-
Notifications
You must be signed in to change notification settings - Fork 151
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Prerelease constraint handling #36
Comments
As a test case: package main
import (
"log"
"github.com/hashicorp/go-version"
)
func main() {
constraint, _ := version.NewConstraint(">= 2.0")
version, _ := version.NewVersion("2.2.0-302bd8b")
ok := constraint.Check(version)
log.Printf("version check: %#v", ok)
}
|
Same problem. |
@jbardin You're probably more intimately familiar with this, do you mind taking a look? |
Hi @nevans, I think the comparison is correctly happening according to the specification. Semantic versioning alone doesn't define the constraint operations, only the comparisons between versions. If you take the examples versions you provided and compare them, you get
Which follows Item 11 in the SemVer spec. The difference from your first example is that when a user adds
The proper constraint to include There is however still a point of contention here. As shown above, NPM would require (Major, Minor, Patch) to match a pre-release, which is the behavior we followed in this package. I did mention Ruby Gems though, and we differ there as they chose to match any (Major Minor, Patch) as long as a prerelease was specified. For example
So Ruby would allow matching between the versions you specified, provided you start with a more precise constraint string. Now we did choose to follow the former behavior rather than the latter, since it provided a safer constraint set without compound statements being provided. Hopefully this clears up the situation, and we can better document this behavior for users. |
Thanks, @jbardin for your explanation. I can see why versioning constraints for selecting compatible libraries would want these constraints for prereleases, even though both Ruby gems and NPM fail the principle of least surprise for me. 😉 I'd expected the For our use case, we needed feature detection on a running server, which means that pre-releases shouldn't be treated as special. At any rate, as per your example, we could use a simple Thanks for taking the time to explain it. |
The logic in fbe76cf doesn't seem quite right. In particular, consider the use-case here: https://github.com/gesellix/couchdb-prometheus-exporter/blob/master/lib/couchdb-client.go#L52-L69 where we are checking if the server version is
>= 2.0
and the reported version is e.g.2.2.0-302bd8b
. Clearly, in this scenario, the constraint check should return true (but it returns false). At the very least, the constraint needs to be able to opt-out of this mode.https://semver.org/#spec-item-11 implies that
2.2.0-pre >= 2.0
should be true:"1.0.0-alpha < 1.0.0-alpha.1 < 1.0.0-alpha.beta < 1.0.0-beta < 1.0.0-beta.2 < 1.0.0-beta.11 < 1.0.0-rc.1 < 1.0.0." I can certainly see the case for not wanting to match pre-releases at all in certain circumstances.
The text was updated successfully, but these errors were encountered: