-46329: Streamline calling sequence a bit. (GH-31465)
* Move handling of bound-methods to PRECALL.

* Remove call_shape.postcall_shrink

* Remove call_shape.callable

* Remove call_shape.callable. Change CALL oparg to match PRECALL oparg.

* Move KW_NAMES before PRECALL.

* Update opcode docs in dis.rst
2022-02-21 18:26:47 +00:00
Mark Shannon cf345e945f
bpo-46329: Change calling sequence (again) (GH-31373)
* Change calling sequence: Add PUSH_NULL. Merge PRECALL_FUNCTION and PRECALL_METHOD into PRECALL.
2022-02-18 17:19:08 +00:00
Eric Snow 81c72044a1
bpo-46541: Replace core use of _Py_IDENTIFIER() with statically initialized global objects. (gh-30928)
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
2022-02-08 13:39:07 -07:00
Mark Shannon 89fd7c3452
bpo-46329: Split calls into precall and call instructions. (GH-30855)
* Add PRECALL_FUNCTION opcode.

* Move 'call shape' varaibles into struct.

* Replace CALL_NO_KW and CALL_KW with KW_NAMES and CALL instructions.

* Specialize for builtin methods taking using the METH_FASTCALL | METH_KEYWORDS protocol.

* Allow kwnames for specialized calls to builtin types.

* Specialize calls to tuple(arg) and str(arg).
2022-01-28 12:42:30 +00:00
Mark Shannon e028ae99ec
bpo-45923: Handle call events in bytecode (GH-30364)
* Add a RESUME instruction to handle "call" events.
2022-01-06 13:09:25 +00:00
Mark Shannon 9f8f45144b
bpo-44525: Split calls into PRECALL and CALL (GH-30011)
* Add 3 new opcodes for calls: PRECALL_METHOD, CALL_NO_KW, CALL_KW.

* Update specialization to handle new CALL opcodes.

* Specialize call to method descriptors.

* Remove old CALL opcodes: CALL_FUNCTION, CALL_METHOD, CALL_METHOD_KW, CALL_FUNCTION_KW.
2021-12-14 18:22:44 +00:00
Gabriele N. Tornetta 2f180ce2cb
bpo-44530: Add co_qualname field to PyCodeObject (GH-26941) 2021-07-07 12:21:51 +01:00
Pablo Galindo 98eee94421
bpo-43950: Add code.co_positions (PEP 657) (GH-26955)
This PR is part of PEP 657 and augments the compiler to emit ending
line numbers as well as starting and ending columns from the AST
into compiled code objects. This allows bytecodes to be correlated
to the exact source code ranges that generated them.

This information is made available through the following public APIs:

* The `co_positions` method on code objects.
* The C API function `PyCode_Addr2Location`.

Co-authored-by: Batuhan Taskaya <isidentical@gmail.com>
Co-authored-by: Ammar Askar <ammar@ammaraskar.com>
2021-07-02 15:10:11 +01:00
Batuhan Taskaya 1b28187a0e
bpo-44313: generate LOAD_ATTR/CALL_FUNCTION for top-level imported objects (GH-26677) 2021-06-30 23:53:36 +01:00
Guido van Rossum 769d7d0c66
bpo-43693 Get rid of CO_NOFREE -- it's unused (GH-26839)
All uses of this flag are either setting it
or in doc or tests for it. So we should be
able to get rid of it completely.
2021-06-23 09:51:44 -07:00
Guido van Rossum 355f5dd36a
bpo-43693: Turn localspluskinds into an object (GH-26749)
Managing it as a bare pointer to malloc'ed bytes is just too awkward in a few places.
2021-06-21 13:53:04 -07:00
Eric Snow 2ab27c4af4
bpo-43693: Un-revert commits 2c1e258 and b2bf2bc. (gh-26577)
These were reverted in gh-26530 (commit 17c4edc) due to refleaks.

* 2c1e258 - Compute deref offsets in compiler (gh-25152)
* b2bf2bc - Add new internal code objects fields: co_fastlocalnames and co_fastlocalkinds. (gh-26388)

This change fixes the refleaks.

https://bugs.python.org/issue43693
2021-06-07 12:22:26 -06:00
Pablo Galindo 17c4edc4e0
bpo-43693: Revert commits 2c1e2583fd and b2bf2bc1ec (GH-26530)
* Revert "bpo-43693: Compute deref offsets in compiler (gh-25152)"

This reverts commit b2bf2bc1ec.

* Revert "bpo-43693: Add new internal code objects fields: co_fastlocalnames and co_fastlocalkinds. (gh-26388)"

This reverts commit 2c1e2583fd.

These two commits are breaking the refleak buildbots.
2021-06-04 17:51:05 +01:00
Eric Snow 2c1e2583fd
bpo-43693: Add new internal code objects fields: co_fastlocalnames and co_fastlocalkinds. (gh-26388)
A number of places in the code base (notably ceval.c and frameobject.c) rely on mapping variable names to indices in the frame "locals plus" array (AKA fast locals), and thus opargs.  Currently the compiler indirectly encodes that information on the code object as the tuples co_varnames, co_cellvars, and co_freevars.  At runtime the dependent code must calculate the proper mapping from those, which isn't ideal and impacts performance-sensitive sections.  This is something we can easily address in the compiler instead.

This change addresses the situation by replacing internal use of co_varnames, etc. with a single combined tuple of names in locals-plus order, along with a minimal array mapping each to its kind (local vs. cell vs. free).  These two new PyCodeObject fields, co_fastlocalnames and co_fastllocalkinds, are not exposed to Python code for now, but co_varnames, etc. are still available with the same values as before (though computed lazily).

Aside from the (mild) performance impact, there are a number of other benefits:

* there's now a clear, direct relationship between locals-plus and variables
* code that relies on the locals-plus-to-name mapping is simpler
* marshaled code objects are smaller and serialize/de-serialize faster

Also note that we can take this approach further by expanding the possible values in co_fastlocalkinds to include specific argument types (e.g. positional-only, kwargs).  Doing so would allow further speed-ups in _PyEval_MakeFrameVector(), which is where args get unpacked into the locals-plus array.  It would also allow us to shrink marshaled code objects even further.

https://bugs.python.org/issue43693
2021-06-03 10:28:27 -06:00
Eric Snow 9f494d4929
bpo-43693: Add _PyCode_New(). (gh-26375)
This is an internal-only API that helps us manage the many values used to create a code object.

https://bugs.python.org/issue43693
2021-05-27 09:54:34 -06:00
Victor Stinner 58a993957a
bpo-44131: Py_FrozenMain() uses PyConfig_SetBytesArgv() (GH-26201)
Moreover, Py_FrozenMain() relies on Py_InitializeFromConfig() to
handle the PYTHONUNBUFFERED environment variable and configure C
stdio streams like stdout (make the stream unbuffered).
2021-05-20 12:08:05 +02:00
Victor Stinner 834498e178
bpo-44131: Fix Makefile for test_frozenmain (GH-26203)
Remove Programs/test_frozenmain.h Makefile target: it ran make
in parallel which caused build errors on LTO+PGO builds.
2021-05-18 08:40:37 +02:00