bpo-29403: Fix mock's broken autospec behavior on method-bound builtin functions - #3

habnabit · 2017-02-10T23:07:39Z

http://bugs.python.org/issue29403

Cython will, in the right circumstances, offer a MethodType instance
where im_func is a builtin function. Any instance of MethodType is
automatically assumed to be a python-defined function (more
specifically, a function that has an inspectable signature), but
_set_signature was still conservative in its assumptions. As a result
_set_signature would return early with None instead of a mock since the
im_func had no inspectable signature. This causes problems deeper inside
mock, as _set_signature is assumed to always return a mock, and
nothing checked its return value.

In similar corner cases, autospec will simply not check the spec of the
function, so _set_signature is amended to now return early with the
original, not-wrapped mock object.

the-knights-who-say-ni · 2017-02-10T23:07:40Z

Hello, and thanks for your contribution!

I'm a bot set up to make sure that the project can legally accept your contribution by verifying you have signed the PSF contributor agreement (CLA).

Unfortunately we couldn't find an account corresponding to your GitHub username on bugs.python.org (b.p.o) to verify you have signed the CLA. This is necessary for legal reasons before we can look at your contribution. Please follow these steps to help rectify the issue:

If you don't have an account on b.p.o, please create one
Make sure your GitHub username is listed in "Your Details" at b.p.o
If you have not already done so, please sign the PSF contributor agreement
If you just signed the CLA, please wait at least a day and then check "Your Details" on bugs.python.org to see if your account has been marked as having signed the CLA (the delay is due to a person having to manually check your signed CLA)
Reply here saying you have completed the above steps

Thanks again to your contribution and we look forward to looking at it!

habnabit · 2017-02-10T23:09:17Z

I signed the CLA when I opened the bug on bpo last week, but my account doesn't say received still: http://bugs.python.org/user25445

ShalbafZadeh · 2017-02-10T23:36:48Z

@habnabit
have you added you github username to your bpo account?
i can't see it
https://bugs.python.org/user6636

habnabit · 2017-02-10T23:38:30Z

@ShalbafZadeh that was an old account; I made a new one last week from a newer openid provider. http://bugs.python.org/user25445 is the new one, which I'm logged into now.

ShalbafZadeh · 2017-02-10T23:42:51Z

@habnabit seems odd

habnabit · 2017-02-11T01:49:12Z

Rebased and pushed, including a test fix. The third-party mock used six, and the stdlib does not include six.

habnabit · 2017-02-11T06:26:14Z

Rebased again!

codecov · 2017-02-11T06:55:24Z

Codecov Report

Merging #3 into master will increase coverage by <.01%.

@@            Coverage Diff             @@
##           master       #3      +/-   ##
==========================================
+ Coverage   82.37%   82.37%   +<.01%     
==========================================
  Files        1427     1427              
  Lines      350948   350959      +11     
==========================================
+ Hits       289088   289104      +16     
+ Misses      61860    61855       -5

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e7ffb99...97bb733. Read the comment docs.

berkerpeksag

I haven't had a time to dive into the whole life cycle of _set_signature() yet, but from a quick glance it looks good to me.

berkerpeksag · 2017-02-22T04:58:29Z

Please add an entry to Misc/NEWS. It should be in the "Library" section.

habnabit · 2017-02-23T22:41:40Z

Okay, rebased and added a NEWS entry.

Yhg1s · 2017-05-22T17:41:52Z

This change looks good to me: returning None does not make sense in the one place where _set_signature is called, and I don't think there's a better solution here. If the code wants to try harder and figure out what signature to use, that would be a change to _get_signature, not _set_signature, and this change would still make sense.

@voidspace, do you want to take a look, or can we merge?

buhtz · 2017-07-19T12:07:32Z

Can we merge this?

Cython will, in the right circumstances, offer a MethodType instance where im_func is a builtin function. Any instance of MethodType is automatically assumed to be a python-defined function (more specifically, a function that has an inspectable signature), but _set_signature was still conservative in its assumptions. As a result _set_signature would return early with None instead of a mock since the im_func had no inspectable signature. This causes problems deeper inside mock, as _set_signature is assumed to _always_ return a mock, and nothing checked its return value. In similar corner cases, autospec will simply not check the spec of the function, so _set_signature is amended to now return early with the original, not-wrapped mock object.

