NewConstraint method faulty behavior

[benjaminbear]

Used version: 3.1.1
When I try to create a constraint with NewConstraint method I get following returns:

good Example
c, cErr := semver.NewConstraint("12.3.4.1234")
c.String() = nil
cErr = "improper constraint: 12.3.4.1234"

faulty behavior
c, cErr := semver.NewConstraint("12.23.4.1234")
c.String() = "12.23.4 1234"
cErr = nil

or

c, cErr := semver.NewConstraint("12.3.34.1234")
c.String() = "12.3.34 1234"
cErr = nil

---

I would expect the same behavior as with "12.3.4.1234", means telling me its not a contraint..

[benjaminbear]

FYI: the used regex to validate the contstraint fails: https://regex101.com/r/3fEGeK/1

[mattfarina]

I can explain what's going on here. 12.3.4.1234 is invalid because it has 4 parts and semver is only 3 parts.

12.3.34 1234 is read as 2 separate constraints. 12.3.34 and 1234. Those are added together. See https://github.com/Masterminds/semver#basic-comparisons

[benjaminbear]

Ah I'm sorry, had to correct the 2 faulty examples, there were copy paste typos.
Is this:
c, cErr := semver.NewConstraint("12.23.4.1234")
c.String() = "12.23.4 1234"
cErr = nil

expected behavior? Its a non semver string with 4 parts like in the first example, but handled differently. The third dot is removed and its a valid range?
I would expect an error here.

Wy is
12.3.4.1234 and
12.23.4.1234
handled differently? the only difference is that the second part is 2 digits

[mattfarina]

I have verified the issue. https://go.dev/play/p/wMEsKXb5KEx

[mattfarina]

@benjaminbear The fix can be found in #186 and the commit message explains what happened to cause this.

I am planning on getting a new release out by the end of this week with some fixes and new functionality.

Feel free to double check my fix and tell me if it didn't work.

[benjaminbear]

Great to hear! Thank you very much @mattfarina