Allow T_BOOL in PyMemberDef definitions

[lillo]

BPO	1720595
Nosy	@loewis, @birkenfeld, @pitrou, @tiran
Files	Python-2.5_T_BOOL.diff: The patch built against a plain Python 2.5 installation

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2008-01-21.21:24:22.475>
created_at = <Date 2007-05-17.09:50:16.000>
labels = ['interpreter-core', 'type-feature']
title = 'Allow T_BOOL in PyMemberDef definitions'
updated_at = <Date 2008-01-21.21:24:22.473>
user = 'https://bugs.python.org/lillo'

bugs.python.org fields:

activity = <Date 2008-01-21.21:24:22.473>
actor = 'georg.brandl'
assignee = 'none'
closed = True
closed_date = <Date 2008-01-21.21:24:22.475>
closer = 'georg.brandl'
components = ['Interpreter Core']
creation = <Date 2007-05-17.09:50:16.000>
creator = 'lillo'
dependencies = []
files = ['7995']
hgrepos = []
issue_num = 1720595
keywords = ['patch']
message_count = 8.0
messages = ['52637', '52638', '59795', '59807', '61362', '61413', '61448', '61450']
nosy_count = 5.0
nosy_names = ['loewis', 'georg.brandl', 'pitrou', 'christian.heimes', 'lillo']
pr_nums = []
priority = 'normal'
resolution = 'accepted'
stage = None
status = 'closed'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue1720595'
versions = ['Python 2.6']

[lillo]

Hello

The attached patch allows to use T_BOOL alongside other common types in the PyMemberDef definitions of your custom types.
The boolean value is assumed to be held in a straight C char variable, and PyMember_SetOne() sets it to 1 or 0 depending on the result of calling PyObject_IsTrue() on the passed value object.

Usage example:

typedef struct ButtonObject
{
	PyObject_HEAD
	char		fToggled;
} ButtonObject;

static PyMemberDef eb_ButtonEventMembers[] = {

{ "toggled", T_BOOL, offsetof(ButtonObject, fToggled), RO, "True if the button is currently toggled" },
{ NULL, 0, 0, 0, NULL }

};

The patch has been built against a plain Python 2.5 installation.

[birkenfeld]

I'd rather not allow any object to be assigned to the attribute, but only a bool. T_INT, for example, doesn't automatically call int() on its values too.

[tiran]

allow or reject it for 2.6, Georg?

[birkenfeld]

The question is also what C type to assume for boolean fields -- char or
int?

[pitrou]

It seems to me that, when a boolean is to be represented, most C
programmers would expect a char rather than an int. Also an int would
make the object larger without any benefit.

As for type checking, Angelo's patch already does a PyBool_Check against
the argument. Is anything more required?

[birkenfeld]

Martin - is a char struct member for T_BOOL fine with you?

[loewis]

I can't see anything wrong with the feature. Of course, the patch itself
lacks changes to the test suite.

Ideally, there would also be changes to the documentation, but as
PyMemberDef seems to be undocumented currently, it's unfair to ask for
documentation with this patch. If documentation was made, it should list
for each T_ constant what the corresponding C type should be.

I think the cast to long (in PyBool_FromLong) is redundant.

[birkenfeld]

Okay, added tests, removed the cast and committed in r60181. If I find
time, I'll add docs for PyMemberDef in the next few days.