Skip to content

Simplify the interpreter's (type, val, tb) exception representation #89874

Description

@iritkatriel
BPO 45711
Nosy @gvanrossum, @terryjreedy, @scoder, @vstinner, @markshannon, @brandtbucher, @da-woods, @iritkatriel
PRs
  • bpo-45711: Use _PyErr_ClearExcState instead of setting only exc_value… #29404
  • bpo-45711: remove unnecessary DUP_TOP and POP in exception handling #29495
  • bpo-45711: assert that the type of exc_info is redundant #29518
  • bpo-45711: Re-bump the magic number and update doc #29528
  • bpo-45711: use exc_value instead of exc_type to determine if exc_info is valid. Add more assertions. #29627
  • bpo-45711: Change exc_info related APIs to derive type and traceback from the exception instance #29780
  • bpo-45711: [asyncio] Normalize exceptions immediately after Fetch, before they are stored as StackItem, which should be normalized #29890
  • bpo-45711: Remove unnecessary normalization of exc_info #29922
  • bpo-45711: Remove type and traceback from exc_info #30122
  • bpo-46219, 46221: simplify except* implementation following exc_info changes. Move helpers to exceptions.c. Do not assume that exception groups are truthy. #30289
  • bpo-45711: move whatsnew entries which are incorrectly listed under N… #30849
  • Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.

    Show more details

    GitHub fields:

    assignee = None
    closed_at = <Date 2022-03-13.19:33:16.899>
    created_at = <Date 2021-11-04.11:29:47.242>
    labels = ['interpreter-core', '3.11', 'performance']
    title = "Simplify the interpreter's (type, val, tb) exception representation"
    updated_at = <Date 2022-03-13.19:33:16.899>
    user = 'https://github.com/iritkatriel'

    bugs.python.org fields:

    activity = <Date 2022-03-13.19:33:16.899>
    actor = 'iritkatriel'
    assignee = 'none'
    closed = True
    closed_date = <Date 2022-03-13.19:33:16.899>
    closer = 'iritkatriel'
    components = ['Interpreter Core']
    creation = <Date 2021-11-04.11:29:47.242>
    creator = 'iritkatriel'
    dependencies = []
    files = []
    hgrepos = []
    issue_num = 45711
    keywords = ['patch']
    message_count = 31.0
    messages = ['405681', '405839', '405849', '406114', '406120', '406209', '406238', '406983', '406984', '407402', '407597', '408043', '408768', '408786', '409453', '411497', '412264', '412265', '412270', '412272', '412273', '412275', '412281', '412289', '412290', '412292', '412303', '412304', '412316', '414482', '415086']
    nosy_count = 8.0
    nosy_names = ['gvanrossum', 'terry.reedy', 'scoder', 'vstinner', 'Mark.Shannon', 'brandtbucher', 'da-woods', 'iritkatriel']
    pr_nums = ['29404', '29495', '29518', '29528', '29627', '29780', '29890', '29922', '30122', '30289', '30849']
    priority = 'normal'
    resolution = 'fixed'
    stage = 'resolved'
    status = 'closed'
    superseder = None
    type = 'performance'
    url = 'https://bugs.python.org/issue45711'
    versions = ['Python 3.11']

    Activity

    1. iritkatriel commented on Nov 4, 2021

      @iritkatriel
      MemberAuthor

      Exceptions are represented in the interpreter as (type, val, tb) triplets which most of the time contain redundant information (the type is the type of val and the tb is also on the exception). This complicates the code and is inefficient as opcodes that manage exceptions push and pop 3 items for each exception.

      We will change the internal representation to be (1) just the exception value if it is normalised and (2) a tuple of the 3 values for the uncommon case where they are all needed.

      See also faster-cpython/ideas#106.

    2. terryjreedy commented on Nov 6, 2021

      @terryjreedy
      Member

      Would there be any change at the Python level?

    3. iritkatriel commented on Nov 6, 2021

      @iritkatriel
      MemberAuthor

      Initially not, neither in python nor in the c api.

      It would be nice to replace PyErr_Fetch/Restore by a version that takes just an exception but that’s a long deprecation.

