[move] Use u256 instead of bigint for parsing #19998
Conversation
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
3 Skipped Deployments
|
| txt.as_bytes(), | ||
| let max_len = match base { | ||
| NumberFormat::Hex => AccountAddress::LENGTH * 2, | ||
| NumberFormat::Decimal => 78, |
There was a problem hiding this comment.
Where is the 78 coming from? Is it possible to express that in terms of AccountAddress::LENGTH?
There was a problem hiding this comment.
Yea! I can do that. It's based off of this: https://stackoverflow.com/questions/43787672/the-max-number-of-digits-in-an-int-based-on-number-of-bits
So would be 241 * AccountAddress::LENGTH / 100 + 1
There was a problem hiding this comment.
Wowza. Yeah I would definitely comment the shit out of that number lol
There was a problem hiding this comment.
Updated it to be a const + I put a link to the stack overflow article on it as that did a much better job than any explanation I can give.
There was a problem hiding this comment.
I'm reconsidering this move to u256 here altogether -- in particular _s in addresses are going to mess this change all up. So will need to look a bit more at it and will land outside of this stack.
tzakian
force-pushed
the
tzakian/new-tt-parser5
branch
from
d2e321f to
7691959
Compare
tzakian
force-pushed
the
tzakian/new-tt-parser6
branch
from
7ed24aa to
04adc54
Compare
tzakian
requested review from
amnn,
emmazzz,
stefan-mysten,
suiwombat and
wlmyng
as code owners
tzakian
force-pushed
the
tzakian/new-tt-parser6
branch
from
04adc54 to
6b941b1
Compare
|
Updated so we handle underscores as well now (in integers) + some minor updates to promote parity between the value tokenizer and type tokenizer. |
2 participants
Description
Switches to using u256 instead of bigint for parsing.
Test plan
CI + additional tests to make sure parity was preserved.