aboutsummaryrefslogtreecommitdiffstats
path: root/docs/abi-spec.rst
diff options
context:
space:
mode:
authorJason Cobb <jason.e.cobb@gmail.com>2018-02-16 23:11:07 +0800
committerAlex Beregszaszi <alex@rtfs.hu>2018-03-31 07:56:36 +0800
commite37b619593cc926873279dc1d5f532a7a13b6fa4 (patch)
treedddcf809cbeb342003891b20690e597d8b1c3f7a /docs/abi-spec.rst
parentbe261ed519e472af071f0b70acfbc277a5861a20 (diff)
downloaddexon-solidity-e37b619593cc926873279dc1d5f532a7a13b6fa4.tar.gz
dexon-solidity-e37b619593cc926873279dc1d5f532a7a13b6fa4.tar.zst
dexon-solidity-e37b619593cc926873279dc1d5f532a7a13b6fa4.zip
Set default fixed point decimal places to 18
Diffstat (limited to 'docs/abi-spec.rst')
-rw-r--r--docs/abi-spec.rst6
1 files changed, 3 insertions, 3 deletions
diff --git a/docs/abi-spec.rst b/docs/abi-spec.rst
index 4d84a7da..98301fdc 100644
--- a/docs/abi-spec.rst
+++ b/docs/abi-spec.rst
@@ -54,7 +54,7 @@ The following elementary types exist:
- ``ufixed<M>x<N>``: unsigned variant of ``fixed<M>x<N>``.
-- ``fixed``, ``ufixed``: synonyms for ``fixed128x19``, ``ufixed128x19`` respectively. For computing the function selector, ``fixed128x19`` and ``ufixed128x19`` have to be used.
+- ``fixed``, ``ufixed``: synonyms for ``fixed128x18``, ``ufixed128x18`` respectively. For computing the function selector, ``fixed128x18`` and ``ufixed128x18`` have to be used.
- ``bytes<M>``: binary type of ``M`` bytes, ``0 < M <= 32``.
@@ -164,9 +164,9 @@ on the type of ``X`` being
- ``int<M>``: ``enc(X)`` is the big-endian two's complement encoding of ``X``, padded on the higher-order (left) side with ``0xff`` for negative ``X`` and with zero bytes for positive ``X`` such that the length is 32 bytes.
- ``bool``: as in the ``uint8`` case, where ``1`` is used for ``true`` and ``0`` for ``false``
- ``fixed<M>x<N>``: ``enc(X)`` is ``enc(X * 10**N)`` where ``X * 10**N`` is interpreted as a ``int256``.
-- ``fixed``: as in the ``fixed128x19`` case
+- ``fixed``: as in the ``fixed128x18`` case
- ``ufixed<M>x<N>``: ``enc(X)`` is ``enc(X * 10**N)`` where ``X * 10**N`` is interpreted as a ``uint256``.
-- ``ufixed``: as in the ``ufixed128x19`` case
+- ``ufixed``: as in the ``ufixed128x18`` case
- ``bytes<M>``: ``enc(X)`` is the sequence of bytes in ``X`` padded with trailing zero-bytes to a length of 32 bytes.
Note that for any ``X``, ``len(enc(X))`` is a multiple of 32.