    4. iritkatriel commented on Nov 10, 2021

      @iritkatriel
      MemberAuthor

      New changeset 05fbd60 by Irit Katriel in branch 'main':
      bpo-45711: Use _PyErr_ClearExcState instead of setting only exc_value to NULL (GH-29404)
      05fbd60

    5. markshannon commented on Nov 10, 2021

      @markshannon
      Member

      New changeset 4cdeee5 by Irit Katriel in branch 'main':
      bpo-45711: remove unnecessary DUP_TOP and POP in exception handling (GH-29495)
      4cdeee5

    6. iritkatriel commented on Nov 12, 2021

      @iritkatriel
      MemberAuthor

      New changeset 8f1b71d by Brandt Bucher in branch 'main':
      bpo-45711: Re-bump the magic number and update doc (GH-29528)
      8f1b71d

    7. iritkatriel commented on Nov 12, 2021

      @iritkatriel
      MemberAuthor

      New changeset de3db14 by Irit Katriel in branch 'main':
      bpo-45711: assert that the type of exc_info is redundant (GH-29518)
      de3db14

    8. iritkatriel commented on Nov 25, 2021

      @iritkatriel
      MemberAuthor

      New changeset c456dfa by Irit Katriel in branch 'main':
      bpo-45711: use exc_value instead of exc_type to determine if exc_info is valid. Add more assertions. (GH-29627)
      c456dfa

    9. iritkatriel commented on Nov 25, 2021

      @iritkatriel
      MemberAuthor

      Following the analysis/discussion on faster-cpython/ideas#106:

      • exc_info is always normalized, so we actually will never need to create a tuple of these three values (exc_curinfo, the exception raised but not yet caught can be unnormalized, but it is not pushed/popped on the stack).

      • We will reduce the interpreter's exc_info representation to just the exception instance.

      • There are two APIs that are impacted, both in non-documented edge cases:

      1. sys.exc_info()[2] can currently be different from sys.exc_info()[1].__traceback__ because changes to the latter (while an except clause is executing) don't show up in the former. This is arguably a bug, we will change it so that the type and traceback are always consistent with the exception.

      2. PyErr_SetExcInfo does no arg checking, and will set exc_info to an inconsistent triplet if you ask it to. However, the exc_value arg must be an exception instance so the only thing you can do is pass in nonsensical args where the type/traceback do not match the exception. This function's purpose is to save/restore exc_info. We will make it ignore the type and traceback and document that change.

    10. iritkatriel commented on Nov 30, 2021

      @iritkatriel
      MemberAuthor

      New changeset 8a45ca5 by Irit Katriel in branch 'main':
      bpo-45711: Change exc_info related APIs to derive type and traceback from the exception instance (GH-29780)
      8a45ca5

    11. iritkatriel commented on Dec 3, 2021

      @iritkatriel
      MemberAuthor

      New changeset 2ff758b by Irit Katriel in branch 'main':
      bpo-45711: [asyncio] Normalize exceptions immediately after Fetch, before they are stored as StackItem, which should be normalized (GH-29890)
      2ff758b

    12. iritkatriel commented on Dec 8, 2021

      @iritkatriel
      MemberAuthor

      New changeset 2109f78 by Irit Katriel in branch 'main':
      bpo-45711: Remove unnecessary normalization of exc_info (GH-29922)
      2109f78

    13. 12 remaining items

    14. iritkatriel commented on Feb 1, 2022

      @iritkatriel
      MemberAuthor

      That commit has significant changes in ceval.c and compile.c. They don't need to be reverted to unbreak cython. I'm working on a PR for a simpler change.

      Have you tried with CYTHON_USE_EXC_INFO_STACK undefined?

    15. iritkatriel commented on Feb 1, 2022

      @iritkatriel
      MemberAuthor

      If this is still the position of cython maintainers:

      cython/cython#4581 (comment)

      then I will need to revert the change until 3.12.

    16. gvanrossum commented on Feb 1, 2022

      @gvanrossum
      Member

      Time to insist on directly communicating with the Cython team (esp. @scoder) and broker some kind of compromise.

    17. vstinner commented on Feb 1, 2022

      @vstinner
      Member

      __Pyx_PyErr_GetTopmostException(PyThreadState *tstate)

      Python provides a *private* _PyErr_GetTopmostException(tstate) function, but Cython reimplements its own function. I'm not sure why.

