Vyper: Meeting - 9th April 2018

Created on 3 Apr 2018 · 5Comments · Source: vyperlang/vyper

Meeting Details

General Bi-Weekly meeting.
Start: Monday 9th April 2018, 14:00 UTC (https://time.is/UTC)
Hangouts: https://meet.google.com/czc-kuep-jda

Agenda

https://github.com/ethereum/vyper/issues/727 - Tuple assign, post or pre beta?
- https://github.com/ethereum/vyper/issues/665
- https://github.com/ethereum/vyper/issues/716 - Integer Division
- https://github.com/ethereum/vyper/issues/737 - Disallow implicit conversion
VIP Discussions:
- https://github.com/ethereum/vyper/issues/711 - Assertions that should never-ever fail
- https://github.com/ethereum/vyper/issues/728 - Allow list of byte arrays
- https://github.com/ethereum/vyper/issues/748 - Remove reduntand syntax for address to contract
- https://github.com/ethereum/vyper/issues/749 - Remove casting in signature
- https://github.com/ethereum/vyper/issues/750 - Require compiler flag for nonstandard primitive types
- https://github.com/ethereum/vyper/issues/539 - None in viper is misleading
- https://github.com/ethereum/vyper/issues/752 - Allow declaring custom units
- https://github.com/ethereum/vyper/issues/732 - Change binary literal format
- https://github.com/ethereum/vyper/issues/764 - uin256 as first class citizen
- shift discussion

meeting

Source

jacqueswww

Most helpful comment

Not sure I will be able to join the call.
Anyway I think we should add #752 to the VIP Discussions.

ben-kaufman on 3 Apr 2018

👍2

All 5 comments

Feel free to comment any additional agenda items here.

jacqueswww on 3 Apr 2018

Not sure I will be able to join the call.
Anyway I think we should add #752 to the VIP Discussions.

ben-kaufman on 3 Apr 2018

👍2

Improving shift function.

shift is compiled to the following code:

   ['if', ['sle', '_s', 0], 
          ['div', '_v', ['exp', 2, ['sub', 0, '_s']]], 
          ['mul', '_v', ['exp', 2, '_s']]]]],

When _s == 0, ['div', '_v', ['exp', 2, ['sub', 0, '_s']]] and ['mul', '_v', ['exp', 2, '_s']]]] are the same. But in the div branch, it need to execute sub which potentially costs more gas. Suggest using slt instead of sle. This will also ease the verification. On the other hand, EIP145 proposes supporting shift opcode directly in EVM. Not sure when it will happen.