Per pythia python#79 python#3 [chat L2032] + supervisor [chat L2034] assignment: the brace-counted scan methodology that produced 100/123 = 81.3% Tier 5 close ratio existed only in chat history (D-1776880902). Anyone re-grepping in two weeks reproduces the same /144-class lapse. This script anchors the methodology in repo: - Brace-balanced extraction of HIRBuilder::emit* method definitions (multi-line signatures + any return type + balanced { } body) - Categorization: stub (≤8 body lines + has hir_builder_emit_*_c call) / partial (>8 body lines + has C call) / pure C++ (no C call) - Output: counts + exhaustive pure-C++ + partial method lists Reference values verified at HEAD a642405 (post-Tier-5-close, push 62): TOTAL: 123 STUBS: 93 PARTIAL: 7 PURE C++: 23 RATIO: 100/123 = 81.3% Closes pythia python#79 python#3 'methodology not committed to repo' substance. Anyone running scripts/count_emit_methods.sh reproduces the canonical baseline without needing chat-search or external memory. Origin lessons embedded in script header: - /144 propagated 2026-04-21 (commit 7783df7) → 2026-04-22 Tier 5 close - Real denominator at HEAD a642405 was 123 (pythia python#78 python#1 catch) - Methodology lives in repo, not chat Authorization chain: - pythia python#79 python#3 surfaced gap: chat L2032 - supervisor python#3 assignment: chat L2034

Per W25 spec §3 Step B + theologian L2189 4-class taxonomy. B-1 batch (~10 lint-pattern externs across 3 small TUs) per spec §3 sequencing (B-0 cfg.cpp d0e8031 validated Class C2 pattern; B-1 scales to Class A + Class C2 mix). PER-TU CLASS CLASSIFICATION + CLEANUP: ssaify_c.c (1 lint extern): hir_func_alloc_register → Class A (hir_c_api.h:52) Fix: add #include hir_c_api.h, delete local extern. Kept non-lint externs (hir_reflow_types, hir_phi_elimination_run) as out-of-W25-scope; tracked separately. licm_c.c (3 lint externs): hir_func_cfg_ptr → Class C2 (hir_instr_c.h:232 static inline, already included via hir_instr_c.h) hir_instr_unlink → Class A (hir_c_api.h) hir_c_insert_before → Class A (hir_c_api.h:726) Fix: add #include hir_c_api.h, delete 3 lint externs. Kept hir_reg_instr (non-lint) as out-of-scope. resolve_kwargs_c.c (5 lint externs): hir_func_cfg_ptr → Class C2 (already accessible) hir_instr_replace_with → Class A hir_c_create_vectorcall_reg → Class A hir_c_create_call_method_reg → Class A hir_c_copy_frame_state → Class A Fix: add #include hir_c_api.h, delete 5 lint externs + duplicate typedef HirFunction (already in hir_c_api.h). Kept hir_register_type (non-lint) as out-of-scope. DIFF: +7/-13 across 3 files (lint extern reduction without API expansion). DEFERRED to B-2: hir_basic_block_c.c (Class C1 — hir_bb_destroy not in any header). Needs API-expansion decision before deletion. PER GATEITEM python#3 HARDENING (post L2174 lapse): testkeeper to verify post-commit, working tree clean. Both HEAD-match AND tree-clean must hold for valid PASS report. Authorization chain: - W25 spec §3 + 4-class taxonomy: theologian L2017 + L2189 - B-1 sequencing per supervisor [chat L2179] - QUERY-BEFORE-PIVOT discipline applied to scope decisions

Per theologian L2430 INVOKE_* Phase 2 spec. Changes: - Python/jit/hir/builder_emit_c.c: new C body hir_builder_emit_invoke_function_c (~115 lines). Mirrors C++ emitInvokeFunction with three paths (A/B/C): - Path A: container_is_immutable + is_function + is_statically_typed → direct InvokeStaticFunction emit (no VectorCall fallback) - Path B: container_is_immutable + is_builtin → tryEmitDirectMethodCall fast path - Path C: VectorCall fallback (or patchable LoadFunctionIndirect path) PY_VERSION_HEX >= 0x030C0000 conditional for __static__.rand special-case folded into hir_builder_is_static_rand_and_try_emit_c bridge (version-guarded inside the bridge, no #if in C body). - Python/jit/hir/builder.cpp: HIRBuilder::emitInvokeFunction becomes a 7-line delegation stub. Adds 4 NEW C bridges (function-target variants of python#2 bridges, since invokeFunctionTarget != invokeMethodTarget): - hir_builder_invoke_function_target_c: 8-field InvokeTarget query (container_is_immutable, is_function, is_statically_typed, is_builtin, callable, func, indirect_ptr, return_type) - hir_builder_try_emit_direct_method_call_for_function_c: path B fast path - hir_builder_setup_static_args_for_function_c: arg_regs via VLA - hir_builder_is_static_rand_and_try_emit_c: PY_VERSION_HEX folded - Python/jit/hir/builder.h: extern "C" forward decls + 4 new friend declarations granting bridges access to private members. C body uses C99 VLA (void *arg_regs[nargs]). CallFlags::Static = 1<<2 hardcoded as 4u (matches hir.h:881-886). Direct C calls to hir_c_create_invoke_static_function_reg + hir_c_create_load_function_indirect_reg + hir_c_create_load_const + hir_c_create_vectorcall_reg (all already exist in hir_c_api.h, no new factories needed). hir_type_from_object handles Type::fromObject conversion in C. Verification: cmake --build target jit/phoenix_jit clean (BUILD_EXIT=0). Sole-path validation pending — INVOKE_FUNCTION only emitted in __static__ modules, standard test suite does not exercise.

