This report has been sent and remains here for reference only.

Gump code was migrated to SVN[1], with Gump metadata continuing to remain in CVS[2] (to be less of a community barrier, until SVN usage has matured as the norm with ASF-wide). At this same time 'traditional' Gump (the Java/Shell/XSLT/Perl combination) was finally archived. Python Gump is re-structuring itself (within SVN) to attempt to become more intuitive to new users.

The Gump PMC voted to allow 'Gumped artifacts' to be publically available in a repository. These artifacts are to be specifically marked as '-gump' so as to attempt remove/reduce any confusion with releases. Also, each directory listing in the repository is to be marked with a disclaimer explaining that the contents can not be trusted, and are not releases. [3]

The Gump PMC voted to grant 'root access' to ajack@apache.org (a non-member) to Brutus. Since Brutus is managed by the Gump team (and is standalone/untrusted) infrastructure felt this did not constitute a security concern.

After some discussion on community@apache.org regarding the need for PMC oversight of mailing lists, the Gump PMC voted to sponsor python@apache.org. The intent is to discuss "Python w.r.t ASF", and allow the various Python projects to collaborate. Unfortuantely, the request has sat (seemingly untouched) in JIRA since August 4th 2004. [5]

Nicola Ken Barozzi has resigned from the Gump PMC. We want to thank him for his past contributions and are happy that he still wants to keep in touch by retaining his committer status.

Gump now can generate XHTML output directly without the need of a Forrest webapp (although still a supported option). This reduction in dependency is intended to ease Gump installation (and hence adoption) by new users.

Gump has received some technical improvements recently. Gump was migrated from Python 2.2 to Python 2.3 (and benefits from many new language features). Multithreading support for CVS|SVN downloads removed latency delays from Gump's total run time. MySQL integration (for historical results data) is currently live. RDF is generated describing projects and their interactions (at dependency level) using a work-in-progress Gump Ontology.

Gump's current focus is on opening itself up to more developers (via documentation) and new users. As part of reaching beyond Java Mono 1.0 has been installed on Brutus and we are collaborating with the NAnt[4] community to take on .NET projects.

Along these lines, and with little to no help from the Gump community, Peter Janes took Python Gump and integrated it with Perforce [6]. He supplied patches that added Perforce repository capabilities to complement CVS and SVN. This patch included documentation [7] [8] and was very high quality work.

Gump now build artifacts and stores them in an online repository, and (as such) can build projects using historical artifacts in it's repository. This will allow Gump to smooth out the problems of 'prerequisite failure' Further, Gump also works with Depot to download artifacts from artifact repositories to automate package processing, and/or allow cascaded Gumps. None of this code is well tested/mature at this point, and some problem reports exist.

  • No labels