Vyper: `AttributeError: 'Attribute' object has no attribute 'id'` on constants
Version Information
- vyper Version (output of
vyper --version): 0.1.0b14+commit.4ac5ef5
OS: osx - Python Version (output of
python --version): 3.7.5 - Environment (output of
pip freeze): same as previous TSTL issues
What's your issue about?
Compiling:
cv0: constant(address) = block.coinbase.balance
produces:
Error compiling: vfile.67808.vy
AttributeError: 'Attribute' object has no attribute 'id'
agroce
All 7 comments
(this is new on latest git pull; confirm some other issues fixed, my regressions mostly down to that CompilerPanic on ranges)
agroce
on 4 Jan 2020
I know this is just a bad error message, at heart, but I don't want to filter out AttributeErrors from the testing so a fix eventually would be good to aid fuzzing.
agroce
on 4 Jan 2020
And "you can't make a constant from a call" would obviously be more user-friendly.
agroce
on 4 Jan 2020
Great find! Hopefully, the fact that we haven't had an issue from you in a while means we're doing better 馃ぃ
fubuloubu
on 5 Jan 2020
You guys have definitely got less to find now.
However, I'm up to some new tricks maybe soon. :)
Have you used python-afl to fuzz vyper any, by the way?
agroce
on 6 Jan 2020
Well, half of us are American, we're mostly fuzzy, and I think at least one of us is a lop, so fairly familiar with it but haven't used it against the codebase yet.
We did start fuzzing the grammar using Hypothesis and Lark though
fubuloubu
on 7 Jan 2020
Ok, you're clean on current tests, and I have to hack AFL and a mutation testing tool before I start a more serious attack :)
Will keep you posted if anything new comes up.
agroce
on 7 Jan 2020
Related issues
lsaether
路
4Comments
jakerockland
路
4Comments
denis-bogdanas
路
3Comments
fubuloubu
路
3Comments
nrryuya
路
4Comments
Most helpful comment
Ok, you're clean on current tests, and I have to hack AFL and a mutation testing tool before I start a more serious attack :)
Will keep you posted if anything new comes up.