|
# TODO: Append (sqlstate) in the error message. This will be come in next minor release. |
Looks like the code here was actually mostly done up which I'm thankful for, I've been using a patch which mostly breaks the exception format but hands over the state and error message back to the software code.
It really looks like its just actually putting in the args to the error in a way that makes sense would have been the issue (projecting), but formally fixing this issue would be desirable to me rather than making a mess of the library code. This would have been an easy PR but I really can't think of a solution that makes sense either, thus my own work is pretty sloppy since I'm only writing it for myself
Its bad
def raise_mysql_exception(data):
errno = struct.unpack("<h", data[1:3])[0]
# https://dev.mysql.com/doc/dev/mysql-server/latest/page_protocol_basic_err_packet.html
# Error packet has optional sqlstate that is 5 bytes and starts with '#'.
errorclass = error_map.get(errno)
if errorclass is None:
errorclass = InternalError if errno < 1000 else OperationalError
err = errorclass(errno)
if data[3] == 0x23: # '#'
# sqlstate = data[4:9].decode()
# TODO: Append (sqlstate) in the error message. This will be come in next minor release.
err.sqlstate = data[4:9].decode("utf-8", "replace")
err.errval = data[9:].decode("utf-8", "replace")
else:
err.errval = data[3:].decode("utf-8", "replace")
raise err-> breakpoint()
(Pdb) e
OperationalError(1644)
(Pdb) e.args
(1644,)
(Pdb) e.__dict__
{'sqlstate': '45000', 'errval': 'Record Delete Blocked'}
(Pdb) e.sqlstate
'45000'
Removing the error message from the args here isn't ideal, i'm aware 😅
PyMySQL/pymysql/err.py
Line 143 in d7bb777
Looks like the code here was actually mostly done up which I'm thankful for, I've been using a patch which mostly breaks the exception format but hands over the state and error message back to the software code.
It really looks like its just actually putting in the args to the error in a way that makes sense would have been the issue (projecting), but formally fixing this issue would be desirable to me rather than making a mess of the library code. This would have been an easy PR but I really can't think of a solution that makes sense either, thus my own work is pretty sloppy since I'm only writing it for myself
Its bad
Removing the error message from the args here isn't ideal, i'm aware 😅