[JENKINS-51594] Test behavior of snapshots vs. RCs, JEP-229 #6
Conversation
| @@ -48,10 +48,6 @@ | |||
| * <p> | |||
| * 'SNAPSHOT' is also allowed as a component, and "N.SNAPSHOT" is interpreted as "N-1.*" | |||
| * | |||
| * <pre> | |||
| * 2.0.* < 2.0.1 < 2.0.1-SNAPSHOT < 2.0.0.99 < 2.0.0 < 2.0.ea < 2.0 | |||
There was a problem hiding this comment.
This was just so far off that I did not see how to rescue it. 2.0.1 < 2.0??
| @@ -180,7 +176,7 @@ public String toString() { | |||
| * Represents a string in the version item list, usually a qualifier. | |||
| */ | |||
| private static class StringItem implements Item { | |||
| private final static String[] QUALIFIERS = {"snapshot", "alpha", "beta", "milestone", "rc", "", "sp"}; | |||
| private final static String[] QUALIFIERS = {"alpha", "beta", "milestone", "rc", "snapshot", "", "sp"}; | |||
There was a problem hiding this comment.
Technically from Maven's PoV -SNAPSHOT happens before -alpha-SNAPSHOT which happens before the release.
IOW your version starts off as 1.1-SNAPSHOT, then goes to 1.1-alpha-1-SNAPSHOT, 1.1-alpha-1, 1.1-alpha-2-SNAPSHOT, 1.1-alpha-2, 1.1-beta-1-SNAPSHOT, 1.1-beta-1, 1.1-rc-1-SNAPSHOT, 1.1-rc-1, ..., 1.1, 1.1-sp-1-SNAPSHOT, 1.1-sp-1, 1.1-sp-2-SNAPSHOT
So if you want to stay consistent with Maven's version handling, this is not the change you want
There was a problem hiding this comment.
Hmm…I am wondering whether we should just be making this library a shim for ComparableVersion. It is not entirely clear to me why it differs at all.
There was a problem hiding this comment.
some "magic" that KK had for * as a version component
There was a problem hiding this comment.
I saw a comment to that effect…but is this “feature” actually being used anywhere that you know of?
There was a problem hiding this comment.
I know not of any usage of that...
There was a problem hiding this comment.
Hmm milestones are usually pre-alpha thus milestone -> alpha -> beta -> rc -> ga but yeah maven ¯_(ツ)_/¯
There was a problem hiding this comment.
|
@jglick FYI this is now conflicted. |
Fixed.
Well it was not “unfinished”, it was more that there was no clear consensus on what the behavior of this library should be or what the historical decisions were. |
There was a problem hiding this comment.
as per Stephenc.
changing the order of rc and snapshot whilst makes sense for our incrementals JEP (222?) it does not match mavens ordering and I am not clear on what if any issues that may cause.
As Snapshots are normally clearly identified and can be filtered this should not cause too many issues (famous last words), but I am not clear on where any checking would actually take place between a snapshot and a non snapshot version.
| @@ -180,7 +176,7 @@ public String toString() { | |||
| * Represents a string in the version item list, usually a qualifier. | |||
| */ | |||
| private static class StringItem implements Item { | |||
| private final static String[] QUALIFIERS = {"snapshot", "alpha", "beta", "milestone", "rc", "", "sp"}; | |||
| private final static String[] QUALIFIERS = {"alpha", "beta", "milestone", "rc", "snapshot", "", "sp"}; | |||
There was a problem hiding this comment.
Right: #6 (comment) jenkinsci/jenkins#3551 makes this less interesting. |
jglick
changed the title
jglick
changed the title
| // Maven considers 99.1.abcd1234abcd to sort before 99.1234deadbeef so we cannot simply use 99. as the branch prefix. | ||
| // Nor can we use 99.1234deadbeef. as the prefix because Maven would compare 5 and 10 lexicographically. | ||
| // 100._. seems to work but is not intuitive. | ||
| // Using changelist.format=%d.v%s behaves better, apparently because then the hash is never treated like a number. |
There was a problem hiding this comment.
JENKINS-51594
See jenkinsci/jenkins#153 where @stephenc was apparently trying to act like Maven by default.