PRIMITIVE_BINARY_OP (Static Python) opcode handler. Per W27a Step A scope. C body: hir_builder_emit_primitive_binary_op_c (~75 lines incl static helpers inlined). Inlines get_primitive_bin_op_kind (PRIM_OP_* → HIR_BOP_* mapping) + is_double_binop (DBL variant detection) directly into a single switch statement. PHX_PRIM_OP_* values hard-coded to mirror cinderx/StaticPython/classloader.h: 70-90 — same pattern as builder_emit_c.c:2862 avoidance of classloader.h in C TU (header pulls in extensive C++-cinderx surface). Static Python ABI is stable; values unchanged across releases. C primitives used: hir_c_create_int_binary_op + hir_c_create_double_binary_op + hir_builder_temps_alloc_stack (all existing). C++ HIRBuilder::emitPrimitiveBinaryOp → 3-line delegation stub. Verification: cmake --build target jit/phoenix_jit clean (BUILD_EXIT=0). Sole-path validation: PRIMITIVE_BINARY_OP only in __static__ modules (emitInvokeNative precedent caveat acknowledged).

PRIMITIVE_COMPARE_OP (Static Python) opcode handler. Per W27a Step A scope. C body: hir_builder_emit_primitive_compare_c (~50 lines). Inlines the PRIM_OP_* → HIR_PCMP_* mapping switch directly. Reuses PHX_PRIM_OP_* defines from python#3 (emitPrimitiveBinaryOp). C primitives used: hir_c_create_primitive_compare + hir_builder_temps_alloc_stack (both existing). C++ HIRBuilder::emitPrimitiveCompare → 3-line delegation stub. Verification: cmake --build target jit/phoenix_jit clean (BUILD_EXIT=0).

BUILD_CHECKED_LIST (Static Python) opcode handler. Per W27b Step A scope. C body: hir_builder_emit_build_checked_list_c (~25 lines). Pops descr + list_size from const_arg tuple, queries preloader type+py_type via existing bridges, emits make_checked_list_reg, fills operands right-to-left from stack. Includes added: cinderx/StaticPython/checked_list.h (Ci_CheckedList_TypeCheck). Header verified C-safe: Ci_CHECKED_LIST_H guard (unique), no Py_OPCODE_H collision (theologian L2515 caveat satisfied). C primitives used: hir_builder_preloader_type + hir_builder_preloader_py_type + hir_c_create_make_checked_list_reg + hir_c_set_operand + hir_builder_temps_alloc_stack (all existing). C++ HIRBuilder::emitBuildCheckedList → 3-line delegation stub. Verification: cmake --build target jit/phoenix_jit clean (BUILD_EXIT=0).

Helper for emitSetupWith + emitBeforeWith. Per W27c Step A scope. C body: hir_builder_emit_setup_with_common_c (~45 lines). Pops manager, allocates enter+exit, emits load_attr_special_reg for both, pushes exit, allocates enter_result, emits vectorcall_reg with CallFlags::None=0u, sets enter as operand 0, sets frame state, emits. PY_VERSION_HEX conditional (_Py_Identifier vs PyObject*) folded into C++ stub: ids are passed as opaque void* downward. is_async drives error message text selection. NEW BRIDGE: hir_builder_emit_setup_with_common_c — out-param returns enter_result via void**. C primitives used: hir_c_create_load_attr_special_reg + hir_c_create_vectorcall_reg + hir_c_set_operand + hir_deopt_set_frame_state + hir_builder_temps_alloc_stack (all existing). C++ HIRBuilder::emitSetupWithCommon → 6-line bridge stub returning Register* via cast of out-param. Verification: cmake --build target jit/phoenix_jit clean (BUILD_EXIT=0). This is the chain-dep root for W27c python#2 (SetupWith) + python#3 (BeforeWith).

BEFORE_WITH/BEFORE_ASYNC_WITH (3.12+) + pre-3.12 BEFORE_ASYNC_WITH opcode handler. Per W27c Step A scope. C body: hir_builder_emit_before_with_c (~17 lines). Calls SetupWithCommon helper (W27c python#1) → push enter_result. Simpler than SetupWith — no SetupFinally call. PY_VERSION_HEX + opcode-branch conditional folded into C++ stub: - pre-3.12: __aenter__/__aexit__, is_async=1 - 3.12+ BEFORE_ASYNC_WITH: __aenter__/__aexit__, is_async=1 - 3.12+ BEFORE_WITH: __enter__/__exit__, is_async=0 ZERO new bridges (uses W27c python#1 helper). C++ HIRBuilder::emitBeforeWith → ~25-line stub (PY_VERSION_HEX + opcode branch + delegation). Verification: cmake --build target jit/phoenix_jit clean (BUILD_EXIT=0).

