Github user neykov commented on the pull request:
https://github.com/apache/incubator-brooklyn/pull/331#issuecomment-63179436
> if i understand correctly, splits on and otherwise ignores non-digits
It doesn't ignore them. If both parts are non-digits then compares plain
strings (L80).
>(and use the natural order comparator so rc10 > rc1 !)
Yes, definitely should use the natural order comparator for strings instead.
> it would also be sensible to have special logic that, given two versions
identical up to a qualifier we prefer the non-qualified item
I though the logic should be the reverse (as it is now) - that's what I
remember users expect? For example 3.3.0-200141115 beats 3.3.0.
> NaturalOrderComparator sorts 0.0.0_SNAPSHOT higher than
0.0.1-SNAPSHOT-20141111114709760
Yes, didn't expect it but turns out that's the case. That's actually the
reason why I decided to go with a separate comparator. First tried existing
version parsers - OSGi's provided Version class and also an external dependency
- semver which provides semantic Version class. Both failed to parse some of
the version strings we have (underscore is not supported as a delimiter in
semver, osgi requires a dot delimiter). They are just too strict for our use.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
