The method _parseInteger of StdDeserializer

[wenzis]

The method _parseInteger of StdDeserializer auto handle JsonToken.VALUE_NUMBER_FLOAT to JsonToken.VALUE_NUMBER_INT,could you support enum to prohibit this feature?
_parseShort、_parseIntPrimitive and so on.

[cowtowncoder]

Just to make sure I understand: you would NOT want a floating point number (like 2.05) to be used for Enum values (rounded as 2)? If so, that should be doable yes.

[wenzis]

Sorry, I didn't say it clearly.
Let me to describe.

first: I have a class
public class Site {
private Integer type
public Integer getType() {
return type;
}
public void setType(Integer type) {
this.type = type;
}
}

Second, there is a json
String json = "{"type":1.2}";

Third: data conversion
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
Site site = OBJECT_MAPPER.readValue(json, Site.class);

the result is successful and I get the site, the value of site type is 1,but I don't think so,
exactly I want a format exception. should I do?

[cowtowncoder]

Different users want different things: many prefer to try very hard to match data; and since (for example) floating point numbers are still numbers, rounding to closest or truncating may make sense in cases where there is mismatch.

So if your specific desire is NOT to allow coercion of floating point numbers to integer numbers, it would make sense to have a new DeserializationFeature, which determines whether such coercion is allowed. This should then be applied to deserialization of various things.

It sounds like this is what you want, then. I don't know when I'll have a chance to work on it, but it will probably not make it in 2.5 yet, but 2.6.

[wenzis]

@cowtowncoder Thanks!

[cowtowncoder]

Added DeserializationFeature.ACCEPT_FLOAT_AS_INT, disabling of which prevents coercion from JSON (etc) floating-point numbers into int/long/BigInteger.

[Skarafaz]

Just noticed this crazy behaviour... no words! Having DeserializationFeature.ACCEPT_FLOAT_AS_INT = true as default is simply insane. Truncating a float means loosing information, so it MUST NOT be the default.

[cowtowncoder]

@Skarafaz you have right to voice your opinion. And I am free to disregard it.
If you want it changed, GO FUCKING CHANGE IT on your set up. Behavior has been here since beginning of the project and I am annoyed by people like you butting in and DEMANDING FOR IMMEDIATE CHANGE. Just go away.