[
https://issues.apache.org/jira/browse/SCM-885?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16474197#comment-16474197
]
Matthieu Brouillard commented on SCM-885:
-----------------------------------------
Another finding I have just made is that jgit implementation behaves as the PR
does (for the exe implementation), ie without startVersion it walks to the root
commits.
So now where is the truth? Do we consider the exe implementation as the truth
and say that the jgit one is buggy and thus provide a fix that will make the
jgit impl to behave like the git one introducing by that a "regression" or do
we consider the opposite exactly as the PR provides?
If you look at the patch attached and run it against master, then you will see
that the jgit impl passes the tck test whereas the git one does not.
So the question is now where to put the regression, in git impl or in jgit impl?
> GitChangeLogCommand is wrong when only endVersion is set
> --------------------------------------------------------
>
> Key: SCM-885
> URL: https://issues.apache.org/jira/browse/SCM-885
> Project: Maven SCM
> Issue Type: Bug
> Components: maven-scm-provider-gitexe
> Affects Versions: 1.9.5
> Reporter: Matthieu Brouillard
> Priority: Major
> Fix For: waiting-for-feedback
>
> Attachments: scm-885-different-behavior-exe-jgit-patch.diff
>
>
> Invoking execution of a GitChangeLogCommand where only the end revision has
> been set produces a wrong out.
> +Actual result:+
> {{git whatchanged --date=iso ..END_REVISION_SHA1 -- PROJECT_PATH}}
> +Expected result:+ only the end revison SHA1 is used without the two dots
> {{git whatchanged --date=iso END_REVISION_SHA1 -- PROJECT_PATH}}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
