Russ did a good job explaining the trade-offs in his keynote at Gophercon Singapore. A bit more accessible than the blog posts. https://youtu.be/F8nrpe0XWRg
-
-
-
I don't find it compelling. Minimal version selection is solving a non-problem in a way that makes new problems.
- 2 more replies
New conversation -
-
-
serious question: is the SAT thing a problem because they need to re-run the dependency solver every build with constraints gathered from every source file in every package, or what?
-
This Tweet is unavailable.
- 1 more reply
New conversation -
-
-
But didn't you know that the versioning algorithm used by nearly every modern package manager is completely unworkable? Learning lessons from other language communities is so passe.
-
I'm guessing you didn't actually read Russ's blog posts.
- 4 more replies
New conversation -
-
-
I’m pretty enthused about it. It matches my strategy for updating vendored dependencies manually, which has been the only reliable dependency management approach I’ve experienced.
-
I’ve worked with several of the largest perforce trees in the world where all dependencies are checked in. When you upgrade dependencies, you want to advance the smallest possible increments of time and make the smallest possible number of changes.
- 3 more replies
New conversation -
-
-
I for one would very much prefer a minimum *number* of versions selection. As in, "if it fulfills dependencies of both crate A and B, pick version x rather than version y for crate A and version z for crate B".
-
It would be ideal if max versions would be selected, but packages provided information on what's the lowest version of their dependencies they can work with. SemVer doesn't really help with that, because you could depend on a fix in a minor version that didn't break compat.
End of conversation
New conversation -
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.