struct.pack error messages are misleading and inconsistent #98248
Description
Bug report
1. Misleading error message
>>> import struct
>>> struct.pack(">Q", -1)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
struct.error: int too large to convert
I don't think -1 is that large to convert to ulonglong, so the error message is wrong. The problem is that -1 is not in the range of ulonglong. The current error message is not helpful for users to debug.
Compared to other error messages:
Code
import struct
for endianness in "<>":
for size in "BHILQ":
try:
fmt = endianness + size
struct.pack(fmt, -1)
except struct.error as e:
print("Error msg of " + fmt + ":", e)
stdout
Error msg of <B: ubyte format requires 0 <= number <= 255
Error msg of <H: ushort format requires 0 <= number <= 65535
Error msg of <I: argument out of range
Error msg of <L: argument out of range
Error msg of <Q: argument out of range
Error msg of >B: ubyte format requires 0 <= number <= 255
Error msg of >H: argument out of range
Error msg of >I: argument out of range
Error msg of >L: argument out of range
Error msg of >Q: int too large to convert
2. Inconsistent error messages when packing into different integral types
See the output above.
A possible solution
I can create a PR to fix the 1st problem. For the 2nd problem, #28178 (comment) and #89197 (comment) said that the inconsistency can be fixed, so I can probably fix this in the same PR.
Your environment
- CPython versions tested on: Python 3.12.0a0 (heads/main:ccab67b, Oct 12 2022, 15:25:20) [GCC 12.2.0] on linux
- Operating system and architecture: Arch Linux 5.19.13.arch1-1, x86-64
- Native endianness: Little-endian
Activity
gh-98248: Normalizing the error messages in function struct.pack (GH-…