* [*] parity-codec -> parity-scale-codec

* [core, model] Decode impls Option -> Result

* [core] typed_cell load expect successful Decode

* [core] Option to Result, remove redundant combinator

* [model] Option to Result

* [*] abbreviate parity-scale-codec

* [core] fix dummy Call Decode impl

* make std crate feature default for all ink crates

* adjust travis CI for builds without std feature enabled

* remove incorrect usages of ink_abi

* apply rustfmt

* remove invalid type-metadata dependencies for ink_core

* [lang] fix Cargo.toml indentations

* adjust Erc20 token examples for new std crate feature

* adjust template and lang example std feature and propagation

* [core] Factor nightly alloc parts out of core into new alloc ink crate

* Further cleanup after factoring out ink_alloc crate

* [alloc] Use symlinks instead of copies

* [ink] Update rust-toolchain from 2019-04-20 to 2019-05-21

* [core,lang] Update & fix Cargo information

* [chores] Update rustfmt version to the current toolchain

* [core] Clean up, fix some warnings

* [core] Remove unneeded dependency

* [core] Ignore return value of Key ops

Fixes a warning

* [core,model] Adjust warning level

* [lang] Minor rustfmt run

* [ink] Add topics to events

For now, ink! will simply leave topics slice empty.

* [core] Add ext_gas_left, ext_gas_price and ext_value_transferred bindings

* [core] Clean up some code by dedup

* [core] Further code clean up

* [core] Simplify SrmlEnv::load impl

Note: No other changes and renamings have been done yet so the crate is currently not in a compilable state.

This can be reverted back to the official parity-codec project once the new version with the changes has been released on crates.io.

Now it has the required features in a branch.

This is required because of some special impls for pDSL that while not being mandatory, they are still important enough to serve usability.

The problem is that env_logger is obviously not no_std compatible.
At first you would think that this is not a problem, since it is only a dev-dependency that is allowed to pull the std lib.
However, due to a Cargo "bug" all dev-dependencies and normal dependencies need to have shared feature flags.
So since env_logger depends on log and pdsl_core depends on log both need to share their feature set which is chosen to be just the one that pulls std - what a mess ...

We now need our own logging facilities that are no_std compatible.

The problem with std::collections::HashMap is that there is no core::collections::HashMap or similar replacement for our no_std environment.
Also hashbrown seems to have superior performance.

This will help writing tests with general setup and tear-down mechanics.
Used to initialize env_logger for every unit test.

Fixed now!

The new system uses similar patterns as .gitignore does.

This should make it work in a default test environment.

This enabled TestEnv and disabled DefaultEnv.
Use this for testing.
Before this commit the standard "test" feature was used, however, its problem is that the test feature is crate-local and cannot be enabled or disabled by dependencies.