python3.12/00323-static-analysis-fixes.patch

From 833c91e1cb9219937c15b6c913e540e7aae07168 Mon Sep 17 00:00:00 2001
From: "Miss Islington (bot)"
 <31488909+miss-islington@users.noreply.github.com>
Date: Wed, 12 Jun 2024 15:57:45 +0200
Subject: [PATCH 1/5] gh-120155: Add assertion to sre.c match_getindex()
 (GH-120402) (#120410)

gh-120155: Add assertion to sre.c match_getindex() (GH-120402)

Add an assertion to help static analyzers to detect that i*2 cannot
overflow.
(cherry picked from commit 42b25dd61ff3593795c4cc2ffe876ab766098b24)

Co-authored-by: Victor Stinner <vstinner@python.org>
---
 Modules/_sre/sre.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/Modules/_sre/sre.c b/Modules/_sre/sre.c
index 0547390..59a5b11 100644
--- a/Modules/_sre/sre.c
+++ b/Modules/_sre/sre.c
@@ -2166,6 +2166,8 @@ match_getindex(MatchObject* self, PyObject* index)
         return -1;
     }
 
+    // Check that i*2 cannot overflow to make static analyzers happy
+    assert(i <= SRE_MAXGROUPS);
     return i;
 }
 
-- 
2.45.2


From eedc88971887bea9f37eab8bb8a1af6f650f88d8 Mon Sep 17 00:00:00 2001
From: "Miss Islington (bot)"
 <31488909+miss-islington@users.noreply.github.com>
Date: Fri, 7 Jun 2024 18:26:03 +0200
Subject: [PATCH 2/5] gh-120155: Fix copy/paste error in
 HAVE_SUBOFFSETS_IN_LAST_DIM() (GH-120228) (#120240)

gh-120155: Fix copy/paste error in HAVE_SUBOFFSETS_IN_LAST_DIM() (GH-120228)

Don't hardcode 'dest' in HAVE_SUBOFFSETS_IN_LAST_DIM() macro of
memoryobject.c, but use its 'view' parameter instead.

Fix the Coverity issue:

Error: COPY_PASTE_ERROR (CWE-398):
Python-3.12.2/Objects/memoryobject.c:273:14: original: ""dest->suboffsets + (dest->ndim - 1)"" looks like the original copy.
Python-3.12.2/Objects/memoryobject.c:274:14: copy_paste_error: ""dest"" in ""src->suboffsets + (dest->ndim - 1)"" looks like a copy-paste error.
Python-3.12.2/Objects/memoryobject.c:274:14: remediation: Should it say ""src"" instead?
GH-  272|       assert(dest->ndim > 0 && src->ndim > 0);
GH-  273|       return (!HAVE_SUBOFFSETS_IN_LAST_DIM(dest) &&
GH-  274|->             !HAVE_SUBOFFSETS_IN_LAST_DIM(src) &&
GH-  275|               dest->strides[dest->ndim-1] == dest->itemsize &&
GH-  276|               src->strides[src->ndim-1] == src->itemsize);
(cherry picked from commit 90b75405260467814c93738a3325645918d4ea51)

Co-authored-by: Victor Stinner <vstinner@python.org>
---
 Objects/memoryobject.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/Objects/memoryobject.c b/Objects/memoryobject.c
index b016804..3c88859 100644
--- a/Objects/memoryobject.c
+++ b/Objects/memoryobject.c
@@ -264,7 +264,7 @@ PyTypeObject _PyManagedBuffer_Type = {
 /* Assumptions: ndim >= 1. The macro tests for a corner case that should
    perhaps be explicitly forbidden in the PEP. */
 #define HAVE_SUBOFFSETS_IN_LAST_DIM(view) \
-    (view->suboffsets && view->suboffsets[dest->ndim-1] >= 0)
+    (view->suboffsets && view->suboffsets[view->ndim-1] >= 0)
 
 static inline int
 last_dim_is_contiguous(const Py_buffer *dest, const Py_buffer *src)
-- 
2.45.2


From d8eaaecbc2823a5d093248467a13111c1f821889 Mon Sep 17 00:00:00 2001
From: Victor Stinner <vstinner@python.org>
Date: Fri, 7 Jun 2024 17:17:06 +0200
Subject: [PATCH 3/5] gh-120155: Fix optimize_and_assemble_code_unit() error
 handling (#120231)

gh-120155: Fix optimize_and_assemble_code_unit() error handling

Don't use 'g' before it's being initialized: don't use the 'error'
label if consts_dict_keys_inorder() failed.

