1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
|
From 1af5b2c34ed4794ae925aa2761d73a816ff21bd2 Mon Sep 17 00:00:00 2001
From: Stuart Archibald <stuartarchibald@users.noreply.github.com>
Origin: https://github.com/numba/numba/pull/8639
Date: Tue, 29 Nov 2022 11:26:38 +0000
Subject: [PATCH 16/20] Fix bytecode iteration to handle EXTENDED_ARG as jump
target.
As title.
---
numba/core/bytecode.py | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/numba/core/bytecode.py b/numba/core/bytecode.py
index 4f5d449dd04..6305751d424 100644
--- a/numba/core/bytecode.py
+++ b/numba/core/bytecode.py
@@ -155,7 +155,15 @@ def _unpack_opargs(code):
arg |= code[i + j] << (8 * j)
i += ARG_LEN
if op == EXTENDED_ARG:
+ # This is a deviation from what dis does...
+ # In python 3.11 it seems like EXTENDED_ARGs appear more often
+ # and are also used as jump targets. So as to not have to do
+ # "book keeping" for where EXTENDED_ARGs have been "skipped"
+ # they are replaced with NOPs so as to provide a legal jump
+ # target and also ensure that the bytecode offsets are correct.
+ yield (offset, OPCODE_NOP, arg, i)
extended_arg = arg << 8 * ARG_LEN
+ offset = i
continue
else:
arg = None
|
