models.fields.IntegerField.to_python() is converting float type numbers, making .clean_fields( ) less reliable

[thiper01@…]

During the full_clean( ) routine, when an IntegerField is being cleaned and the code calls the field's to_python( ) method, if the value of the field is a float, it will just be converted to an integer using int(value). As it is characteristic of python's native int( ) function, it converts any float to int regardless of decimal places and approximations necessary, allowing for examples like this to happen:

exModel = ModelWithIntegerField(value=1.2)
>>   exModel.value     #1.2
exModel.clean_fields() #no validation error raised
>>   exModel.value     #1

I don't know if this behaviour is because of ticket #24229, but I believe there are better ways than just allowing any number through IntegerField verification as long as it's a number, regardless of how it affects the output.

My suggestions to improve this are:
1# Check first if it's a float type then verify if the numbers after the decimal point are 0 if value%1 == 0, if they are, no problem converting, if not raise a TypeError as it would distort the original value without warning.

[thiper01@…]

​https://github.com/django/django/pull/18033

[David Sanders]

Hi there,

The issue tracker is for reporting bugs with Django, whereas this request is change in known behaviour that folks may be relying upon. Please start a discussion on the Django forum if you strongly feel this needs to change: ​https://www.djangoproject.com/community/ If the Django community agrees then this ticket can be reopened.

Closing per https://code.djangoproject.com/wiki/DevelopersMailingList#BeenWONTFIXed