Fix the Coverity issue:

Error: UNINIT (CWE-457):
Python-3.12.2/Python/compile.c:7670:5: skipped_decl: Jumping over declaration of ""g"".
Python-3.12.2/Python/compile.c:7714:5: uninit_use_in_call: Using uninitialized value ""g.g_block_list"" when calling ""_PyCfgBuilder_Fini"".
Python-3.12.2/Python/compile.c:7714:5: uninit_use_in_call: Using uninitialized value ""g.g_entryblock"" when calling ""_PyCfgBuilder_Fini"".
  7712|       Py_XDECREF(consts);
  7713|       instr_sequence_fini(&optimized_instrs);
  7714|->     _PyCfgBuilder_Fini(&g);
  7715|       return co;
  7716|   }
---
 Python/compile.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/Python/compile.c b/Python/compile.c
index 40335f6..0cd8d60 100644
--- a/Python/compile.c
+++ b/Python/compile.c
@@ -7682,7 +7682,7 @@ optimize_and_assemble_code_unit(struct compiler_unit *u, PyObject *const_cache,
     PyCodeObject *co = NULL;
     PyObject *consts = consts_dict_keys_inorder(u->u_metadata.u_consts);
     if (consts == NULL) {
-        goto error;
+        return NULL;
     }
     cfg_builder g;
     if (instr_sequence_to_cfg(&u->u_instr_sequence, &g) < 0) {
-- 
2.45.2


From 85f555bd5d59cf5a29c8f7ef410b0f78a0b839cc Mon Sep 17 00:00:00 2001
From: "Miss Islington (bot)"
 <31488909+miss-islington@users.noreply.github.com>
Date: Mon, 10 Jun 2024 12:12:25 +0200
Subject: [PATCH 4/5] gh-120155: Fix Coverity issue in zoneinfo load_data()
 (GH-120232) (#120311)

gh-120155: Fix Coverity issue in zoneinfo load_data() (GH-120232)

Declare the 'rv' varaible at the top of the load_data() function to
make sure that it's initialized before the first 'goto error' which
uses 'rv' (return rv).

Fix the Coverity issue:

Error: UNINIT (CWE-457):
Python-3.12.2/Modules/_zoneinfo.c:1233:5: skipped_decl: Jumping over declaration of ""rv"".
Python-3.12.2/Modules/_zoneinfo.c:1284:5: uninit_use: Using uninitialized value ""rv"".
  1282|       }
  1283|
  1284|->     return rv;
  1285|   }
  1286|
(cherry picked from commit b90bd3e5bbc136f53b24ee791824acd6b17e0d42)

Co-authored-by: Victor Stinner <vstinner@python.org>
---
 Modules/_zoneinfo.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/Modules/_zoneinfo.c b/Modules/_zoneinfo.c
index 8fc8616..47e4012 100644
--- a/Modules/_zoneinfo.c
+++ b/Modules/_zoneinfo.c
@@ -954,6 +954,7 @@ end:
 static int
 load_data(zoneinfo_state *state, PyZoneInfo_ZoneInfo *self, PyObject *file_obj)
 {
+    int rv = 0;
     PyObject *data_tuple = NULL;
 
     long *utcoff = NULL;
@@ -1230,7 +1231,6 @@ load_data(zoneinfo_state *state, PyZoneInfo_ZoneInfo *self, PyObject *file_obj)
         }
     }
 
-    int rv = 0;
     goto cleanup;
 error:
     // These resources only need to be freed if we have failed, if we succeed
-- 
2.45.2


From 94346cac4ee78d624b0dbe9d67f5dc2a03fa2a4f Mon Sep 17 00:00:00 2001
From: "Miss Islington (bot)"
 <31488909+miss-islington@users.noreply.github.com>
Date: Tue, 25 Jun 2024 19:40:08 +0200
Subject: [PATCH 5/5] gh-120155: Fix Coverity issue in parse_string()
 (GH-120997) (#121006)

gh-120155: Fix Coverity issue in parse_string() (GH-120997)
(cherry picked from commit 769aea332940f03c3e5b1ad9badd6635c1ac992a)

Co-authored-by: Victor Stinner <vstinner@python.org>
---
 Parser/string_parser.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/Parser/string_parser.c b/Parser/string_parser.c
index 65c320c..164f715 100644
--- a/Parser/string_parser.c
+++ b/Parser/string_parser.c
@@ -226,9 +226,14 @@ _PyPegen_parse_string(Parser *p, Token *t)
         PyErr_BadInternalCall();
         return NULL;
     }
+
     /* Skip the leading quote char. */
     s++;
     len = strlen(s);
+    // gh-120155: 's' contains at least the trailing quote,
+    // so the code '--len' below is safe.
+    assert(len >= 1);
+
     if (len > INT_MAX) {
         PyErr_SetString(PyExc_OverflowError, "string to parse is too long");
         return NULL;
-- 
2.45.2