I wish that people would spend a tenth of the time actually profiling Cargo as they do talking about theoretical performance of dependency resolution.
-
-
Replying to @pcwalton
With a potentially exponential-time algorithm, benchmarks are not that enlightening: they tell you that simple cases are fast, but don't give much hint as to what sorts of inputs might cause execution times to explode and whether those inputs are likely.
2 replies 0 retweets 0 likes -
Replying to @NYarvin
Fortunately we have lots of experience to demonstrate that version resolution is virtually never a problem in practice.
2 replies 0 retweets 2 likes -
If only there was some sort of real world project with a ton of dependencies to check if performance of dependency resolution is actually a problem or not.
1 reply 0 retweets 2 likes -
Is exponential behavior not actually a risk here? Because if it is a risk, experience only tells you that you've not yet fallen over the exponential cliff, not that the cliff isn't out there waiting for you.
2 replies 0 retweets 0 likes -
The idea that maybe, someday, an algorithm for solving a problem might theoretically cause a performance problem is a terrible reason to not even try to solve that problem in the first place.
1 reply 0 retweets 1 like
Wow, Twitter really degrades this sort of conversation. I mean, I already said in other tweets that I thought that waiting until a problem arises seems perfectly reasonable here. This is just a point of curiosity as to what the future might hold.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.