Hi, [cc infra-dev@, FYI]
There's an experimental Gerrit installation at git.apache.org, and I've been thinking of ways for us to leverage it as a possible way of hosting native Git repositories at Apache. To do that, we'd need some candidate projects to iron out all the open issues and potential problems. Doing that within a "normal" Apache project is a bit troublesome as the code that's eventually targeted for a release wouldn't have an svn history behind it. At Labs however we have codebases that are not targeted to be released (except after migrating from away from the Labs), and thus don't have as strict operational requirements. So I was thinking that it would make sense to use a Labs codebase as a guinea pig for native Git support. What I'd like to do is to start up a new Lab for an idea I recently had (see below), but use Git instead of Subversion for version control. The obvious alternative is to simply do this on Github, but I'd like to see how this could work in Apache. The canonical repository of this codebase would live on git.apache.org and all commits would go through Gerrit to better keep track of the origin and oversight of the changes. I'd set up commit notifications etc. to go to Labs just like with svn. If the lab ended up attracting more interest, we'd take it to the Incubator as an externally developed codebase and do a normal IP clearance on it before importing it to Apache svn. WDYT? BR, Jukka Zitting <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"; xmlns="http://usefulinc.com/ns/doap#"; xmlns:foaf="http://xmlns.com/foaf/0.1/"; xmlns:labs="http://labs.apache.org/doap-ext/1.0#"; xmlns:projects="http://projects.apache.org/ns/asfext#";> <Project rdf:about="http://labs.apache.org/labs#oak";> <name>Oak</name> <shortname>oak</shortname> <shortdesc xml:lang="en">HTTP-based hierarchical storage system</shortdesc> <description xml:lang="en">Oak is a system for storing and managing a hierarchy of web resources. An Oak system is accessed using HTTP, possibly with extensions like WebDAV or AtomPub. Each resource stored in Oak is associated with a media type, and pluggable converters are used to generate different representations of the resources. For example, a PHP resource would be executed on the server side when requested as text/html. Pluggable map-reduce tasks are used for things like search indexes and resource summaries. The managed resource hierarchy is stored in a distributed hash table or an equivalent cloud storage backend.</description> <homepage rdf:resource="http://labs.apache.org/oak/"/> <license rdf:resource="http://usefulinc.com/doap/licenses/asl20"/> <created>2010-6-14</created> <labs:status>active</labs:status> <maintainer> <foaf:Person rdf:about="http://people.apache.org/~jukka/#me";> <foaf:name>Jukka Zitting</foaf:name> <foaf:homepage rdf:resource="http://people.apache.org/~jukka/"/> <foaf:mbox_sha1sum>79c84b7b6c8440f312aac4557938d9c3fbcb1384</foaf:mbox_sha1sum> </foaf:Person> </maintainer> <repository> <!-- TODO: How does a GITRepository look like? --> <SVNRepository> <location rdf:resource="http://svn.apache.org/repos/asf/labs/oak/"/> <browse rdf:resource="http://svn.apache.org/viewvc/labs/oak/"/> </SVNRepository> </repository> <programming-language>clojure</programming-language> <programming-language>javascript</programming-language> </Project> </rdf:RDF> --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
