The bug has already been reported a year+ ago at
https://issues.apache.org/jira/browse/NPANDAY-322
It's resolved with "fixed at 1.2.1", so I added part of my previous e-mail
as a comment (I can't re-open the issue in Jira, why?).


Cheers,
Stoyan

On 11/1/11 10:44 PM, "Stoyan Damov" <[email protected]> wrote:

>Hi guys,
>
>First, sorry if the I'm not following the "protocol" for sending a single
>thing in one e-mail (e.g. this is a question, a bug report, and a feature
>request) but I couldn't find a "How to send e-mails to this list" link.
>
>Our company is evaluating NPanDay for use by our the .NET team (the Java
>team is already using maven), mainly to get rid of NuGet as a "dependency
>management" system with which we're not at all happy, but also to have a
>common toolset (.NET tools are years behind the ones for Java).
>From my little experience with NPanDay I believe it's the cure we were
>looking for (we were interested in NPanDay for quite some time but because
>we're using .NET framework 4 & VS 2010 previous versions were not worth
>evaluating).
>
>I found some problems (described below) in NPanDay.VisualStudio.Plugin
>(mainly in ReferenceManager) which I Q&D fixed. However, I'd like to
>actually do a more thorough code review & make the best possible fix, and
>then propose it.
>I couldn't find what's the "ceremony" of submitting a patch, hence I'm
>asking on this list. Sorry if that's not the right place for that.
>
>The first bug is in "Resync references" which calls
>ReferenceManager.CopyArtifact:
>
>if (!artifact.FileInfo.Exists || artifact.Version.EndsWith("SNAPSHOT"))
>{
>    if 
>(!NPanday.ProjectImporter.Digest.Model.Reference.DownloadArtifact(artifact
>,l
>ogger))
>    {
>        ReferenceErrorEventArgs e = new ReferenceErrorEventArgs();
>        e.Message = string.Format("Unable to get the artifact {0} from any
>of your repositories.", artifact.ArtifactId);
>        onError(e);
>        return;
>    }
>}
>
>The check for SNAPSHOT above forces "Resync references" to try and
>download
>the artifact off the remote repository.
>If the artifact is not yet installed (which is common if I'm developing a
>new module but don't want to commit/push it yet), what happens is that:
>1. The resync fails (which is bad) and
>2. If the artifact is installed in the local repo, it gets deleted (which
>is
>worse)
>
>What should happen (for SNAPSHOTs) instead is this:
>
>1. Try to get timestamp of artifact in remote repo. If that fails, it
>shouldn't be a problem, the artifact might not be deployed (yet).
>2. Try to get timestamp of artifact in local repo. If that fails, then
>this
>is an error and should be reported.
>3. Get the timestamp of the local file (in .references/...)
>4. If there's a remote timestamp and it's greater than the local one,
>update
>the local repo's artifact.
>5. If the local repo's timestamp (possibly updated at point 4) is greater
>than the timestamp of the local file, update the file in .references/...
>
>For steps 1 and 2 use the timestamp in maven-metadata.xml (in
>ReferenceManager.copyToReferenceFolder there's already a TODO which
>suggests
>to use "<metadata>/<versioning>/<lastUpdated>" for the timestamp).
>Currently, the local repo's timestamp is considered to be the file's last
>write time (not even .LastWriteTimeUtc, which is another bug).
>
>To make it clear, since we're all developers, here's the algo in pseudo
>code
>(this might not compile, I'm writing it in Outlook):
>
>DateTime remoteRepoTimestamp, localRepoTimestamp, localFileTimestamp;
>bool hasRemoteTimestamp =TryGetArtifactRemoteRepositoryTimestamp(artifact,
>logger, out remoteRepoTimestamp);
>bool hasLocalRepoTimestamp =
>TryGetArtifactLocalRepositoryTimestamp(artifact, logger, out
>localRepoTimestamp);
>localFileTimestamp = GetLocalFileTimestamp(artifact.FileInfo, logger);
>if (hasRemoteTimestamp && hasLocalRepoTimestamp && remoteRepoTimestamp >
>localRepoTimestamp)
>{
>UpdateLocalRepoArtifact(); // ==
>NPanday.ProjectImporter.Digest.Model.Reference.DownloadArtifact(artifact,
>logger)
>// ^^ error handling above omitted for brevity
>localRepoTimestamp = remoteRepoTimestamp;
>}
>if (hasLocalRepoTimestamp && localRepoTimestamp > localFileTimestamp)
>{
>copyToReferenceFolder(artifact, referenceFolder);
>} 
>
>Something like this.
>
>
>The 2nd thing is that there's a pretty good case to have "Resync
>references
>from *local* repo" and "Resync references" (from remote, which is
>described
>in the bug above).
>"Resync references from local repo" should skip the remote repo. Here's
>the
>valid case:
>
>I'm working on X and Y. I make a change in X and "maven install" it (in
>the
>local repo). I don't push the changes to the SCM.
>Y has a dependency on X and I want to resync the references so that I get
>the updated X dependency. Currently I can't do that.
>
>I can't (don't want to) push X's changes and wait on the build server to
>deploy the new snapshot because I don't want to "hurt" the rest of the
>developers with something which might still have bugs in it.
>
>Hope both the 1st bug & the feature request (it's more a feature request)
>are clear. Let me know if I should elaborate.
>
>One more thing. No offense but the current code base (I have only looked
>at
>the .NET code so far) needs some cleaning up (TODOs, XML "parsing", etc.).
>Let me know if you're interested in a code review and to who I should
>e-mail
>it, also I'll be more than glad to do most of the clean up myself if I'd
>be
>allowed to push/propose changes.
>
>Cheers,
>
>
>Stoyan Damov
>
>Head of Development
>Pirinsoft Bulgaria Ltd.
>
>40 Atanas Moskov str., fl.2
>Business Center Vitosha
>1715 Sofia, Bulgaria
>
>Tel: +359 2 421 85 10
>Fax: +359 2 421 85 15
>Mobile: +359 892 493 336
>E-mail: [email protected]
>
>This communication (including any attachments) contains proprietary
>information some or all of which may be legally privileged. Unless you are
>the intended recipient (or authorized to receive for the intended
>recipient), you must not print, retain, use, copy, distribute or disclose
>to
>anyone this message or any information contained in it. If you receive
>this
>e-mail in error, please notify the author immediately by replying to this
>e-mail and then delete it from your system. E-mail transmission cannot be
>guaranteed to be secure or error-free as information could be intercepted,
>corrupted, lost, destroyed, arrive late or incomplete, or contain viruses.
>The sender therefore does not accept liability for any errors or omissions
>in the contents of this message, which arise as a result of e-mail
>transmission.
>
>


Reply via email to