MATCH_CLASS (mainline) opcode handler. Per W27d Step A scope. C body: hir_builder_emit_match_class_c (~70 lines). Pops names/type/subject. Loads nargs as TCUInt64 const. Emits MatchClass → attrs_tuple. Refines to TOptTupleExact = TTupleExact|TNullptr (constructed via hir_type_union per theologian L2544 fallback — HIR_TYPE_OPTTUPLEEXACT not defined). Cond-branch: true_block refines to TTupleExact + (pre-3.12) loads Py_True; false_block emits CheckErrOccurred + (pre-3.12 Py_False+assign-subject; 3.12+ Py_None +assign-none). Converges at done. ZERO new bridges. C primitives: hir_c_create_load_const + hir_c_create_match_class_reg2 + hir_c_create_refine_type_reg + hir_c_create_cond_branch + hir_c_create_check_err_occurred_reg + hir_c_create_assign_reg + hir_c_create_branch_cpp + hir_type_from_cuint + hir_type_union + hir_type_from_object + hir_func_alloc_register + hir_cfg_alloc_block + hir_builder_temps_alloc_stack (all existing). C++ HIRBuilder::emitMatchClass → 4-line delegation stub. Verification: cmake --build target jit/phoenix_jit clean (BUILD_EXIT=0). Sole-path: MATCH_CLASS is mainline-tested per theologian L2544 coverage observation.

…ERO C++ SEQUENCE_GET (Static Python) opcode handler. FINAL emit-method conversion in W27 burndown. Per W27d Step A scope. C body: hir_builder_emit_sequence_get_c (~85 lines). Same shape as emitSequenceSet (W27c python#5) + SEQ_SUBSCR_UNCHECKED bit gating + emitLoadArrayItem (vs StoreArrayItem): - oparg & SEQ_SUBSCR_UNCHECKED set → skip CheckSequenceBounds, raw idx - else CheckSequenceBounds + adjusted_idx - Per (oparg & ~UNCHECKED): SEQ_LIST/INEXACT/CHECKED_LIST → load_field 'ob_item' at PyListObject offset + TCPtr; SEQ_ARRAY_INT64 → load_const offset + load_field_address - LoadArrayItem with element_type_from_seq_type(oparg) PHX_SEQ_SUBSCR_UNCHECKED = (1<<3) = 8 hard-coded inline. PHX_SEQ_* base values reused from W27c python#5 (PHX_SEQ_LIST/TUPLE/LIST_INEXACT/ CHECKED_LIST/ARRAY_INT64). ZERO new bridges. C primitives: hir_c_create_load_field_reg + hir_c_create_guard_is_reg + hir_c_create_refine_type_reg + hir_c_create_check_seq_bounds_reg + hir_c_create_load_field_address_reg + hir_c_create_load_array_item_reg + hir_c_create_load_const + hir_type_from_cint + hir_builder_temps_alloc_stack (all existing). C++ HIRBuilder::emitSequenceGet → 4-line delegation stub. Verification: cmake --build target jit/phoenix_jit clean (BUILD_EXIT=0). Sole-path: SEQUENCE_GET is __static__-only per accepted-residual L2531. 🎯 W27d CLOSURE — ZERO C++ ACHIEVED for emit-methods: W27d python#1: emitCopyFreeVars W27d python#2: emitGetYieldFromIter W27d python#3: emitMatchClass W27d python#4: emitMatchMappingSequence W27d python#5: emitSequenceGet (THIS) Pure-C++ count: 5 → 0 / 123 (100% per scripts/count_emit_methods.sh). Total W27 burndown: 20 → 0 across 4 batches (W27a/b/c/d). W27e residual paragraph (theologian L2533) attached to closure announcement to Alex (10 __static__-only methods accepted-residual per Phoenix-doesn't- expose-Static-Python framing).

Per supervisor L2565 + theologian L2569 ACK. Reduces C++ stub from ~13 lines to 8 lines (qualifies as STUB). Strategy: change bridge convention from caller-allocates-out to callee-allocates-out. C body of bridge now allocates 'out' internally via hir_builder_temps_alloc_stack. Both call sites (C++ stub at builder.cpp + C-body call from emit_any_call_c) updated to drop the explicit out alloc. Bridge sig change: void *out param removed (was 4th arg). Falsification discipline: change is mechanically validated by build (callers updated in same commit) — no signature drift surface beyond the tightening. Verification: cmake --build target jit/phoenix_jit clean (BUILD_EXIT=0). Pure-C++ count unchanged (still 0); PARTIAL count drops 9→8.

Per supervisor [chat L2626] + L2742 + librarian L2737 (close pythia python#87 python#3 ephemera-gap durably). Captures empirical W26 push 84 incident (95c9f9b → 1553c14 fix): - Both cinder_opcode_ids.h + Include/opcode.h use Py_OPCODE_H guard - Include order silently shadows whichever loads second - W26 case: BINARY_OP_ADD_INT undefined → attr_probe HIR regression caught by W21 golden trip-wire Three resolution options enumerated: - (A) rename cinder_opcode_ids.h guard to CINDERX_OPCODE_IDS_H — recommended, eliminates collision class - (B) static_assert in cinder_opcode.h to detect collision at compile — hardens detection without fixing - (C) header-comment only — current state, NOT recommended (chat- ephemera proven insufficient) Related W29 candidate noted: PHX_PRIM_OP_* / PHX_PRIM_UOP_* hard-coded in builder_emit_c.c lines 3727-3746 with no static_assert binding to classloader.h authoritative #defines (per pythia python#89 python#3 re-issue). DEFERRED per supervisor — schedule post-Batch-2 burndown, before any upstream sync touching Include/opcode.h or classloader.h.

