[EVM] EVM Compliance Issues

[smblucker]

The following are a compiled list of issues related to the EVM; it includes all issues compiled in #588 by @silasdavis as well as subsequent issues:

Table of Contents:

• EIP Implementation Issues
• Bad/Invalid Jumpdest Issues
• Invalid Opcodes
• Chain Stopping Issues

EIP Implementation Issues
Byzantium:

• INVALID opcode - #601

• STATICCALL opcode - #594

• REVERT opcode - #600

• RETURNDATASIZE and RETURNDATACOPY opcodes- #705

Constantinople:

• SHL, SHR, and SAR opcodes - #705

Bad/Invalid Jumpdest Issues

• String concatenation issue - #300, see failing test: https://github.com/monax/bosmarmot/blob/develop/legacy-contracts.js/failing%20tests/%2327.js

• Subcurrency test causes panic in EVM tool but not burrow - #675

Invalid Opcodes

• Bad EVM interpretation on non-opcode - #526, see failing test: https://github.com/monax/bosmarmot/blob/develop/legacy-contracts.js/failing%20tests/WebSocket.js

Chain Stopping Issues

• Storage/Merkle related panic #626

• Storage/BlockCache related panic #627

[kyriediculous]

Is there an ETA on exceptions rather than using if-else-> Boolean ?

What version of solc is burrow using atm?

Thank you in advance!

[compleatang]

@kyriediculous we test against solc 0.4.22 at the moment.

[kyriediculous]

@compleatang That's super awesome !

That means the five-types model might actually work well now that we can return dynamic types from functions (despite from gas cost maybe)

[compleatang]

hahaha. well, considering we've been writing five-types models since before solc existed, I certainly hope so :)

[compleatang]

Also @kyriediculous we have a test for dynamic arrays at least here -> https://github.com/monax/bosmarmot/pull/39

We'd love some help writing a test for dynamic types if you had a moment....

[silasdavis]

Finally fixed by #940 and #953