Experimental Gump RSS feed
At the moment, I'm not spending much time on this, but I did spend enough to capture the module as the subject...
Posted by Sam Ruby at
A newsletter received from SourceForge.net Team last month:
"New RSS Feeds.
---------------------------------------------
A convenient way to keep up-to-date with the SourceForge.net site,
project news, activity information, and new projects is now available.
SourceForge.net provides a number of RSS feeds that may be accessed
using an RSS reader or aggregator (some content management systems also
allow you include data from an RSS feed on your website, much as
Slashdot and Freshmeat do in the right sidebar of their layout). Using
these RSS feeds, you can keep up-to-date with the latest project news
and file releases, Site Status updates and SourceForge.net statistics
(such as top projects).
Project-specific RSS feeds may now be accessed using the "View list of
RSS feeds available for this project" link on the Summary page for each
project. Information about the full set of the available
SourceForge.net RSS feeds (14 feeds, in all), including information on
the software needed to view RSS feed data and a complete list of the
options available for each RSS feed, may be found at:
https://sourceforge.net/docman/display_doc.php?docid=15483&group_id=1"
RAN: RSS for dependency management?
Sam Ruby: Experimental Gump RSS feed Here's some ant to GET the Ant task extensions I tend to rely on: <property name="gob.href" value="http://www.dehora.net/code/gob"/> <property name="gob.href.core"... [more]Trackback from Bill de hÓra at
Hi Sam,
fyi: I've been working on an RDF vocabulary/module/schema for project/issue management. The write-up's still sketchy but I think the terms are now pretty much in place:
Danny, as this matures, I certainly would be up for writing an XSLT transform which would produce such information from Gump definitions.
Two points to think about. First, what you describe is IMHO pretty much orthogonal to RSS. The hard part about producing Gump RSS is not the serialization but capturing the time dependent aspects of the information. I can expand on this if you like.
The main problem I see with an RDF definition of project is identity. I could certainly produce an RDF description for all the projects that Gump builds. It may be the only description you get for junit. But there undoubtably will be another description for Maven. Which one 'wins'?
Posted by Sam Ruby at"...pretty much orthogonal to RSS..." - that's certainly true when RSS is only considered as a syndication format, but I reckon there's a lot of potential when viewed from the RDF model point of view. e.g. the classic bugtracker web app is essentially the same system as a blogging app such as MT, and a common tools could be used across the language. There's also a lot of potential for reusing the RSS tools with other languages, to e.g. view bug reports in a newsreader.
re. time dependent aspects of Gump - please do expand (I'd better do some background reading too).
I don't really understand the difficulty re. identity, which probably means I'm missing something here. The way I'd approach this is to include Maven's project name as the prj:name (from my shiny new schema) and give it a prj:version etc. Depending on how it was organised, the project could either have a per-version URI, or be represented as a blank node identified by its characteristics - name, date, version etc.
Posted by Danny atEvery day a gump run is made. (Actually, there are multiple gump runs on multiple machines, but lets focus on just one for the moment).
Last night's run produced 345 pieces of information about projects and their ability to work with others. But is it all news?
Last night,
org.apache.maven.project.ProjectInheritanceTest failed. Is this news? I think not.
Had the maven build been held up due to a prereq on one day and then failed in the same way the next, it still wouldn't have been news.
In fact, had it been held up due to a prereq for a full week and then failed, it still wouldn't be news. But a week back is as far as I keep data online. So if it had been held up due to a prereq for a full week, I have no way of reliably determining if today's build status is news or not.
It is often the case that knowing which pieces of information is relevant is more important than having all data expressed in a canonical form.
Posted by Sam Ruby atHmm - thought a bit...why not just feed the changes?
i.e.
Monday ok
Tuesday fail
news
Monday fail
Tuesday ok
news
Monday fail
Tuesday fail
silence
Monday ok
Tuesday ok
silence
In essence that's what I do.
One added twist: if there is a prereq failure, there is no data for that date. So that date doesn't count.
So, for every project there is a sequence of up to seven data points. If they are all the same, then then there is no news. If there is a difference, the divide the stream into consecutive sequences where the status is the same. Select the last such sequence. Select the first item in that sequence. That's the newsworthy item for that project.
Source: gumprss.py
Posted by Sam Ruby atThat's a great little exercise for RDF modelling/manipulation!
I hope you don't mind, I'll pass this on to the rdf-calendar list, it might
well make a good test case.
RSS Action
There's something in the air over RSS. This weekend's frenzy of 3PDNA's has carried over today, with the author of Syndirella expressing that he may stop development, and a host of blog posts about SharpReader (which I like, but as long as there's...Excerpt from Ted Leung on the air at
RSS Action
There's something in the air over RSS. This weekend's frenzy of 3PDNA's has carried over today, with the author of Syndirella expressing that he may stop development, and a host of blog posts about SharpReader (which I like, but as long as there's...Excerpt from Ted Leung on the air at
Never satisfied... can we get one RSS feed per project?
Posted by James Snell at