…r.cpp Per supervisor 21:27:42Z + theologian 21:28:13Z methodology cross-check: 4 caller-search patterns (word-boundary + fn-pointer + Lib/test substring + broad), 1 caller (FORMAT_SIMPLE bytecode dispatch at builder.cpp:2457), argument-forwarding equivalence verified. Caller-rewrite: FORMAT_SIMPLE case now invokes hir_builder_emit_format_simple_c(&tc, current_func_, this) directly instead of going through the C++ method wrapper. Extern "C" decl hoisted to file-scope (builder.cpp:56) so the dispatch loop has visibility before the method's original location. Numstat (verified pre-commit per shepard 20:41:46Z + supervisor 20:16:25Z discipline): builder.cpp +4/-8 (NET -4; 3-line wrapper deleted + 4-line comment block deleted + 1-line extern decl hoisted to file scope + 4-line comment hoist NET -4 — see split below) builder.h +0/-1 (NET -1; method declaration removed) Bundle: -5L net, ZERO new bridges (existing C function + existing extern decl, just relocated). Pattern validated for Phase 1 batch-of-N scaling per supervisor 21:26:49Z 'Phase 1 python#3 will be batch-of-5+ aggressive across simple 1-caller methods'. Verified at HEAD post-fix: test_phoenix_jit_comparisons + controlflow + autocompile + partial_conversions: 369/369 PASS (no regression). W44 DO-NOT-USE caller gate: PASS (2 markers, 0 violations).

Per supervisor 21:39:13Z + theologian 21:39:06Z methodology cross-check PASS for batch-of-5: emitBuildCheckedList + emitBuildCheckedMap + emitSequenceGet + emitSequenceSet + emitGetYieldFromIter. All 5 had 1 caller in builder.cpp dispatch + 0 fn-pointer + 0 Lib/test substring (caller-search per supervisor 21:13:23Z scope: full repo + .py + Lib/test). All 5 were pure delegation stubs (no logic before C call). Pattern (validated by Phase 1 python#2 emitFormatSimple precedent): 1. Hoist extern "C" decl to file scope (builder.cpp:54-62) 2. Rewrite caller in dispatch switch to call C function directly 3. Delete C++ method wrapper + decl emitBeforeWith DROPPED from batch (25-line PY_VERSION_HEX-conditional logic violates 0-bridge constraint per theologian 21:39:06Z; defer to single-method batch later). emitGetYieldFromIter caller-rewrite preserves required casts per theologian warning 21:39:06Z: static_cast<int>(co_flags) + static_cast<void*>(&PyCoro_Type). Numstat (verified pre-commit per shepard 20:41:46Z + supervisor 20:16:25Z discipline): builder.cpp +14/-53 (NET -39; 5 wrapper bodies deleted + 5 extern decls migrated to file-scope comment block) builder.h +0/-13 (NET -13; 5 method declarations removed) Bundle: -52L net, ZERO new bridges (all 5 C functions + extern decls already existed; just relocated externs to file-scope). Verified at HEAD post-fix: test_phoenix_jit_comparisons + controlflow + autocompile + partial_conversions: 369/369 PASS (no regression). W44 DO-NOT-USE caller gate: PASS (2 markers, 0 violations). Phase 1 burndown advances: builder.cpp 4900→4859 (-41 lines beyond Phase 1 python#2 baseline). Pattern repeats cleanly for next batch on the remaining 1-caller / pure-delegation methods.

…oadSpecial/MatchClass/MatchMappingSequence/Send) Per supervisor 21:45:14Z + theologian 21:45:17Z methodology cross-check PASS for batch-of-5: emitCopyFreeVars + emitLoadSpecial + emitMatchClass + emitMatchMappingSequence + emitSend. Methods deleted (all bytecode-dispatch overloads, NOT the inline helpers in HIRBuilder class body that share the same name with Register*-based signatures — overload distinction verified by theologian on disk: helper at builder.cpp:710/807/859 vs dispatch at :4642/4679/4728). emitSetupWith DROPPED from batch (25-line PY_VERSION_HEX-conditional enter_id/exit_id logic = Cat-B per emitBeforeWith precedent in python#3). 6 caller-rewrites (multi-caller methods: emitCopyFreeVars × 2 at 1065+1878, emitMatchMappingSequence × 2 at 2408+2412; single-caller methods: emitLoadSpecial+emitMatchClass+emitSend × 1 each). Numstat (verified pre-commit per shepard 20:41:46Z + supervisor 20:16:25Z discipline): builder.cpp +18/-60 (NET -42; 5 wrapper bodies + 5 inline extern decls deleted, +18 from caller-rewrite expansions and 5 file-scope extern decls for the new C dispatch targets) builder.h +0/-12 (NET -12; 5 method declarations removed) Bundle: -54L net, ZERO new bridges (5 existing C functions reused via file-scope extern hoists; matches Phase 1 python#3 -52L precedent). Verified at HEAD post-fix: test_phoenix_jit_comparisons + controlflow + autocompile + partial_conversions: 369/369 PASS (no regression). W44 DO-NOT-USE caller gate: PASS. Phase 1 cumulative burndown: -35 (python#1+python#1.5) + -5 (python#2) + -52 (python#3) + -54 (python#4) = -146L total. builder.cpp 4859→4817.

