wrong results when IF function along with float argument

[jingyugao]

Bug Report

Please answer these questions before submitting your issue. Thanks!

1. What did you do?
   If possible, provide a recipe for reproducing the error.
   SELECT IF(0.1, 1, 2) FROM t1;

2. What did you expect to see?
   +---------------+
   | IF(0.1, 1, 2) |
   +---------------+
   | 1 |
   +---------------+
   1 row in set (0.00 sec)

3. What did you see instead?
   +---------------+
   | IF(0.1, 1, 2) |
   +---------------+
   | 2 |
   +---------------+
   1 row in set (0.00 sec)

4. What version of TiDB are you using (tidb-server -V or run select tidb_version(); on TiDB)?
   Release Version: None
   Git Commit Hash: None
   Git Branch: None
   UTC Build Time: None
   GoVersion: None
   Race Enabled: false
   TiKV Min Version: 2.1.0-alpha.1-ff3dd160846b7d1aed9079c389fc188f7f5ea13e
   Check Table Before Drop: false

[jingyugao]

In tidb, almost all builtIn*Sig function use the result of evalInt to check true or false.
This is not ok when args is float64 which is less than 1, it will be convert to 0 by evalInt.
So we should replace evalInt with evalReal.(By the way,if the argis to small and is out of the float64 accuracy,such as 1e-1000, it should be converted to 0)

[zz-jason]

evalInt means this expression signature should return an int result. For this case, we should implement evalInt. But in order to get the correct result, we should calling args[0].EvalReal() instead of args[0].EvalInt().

[jingyugao]

evalInt means this expression signature should return an int result. For this case, we should implement evalInt. But in order to get the correct result, we should calling args[0].EvalReal() instead of args[0].EvalInt().

Yes,I have fix the logic op #11587 many builtinfunc doesnt implement evalReal, so it may be difficult to handle.

[SunRunAway]

CASE has a similar problem, I've created another issue #15291