From 22a2f5dd1d22fa7577278883cfd9c2f96c152c28 Mon Sep 17 00:00:00 2001 From: Chris Ward Date: Wed, 18 Jul 2018 17:03:30 +0200 Subject: Clarify truncation with addresses --- docs/types.rst | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/docs/types.rst b/docs/types.rst index 566c4c19..fbc83972 100644 --- a/docs/types.rst +++ b/docs/types.rst @@ -104,6 +104,14 @@ Operators: * ``<=``, ``<``, ``==``, ``!=``, ``>=`` and ``>`` +.. warning:: + If you convert a type that uses a larger byte size to an ``address``, for example ``bytes32``, then the ``address`` is truncated. + To reduce conversion ambiguity version 0.4.24 and higher of the compiler force you make the truncation explicit in the conversion. + Take for example the address ``0x111122223333444455556666777788889999AAAABBBBCCCCDDDDEEEEFFFFCCCC``. + + You can use ``address(uint160(bytes20(b)))``, which results in ``0x111122223333444455556666777788889999aAaa``, + or you can use ``address(uint160(uint256(b)))``, which results in ``0x777788889999AaAAbBbbCcccddDdeeeEfFFfCcCc``. + .. note:: Starting with version 0.5.0 contracts do not derive from the address type, but can still be explicitly converted to address. -- cgit