Per supervisor 21:54:07Z + theologian 21:53:58Z methodology cross-check PASS for batch-of-6: emitPrimitiveLoadConst + emitPrimitiveBox + emitPrimitiveUnbox + emitPrimitiveBinaryOp + emitPrimitiveCompare + emitPrimitiveUnaryOp. All 6 had 1 caller each in builder.cpp dispatch (lines 2005/2009/2013/ 2017/2021/2025) + Lib/test/ scope EMPIRICALLY VERIFIED clean per theologian 21:53:58Z (grep -E 'emitPrimitive(LoadConst|Box|Unbox| BinaryOp|Compare|UnaryOp)' Lib/test/test_phoenix_*.py = 0 references). Methods deleted (all bytecode-dispatch overloads). Inline helpers in HIRBuilder class body with Register*-based signatures (lines 944, 733, 456, 978) STAY — distinct overloads. tc.emitPrimitiveBox / tc.emitPrimitiveUnbox at lines 3875/3884 are TranslationContext class methods (different class) — NOT affected. 6 caller-rewrites (lines 2005, 2009, 2013, 2017, 2021, 2025) at bytecode dispatch. Numstat (verified pre-commit per shepard 20:41:46Z + supervisor 20:16:25Z discipline + supervisor 21:54:07Z 'commit-message count = 6 caller-rewrites avoid python#4 annotation typo'): builder.cpp +14/-62 (NET -48; 6 wrapper bodies + 6 inline extern decls deleted, +14 from caller-rewrite expansions and 6 file-scope extern decl hoists) builder.h +0/-18 (NET -18; 6 method declarations removed, 3 lines per declaration) Bundle: -66L net, ZERO new bridges (6 existing C functions reused via file-scope extern hoists). Verified at HEAD post-fix: test_phoenix_jit_comparisons + controlflow + autocompile + partial_conversions: 369/369 PASS (no regression). W44 DO-NOT-USE caller gate: PASS (2 markers, 0 violations). Phase 1 cumulative burndown: -35 (python#1+python#1.5) + -5 (python#2) + -52 (python#3) + -54 (python#4) + -66 (python#5) = -212L total. builder.cpp 4817→4769.

…nix.sh Adds Step 1g per spec §2.7.3 python#3, invoking scripts/w45_section_3_5_derivation_drift.sh --strict after Step 1e (W44 DO-NOT-USE caller gate). Step 1f reserved for future W45 §1-§2 bridge-sig falsifier integration. Pattern mirrors Step 1e (W44): capture output, exit code, GATE_PASS flag, FAILURES annotation. ~15L diff. Authorization: supervisor 00:22:51Z post-push priority (1) per spec §2.7.3 python#3 integration step; ~10L estimate (actual 15L for output capture + failure annotation parity with Step 1e).

Theologian 01:53:41Z draft per shepard 01:53:07Z + pythia python#104 (3) + supervisor 01:53:55Z GO. Adds proactive complement to existing gate python#3 (HEAD==binary AND tree-clean during build, REACTIVE detection at build-symptom). 4-step discipline for multi-file (≥2) edit sequences: 1. Pre-edit baseline snapshot 2. Mid-edit integrity check (HALT on unexpected change) 3. Pre-commit verification (HALT if staged diff diverges) 4. HALT response (clean-HEAD restore + observed-but-not-attributed report; do NOT auto-resume) Rationale: 5+ prior incidents caused gate failures + false BUILD PASS + ARM64 build BLOCK (D-1775810621, D-1775669703, D-1776414469, D-1776434533, D-1776887480/D-1776890644 per librarian 01:46:19Z). Triggered by 01:23Z external file-state revert during Tier 8 pilot Phase A execution (generalist 01:24:43Z observed-but-not-attributed HALT + clean-HEAD restore demonstrated the prescribed response). Doc-only +14L. No §3.5 BUILD MODE per touched-files rule (no builder*.{cpp,h,c} touched).