      #74716 proposes adding PyErr_GetActiveException() function which has no parameter, but Cython __Pyx_PyErr_GetTopmostException() has a tstate parameter.

      Simplified example numpy/random/_mt19937.c code:

      static CYTHON_INLINE void
      __Pyx__ExceptionSave(PyThreadState *tstate, PyObject **type,
                           PyObject **value, PyObject **tb)
      {
          _PyErr_StackItem *exc_info = __Pyx_PyErr_GetTopmostException(tstate);
          *type = exc_info->exc_type;
          *value = exc_info->exc_value;
          *tb = exc_info->exc_traceback;
          Py_XINCREF(*type);
          Py_XINCREF(*value);
          Py_XINCREF(*tb);
      }
      
      static CYTHON_INLINE void
      __Pyx__ExceptionReset(PyThreadState *tstate, PyObject *type,
                            PyObject *value, PyObject *tb)
      {
          PyObject *tmp_type, *tmp_value, *tmp_tb;
          _PyErr_StackItem *exc_info = tstate->exc_info;
          tmp_type = exc_info->exc_type;
          tmp_value = exc_info->exc_value;
          tmp_tb = exc_info->exc_traceback;
          exc_info->exc_type = type;
          exc_info->exc_value = value;
          exc_info->exc_traceback = tb;
          Py_XDECREF(tmp_type);
          Py_XDECREF(tmp_value);
          Py_XDECREF(tmp_tb);
      }

      Cython saves/restores the current exception of tstate. Maybe we need to provide a high-level API for that as well?

    18. iritkatriel commented on Feb 1, 2022

      @iritkatriel
      MemberAuthor

      This is a backport of @scoder's patch to 0.29.x. (I don't know if this is helpful).

      https://github.com/cython/cython/compare/master...iritkatriel:exc_info?expand=1

    19. iritkatriel commented on Feb 1, 2022

      @iritkatriel
      MemberAuthor

      #74716 proposes adding PyErr_GetActiveException() function which has no parameter, but Cython __Pyx_PyErr_GetTopmostException() has a tstate parameter.

      I've now updated it to follow the pattern of other functions, where the is a private function that takes tstate and the public function calls it.

      So it adds in Include/pyerrors.h

      PyAPI_FUNC(PyObject*) PyErr_GetActiveException(void);
      PyAPI_FUNC(void) PyErr_SetActiveException(PyObject *);

      and in Include/cpython/pyerrors.h

      PyAPI_FUNC(PyObject*) _PyErr_GetActiveException(PyThreadState *);
      PyAPI_FUNC(void) _PyErr_SetActiveException(PyThreadState *, PyObject *);
    20. da-woods commented on Feb 1, 2022

      da-woodsmannequin
      Mannequin

      Just a quick comment on Cython and these changes:

      Cython 0.29 can build itself on Python 3.11a4 with CFLAGS="-DCYTHON_FAST_THREAD_STATE=0 -DCYTHON_USE_EXC_INFO_STACK=0" python3.11 setup.py build_ext.

      I think there's some coroutine code left that isn't fixed by those flags, but a large chunk of Cython libraries won't use coroutines and so will work fine with these flags. Hopefully that unblocks some stuff for you.

    21. vstinner commented on Mar 3, 2022

      @vstinner
      Member

      lxml does crash on the current Cython 0.29.x development branch.

    22. iritkatriel commented on Mar 13, 2022

      @iritkatriel
      MemberAuthor

      We have agreed on python-dev [1] that the cpython changes will not be reverted, and the issue will be fixed in cython. So I am closing this again.

      [1] https://mail.python.org/archives/list/python-dev@python.org/message/BHIQL4P6F7OPMCAP6U24XEZUPQKI62UT/

    23. transferred this issue fromon Apr 10, 2022
    24. added a commit that references this issue on Dec 8, 2023
    Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

    Metadata

    Metadata

    Assignees

    No one assigned

      Labels

      3.11only security fixesinterpreter-core(Objects, Python, Grammar, and Parser dirs)performancePerformance or resource usage

      Projects

      No projects

        Milestone

        No milestone

        Relationships

        None yet

        Development

        No branches or pull requests

        Issue actions