Do not add members of address to contracts in experimental 0.5.0 #2981
Conversation
axic
changed the title
|
Needs tests, but this must be the simplest 0.5.0 change :) |
axic
force-pushed
the
no-address-overload
branch
from
5830277 to
34ba0d0
Compare
axic
changed the title
axic
force-pushed
the
no-address-overload
branch
from
34ba0d0 to
f734305
Compare
|
@chriseth this is ready |
| @@ -1660,7 +1660,9 @@ MemberList::MemberMap ContractType::nativeMembers(ContractDefinition const*) con | |||
| &it.second->declaration() | |||
| )); | |||
| } | |||
| addNonConflictingAddressMembers(members); | |||
| // In 0.5.0 address members are not populated into the contract. | |||
| if (!_contract->sourceUnit().annotation().experimentalFeatures.count(ExperimentalFeature::V050)) | |||
There was a problem hiding this comment.
Should we assert for _contract != nullptr or turn on 0.5.0 only if it was supplied?
There was a problem hiding this comment.
_contract is different from m_contract. The parameter only specifies the current scope, while m_contract is the actual x part in x.member.
There was a problem hiding this comment.
Cool, didn't notice the type has m_contract.
There was a problem hiding this comment.
Apparently m_contract doesn't behave correctly in this scenario for some reason.
There was a problem hiding this comment.
Oh I'm sorry. Of course it has to be the source unit where it is used in, not where it was defined.
There was a problem hiding this comment.
You initial comment was 100% correct.
docs/types.rst
Outdated
| @@ -107,6 +107,9 @@ Operators: | |||
|
|
|||
| * ``<=``, ``<``, ``==``, ``!=``, ``>=`` and ``>`` | |||
|
|
|||
| .. note:: | |||
| Starting with version 0.5.0 contracts do not derive from the address type. | |||
There was a problem hiding this comment.
... but are still convertible to address.
axic
force-pushed
the
no-address-overload
branch
2 times, most recently
from
12c6edb to
5257f5e
Compare
|
@chriseth updated, ready to merge |
axic
force-pushed
the
no-address-overload
branch
2 times, most recently
from
5f4a3f8 to
90b48ec
Compare
axic
force-pushed
the
no-address-overload
branch
from
90b48ec to
09276cb
Compare
|
@chriseth passes tests now |
|
I hope we don't hit the assert... |
Fixes #2683. Depends on #2936.