…xtension Per supervisor 02:35:05Z (PIR redirect to theologian) + 02:37:20Z (GO atomic commit + Tier 8 Phase A resume directive). Theologian 02:36:49Z PIR delivery + 4-section CLAUDE.md amendment. PIR scope (docs/2026-04-24-pre-edit-revert-pir.md, 132L NEW): (1) Automated-detection feasibility — 3 options analyzed: (a) inotify file-watcher daemon: feasible but multi-session NBS-suite extension; DEFERRED to W48 (b) pre-commit git hook: REJECTED (duplicates gate python#3) (c) agent-side mtime-checkpoint discipline: ADOPTED (~5L per agent edit-loop; in-session feasible) (2) Per-incident root-cause attribution for 5 priors + this incident: D-1775810621 / D-1775669703 / D-1776414469 — SYMPTOM-only D-1776434533 — DIRECTIVE-only (Alex 'always full commit checkouts') D-1776887480/D-1776890644 — GATE-hardening (gate python#3 added) 2026-04-24T01:23Z — SYMPTOM-only (observed-but-not-attributed) TOTAL: 6 incidents, 0 root-cause attributions (3) Class CONFIRMED 'undiagnosed-recurring' (already labeled in CLAUDE.md amendment e8a83df per pythia python#105) CLAUDE.md mtime-checkpoint extension (+1L net, rule 1 enhancement): Pre-edit baseline now also captures per-touched-file mtime via `stat -c %Y %n`; verify before each subsequent write within edit sequence. mtime mismatch = external revert; HALT. Mechanizes detection vs requiring agent to notice mid-edit. Bundle scope: doc-only (PIR file + CLAUDE.md edit), no §3.5 BUILD MODE per touched-files rule. Atomic per supervisor 02:37:20Z. Phase 3 closure-amendment (docs/tier7-phase3-closure-summary.md) remains STAGED on disk per theologian 02:36:49Z + supervisor 02:37:20Z ('RETRACT only on successful Phase A resume push'). NOT committed this commit. Tier 8 Phase A resume armed post-push 29 per supervisor 02:37:20Z GO (R-retry) — PIR + mtime-checkpoint discipline mechanizes recurrence detection; don't infinitely block on Alex when discipline armed.

…amendment Per supervisor 02:48:39Z atomic doc-only bundle. Closes the 6→7 escalation triggered by Tier 8 pilot Phase A (R-retry) recurring revert at incident python#7. W48 spec (docs/w48-nbs-inotify-file-watcher-spec.md, +194L NEW): Theologian 02:48:25Z spec per W42/W44/W45 spec pattern. Elevates PIR §2.1 option (a) inotify file-watcher from DEFERRED to ACTIVE workstream. Mechanism: pyinotify daemon as NBS-suite sidecar; narrow watch per agent edit-set; event correlation with agent Write tool calls; forensic snapshot for root-cause attribution (closes pythia python#105 (3) zero-attribution gap). 4-escalation history honest framing: Alex directive D-1776434533 → gate python#3 D-1776887480 → 4-step discipline 03a0dcb → mtime-checkpoint 48339e2 → W48. Each progressively detective, none preventive. Cost ~3-5 sessions multi-session NBS-suite extension. Acceptance: 7 items including CLAUDE.md rule 5 addition + Tier 8 pilot retry under W48. PIR amendment (docs/2026-04-24-pre-edit-revert-pir.md, +21/-2): - Incident table: appended incident python#7 (2026-04-24T02:45Z, same pattern as python#6, mtime-checkpoint detected but did not prevent) - Conclusion: 7 incidents (was 6), 0 root cause attributions - NEW 'Incident python#7 ESCALATION' subsection: documents (R-retry) reasoning empirically INVALIDATED + pythia python#105 'fever has name' validated twice - NEW 'W48 ESCALATION' subsection: cross-links W48 spec landing + Tier 8 pilot Phase A retry GATED on W48 landing Doc-only bundle, no §3.5 BUILD MODE per touched-files rule. Closes the post-incident-python#7 reactive cycle: closure-amendment shipped at push 30 (f6328cd), W48 + PIR ship at push 31. Authorization: theologian 02:48:25Z (W48 spec) + supervisor 02:48:39Z (atomic bundle GO).

