Fix repoUrl generating wrong path, following links when it shouldn't

[autumnfound]

Source: https://bugs.eclipse.org/bugs/show_bug.cgi?id=569607

When generating the repoUrl using the filepath, symlinks are being followed causing bad project URLs to be generated. New solution should be more robust when dealing with symlinks to retain functionality.

[mbarbero]

IMO, we should not look for the path of the repo on disk. We should just get the current repo name (e.g. ajdt/org.eclipse.ajdt or statet/org.eclipse.statet-commons) as this is what is eventually used to generate the urls of the repo:

• https://git.eclipse.org/r/ (<- anonymous url)
• https://git.eclipse.org/r/a/ (<- authenticated url)
• git://git.eclipse.org/gitroot/
• ssh://git.eclipse.org:29418/statet/

I guess you'll have to first update the logic in https://github.com/EclipseFdn/git-eca-rest-api

[autumnfound]

I've made 2 changes to affect this update, one here(#33) and one at EclipseFdn/git-eca-rest-api#42. These changes together will send statet/org.eclipse.statet-commons, and be able to read this change on the API side. To achieve this, projects will have .git removed from repo URLs for Gerrit repos (if present) to normalize the URLs. This depends on project repos being named in a way that matches the repo URL.

i.e, for ajdt/org.eclipse.ajdt, the expected repoURL is /gitroot/ajdt/org.eclipse.ajdt

[jonahgraham]

The recently updated change pushed out yesterday is causing problems - see my email on cross-project. The problem may be down to simrel's special status?

[jonahgraham]

PS. Not just simrel that is failing for me, the CDT website also is not calling me a committer.

$ git push origin HEAD:refs/for/master
Enumerating objects: 10, done.
Counting objects: 100% (10/10), done.
Delta compression using up to 8 threads
Compressing objects: 100% (7/7), done.
Writing objects: 100% (7/7), 898 bytes | 898.00 KiB/s, done.
Total 7 (delta 5), reused 0 (delta 0), pack-reused 0
remote: Resolving deltas: 100% (5/5)
remote: Processing changes: refs: 1, done    
remote: commit 1d252fc: ----------
remote: commit 1d252fc: Reviewing commit: 1d252fcc
remote: commit 1d252fc: Authored by: Jonah Graham <jonah@kichwacoders.com>
remote: commit 1d252fc: 
remote: commit 1d252fc: Reviewing commit: 1d252fcce2614b283d8b84e65c7063b5f189b1ce
remote: commit 1d252fc: Authored by: Jonah Graham <jonah@kichwacoders.com>
remote: commit 1d252fc: The author is not a committer on the project.
remote: commit 1d252fc: The author has a current Eclipse Contributor Agreement (ECA) on file.
remote: commit 1d252fc: The author has not signed-off on the contribution.
remote: If there are multiple commits, please ensure that each commit is signed-off.
remote: commit 1d252fc: 
remote: commit 1d252fc: Please see http://wiki.eclipse.org/ECA
To ssh://git.eclipse.org:29418/www.eclipse.org/cdt
 ! [remote rejected] HEAD -> refs/for/master (commit 1d252fc: The contributor must sign-off on the contribution.)
error: failed to push some refs to 'ssh://git.eclipse.org:29418/www.eclipse.org/cdt'

[autumnfound]

Committer status is reserved for projects that are tracked as a repo connected to a project. Typically the websites for projects are maintained as a part of the project, but are not directly tied. This is why there is no committer status for the CDT site at least. I talked with @chrisguindon earlier and he had mentioned that SimRel is in a similar boat, and would require sign offs.

[autumnfound]

This was fixed a couple of days ago, but wasn't closed, closing now.

[jonahgraham]

I don't know where I am supposed to make such comments - so as this is now closed I have continued in https://bugs.eclipse.org/bugs/show_bug.cgi?id=569703#c21

[chrisguindon]

+1 We are using https://bugs.eclipse.org/bugs/show_bug.cgi?id=569703#c21 to capture feedback.