cypix314123 That’s interesting and puzzling. How are they going to implement it? Seems to me there will be lots of extra new syntax, which would make contracts much, much less readable than just regular comments.
@@magno5157 I don't really know. I think they will use the C++ attribute syntax. I stop working with C++ at C++14 and I heard that about contracts on C++ podcast. I found this www.modernescpp.com/index.php/c-core-guidelines-a-detour-to-contracts int push(queue& q, int val) [[ expects: !q.full() ]] [[ ensures !q.empty() ]]{ ... [[assert: q.is_ok() ]] ... }
it's because how easily hackable and dynamic python is, you can even do things like modify the AST of a function at runtime, or pull the stack frame out and inspect it at runtime, which makes it possible to write some wacky contracts that's basically impossible in many other languages
Yawn. Yet another talk with contrived examples designed to justify the existence of non-necessary packages that are not worth the tiny benefit they provide.
There were a few gems here that I want to commit to memory. Will be rewatching this one.
This is brutally impressive, thank you so much!
6:42 lol ..’huge negative numbers’ ...g1. made me chuckle a lil’.
This is cool!
I got to the part where he brought granola bars and just had to like the video on principle.
I've benn interested by contract programming paradim a year ago. Now i'm interested again :)
Didn't realize programming by contracts was still unknown to a large number of programmers. I thought they were already pretty standard.
Well now that C++20 will get them as language feature, few more will now.
cypix314123 That’s interesting and puzzling. How are they going to implement it? Seems to me there will be lots of extra new syntax, which would make contracts much, much less readable than just regular comments.
@@magno5157 I don't really know. I think they will use the C++ attribute syntax. I stop working with C++ at C++14 and I heard that about contracts on C++ podcast.
I found this www.modernescpp.com/index.php/c-core-guidelines-a-detour-to-contracts
int push(queue& q, int val)
[[ expects: !q.full() ]]
[[ ensures !q.empty() ]]{
...
[[assert: q.is_ok() ]]
...
}
@@cypix314123 wow C++ is becoming even more pathetic soup of punctuation marks
Great talk.
Why "especially" Python? I'd be interested to hear more about that.
The talk was at PyCon (Python Conference), there isn't anything specific that makes this constrained to Python.
it's because how easily hackable and dynamic python is, you can even do things like modify the AST of a function at runtime, or pull the stack frame out and inspect it at runtime, which makes it possible to write some wacky contracts that's basically impossible in many other languages
where did he import @ensure from?
from contracts
The module is: pypi.org/project/dpcontracts/
Git Repo: github.com/deadpixi/contracts
2:57 beyond unit tests
"...better than any testing framework out there...any language..." Ok. Hypothesis is almost an exact clone of quickcheck for haskell. So, dude, nah.
He did say 'one of the best', to be fair, and considering the limitations of Python when compared to Haskell, that's a significant achievement.
Yawn. Yet another talk with contrived examples designed to justify the existence of non-necessary packages that are not worth the tiny benefit they provide.