ROOT CAUSE FIX for the 8-incident undiagnosed-recurring file-state revert class (D-1776998268 + 7 priors). Per supervisor 04:14:27Z post-incident-python#8 ROOT CAUSE IDENTIFICATION: the §3.5 falsifier's restore_files() trap was the "external reverter" across all 8 incidents. It ran `git checkout HEAD --` on TOUCHED_FILES on script exit; if those files had unstaged modifications BEFORE the script ran (e.g., agent's in-flight Tier 8 Phase A content), the trap blew them away. Empirical validation (this session 04:13:36Z): - Test python#6 + python#7 (trivial content, single-mechanism): no revert - Tests python#5/python#7 idle: no revert - cmake build alone: no revert - §3.5 invocation on staged Phase A: REVERT (PhxExceptionTable count 13 → 0; mtime bumped 1777002918 → 1777003987; builder_state_c.h no longer M in git status) 8 prior incidents now retroactively attributed to §3.5 trap firing during build/gate cycles concurrent with agent Tier 8 attempts. NO external actor (no VS Code, no Alex, no formatter, no cron). W48 inotify daemon NOT NEEDED. (Iso-A) host relocation NOT NEEDED. Class size CLOSED — class type changed from "undiagnosed-recurring" to "self-inflicted via §3.5 trap, RESOLVED". Fix mechanism: snapshot per-file content to /tmp on first snapshot_file_if_new() call (per fixture mutation start). restore_files() copies from snapshot (not HEAD), preserving pre-script unstaged state. Snapshots cleaned + FILE_SNAPSHOTS array unset post-restore. Re-validation (this session post-fix): - §3.5 BUILD MODE 4/4 PASS - post-restore build [OK] - Phase A content (PhxExceptionTable count 13) SURVIVED §3.5 invocation - ALL 6 Tier 8 staged files still M Doc-only (script-only) push, no §3.5 BUILD MODE per touched-files rule (this IS the §3.5 script being amended; meta-circular but doc-only gate appropriate). Authorization: supervisor 04:14:27Z FIX directive post-ROOT-CAUSE identification (generalist 04:13:58Z catch). Pythia python#105 'fever has name infection still spreads' was structurally correct re self-inflicted infrastructure — but the infection wasn't external-class undiagnosable; it was OUR OWN script's exit trap trampling unstaged work. 5 detection layers (Alex directive D-1776434533, gate python#3 D-1776887480, 4-step Pre-Edit WT Integrity D-1776995670, mtime-checkpoint D-1776998268, W48 spec D-1776999077) added to detect ourselves. Audit own scripts BEFORE external-actor hypotheses (per new feedback memory entry).

the-knights-who-say-ni added the CLA not signed label Feb 10, 2017

habnabit force-pushed the fix-mock-builtin-methods-for-cython branch from c8bc9c9 to 9148028 Compare February 11, 2017 01:47

habnabit force-pushed the fix-mock-builtin-methods-for-cython branch 2 times, most recently from e12a088 to 97bb733 Compare February 11, 2017 06:25

berkerpeksag added backport to 3.5 labels Feb 11, 2017

Mariatta removed the CLA not signed label Feb 15, 2017

the-knights-who-say-ni added the CLA signed label Feb 15, 2017

berkerpeksag reviewed Feb 22, 2017

View reviewed changes

berkerpeksag requested a review from voidspace February 22, 2017 04:57

habnabit force-pushed the fix-mock-builtin-methods-for-cython branch from 97bb733 to c5b3653 Compare February 23, 2017 22:40

auvipy approved these changes Mar 24, 2017

View reviewed changes

habnabit and others added 2 commits July 20, 2017 02:33

Add blurb entry

989e99a

berkerpeksag force-pushed the fix-mock-builtin-methods-for-cython branch from 75ad0f4 to 989e99a Compare July 19, 2017 23:35

Move test to correct test case

4f4ed02

berkerpeksag approved these changes Jul 19, 2017

View reviewed changes

berkerpeksag merged commit 856cbcc into python:master Jul 20, 2017

serhiy-storchaka mentioned this pull request Jun 30, 2026

Inconsistency between quopri.decodestring(), email.quoprimime.decode() and binascii.a2b_qp() #62222

Open

This was referenced Jul 18, 2026

free threading: struct iter_unpack is not memory safe #154013

Closed

use-after-free: functools partial #154189

Closed

serhiy-storchaka mentioned this pull request Jul 23, 2026

SocketHandler silently drops log messages on re-connect #84532

Open

chongwick mentioned this pull request Aug 19, 2026

segv in ceval #156074

Closed

gabyfle mentioned this pull request Sep 1, 2026

[3.13] gh-132396: Resolve 'redefinition of unused name' errors in Lib/test/ (GH-132397) #132699

Merged

This was referenced Sep 11, 2026

Heap-buffer-overflow in the hz incremental decoder on a trailing ~ #157325

Open

mmap.mmap.__setitem__ crashes when concurrently resized #157335

Open

Heap UAF in io.TextIOWrapper read path via reentrant detach() #157363

Open

Uh oh!

Uh oh!

Conversation

habnabit commented Feb 10, 2017

Uh oh!

the-knights-who-say-ni commented Feb 10, 2017

Uh oh!

habnabit commented Feb 10, 2017

Uh oh!

ShalbafZadeh commented Feb 10, 2017

Uh oh!

habnabit commented Feb 10, 2017

Uh oh!

ShalbafZadeh commented Feb 10, 2017

Uh oh!

habnabit commented Feb 11, 2017

Uh oh!

habnabit commented Feb 11, 2017

Uh oh!

codecov Bot commented Feb 11, 2017

Codecov Report

Uh oh!

berkerpeksag left a comment

Choose a reason for hiding this comment

Uh oh!

berkerpeksag commented Feb 22, 2017

Uh oh!

habnabit commented Feb 23, 2017

Uh oh!

Yhg1s commented May 22, 2017

Uh oh!

buhtz commented Jul 19, 2017

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

10 participants