+1 I'd like to see us meet our deadline too. A week of testing on a
frozen code base would be good, especially if we post RC-1 to users@ to
get them to help. Maybe I was asleep at the wheel, but 0.3 went out a
bit too fast from my perspective and the synthetic control kmeans bug
has been biting me on the butt ever since <g>. It would also be good to
verify that all the wiki text is up to date during this period.
On 10/1/10 12:14 PM, Sean Owen wrote:
I figure we want to take a moment from Monday to test and polish, sure.
Let's start Monday if possible and release a week later if all is well.
While the board report sure isn't a deadline it seems entirely possible to
get this done by mid-month.
On Fri, Oct 1, 2010 at 5:11 PM, Jeff Eastman<[email protected]>wrote:
Are we going to go directly from code freeze to ship without a period of
testing and polishing? In addition to the unit tests which pass, we have a
number of examples that ought to be cluster-tested on a stable release
candidate before we burn bits. Or is your intent to vote on RC-1?
On 515, the problem still manifests. I've tried changing:
params.set("output", "output/frequentpatterns");
to:
params.set("output",
getTestTempDir("output/frequentpatterns").getAbsolutePath());
which causes 2 of 3 tests to fail so I'm concerned there is something else
hardwired inside of PFPGrowth. I'm still investigating...
On 10/1/10 11:33 AM, Sean Owen wrote:
Down to 3 issues left for 0.4. Note that we have to report board status in
2
weeks -- would be great to report 0.4 is released as planned.
MAHOUT-287: Robin should this just be punted? no action in some time
MAHOUT-515: Is this actually done? or minor? either way seems like nothing
crucial left for 0.4?
MAHOUT-271: There's a patch -- if Joe doesn't confirm it's ready by
tomorrow, just going to push it
I'll call for a vote quite shortly on 0.4, so do take a moment to do your
last checks and such that you're ready, now, if you can.
*T**Key**Summary**Pr* *Status**Updated**Created*[image:
Improvement]<https://issues.apache.org/jira/browse/MAHOUT-287>
MAHOUT-287<https://issues.apache.org/jira/browse/MAHOUT-287>Bayes
Classifier should use Vector as
input<https://issues.apache.org/jira/browse/MAHOUT-287>[image:
Major][image: Open] Open26/Sep/1010/Feb/10[image:
Bug]<https://issues.apache.org/jira/browse/MAHOUT-515>
MAHOUT-515<https://issues.apache.org/jira/browse/MAHOUT-515>PFPGrowthTest
has a hard-wired reference to an existing output
directory<https://issues.apache.org/jira/browse/MAHOUT-515>[image:
Major][image: Open] Open30/Sep/1030/Sep/10[image:
Improvement]<https://issues.apache.org/jira/browse/MAHOUT-271>
MAHOUT-271<https://issues.apache.org/jira/browse/MAHOUT-271>Make
WikipediaDatasetCreatorMapper fuzzy category match respect word
boundaries<https://issues.apache.org/jira/browse/MAHOUT-271>[image:
Trivial][image: Patch Available] Patch Available30/Sep/1028/Jan/10
