@sig - thanks for finding the bugs - the more bugs found, the higher quality of ESME - and then everyone wins.
If you want to get a head start on testing the new RC - just take a look here: https://svn.apache.org/repos/asf/incubator/esme/tags/apache-esme-1.0-incubating/ I'll do some tests tomorrow during the day and then start a vote in the evening. D. On Sun, Feb 21, 2010 at 6:22 PM, Sig Rinde <[email protected]> wrote: > Great work folks, interesting process to follow I must say I also find it interesting - especially comparing the release process in a "community-based" / ASF project to releases in the corporate world. > > And sorry for messing up, keep in mind for next time I'm really good > at stumbling over bugs, i.e. mess up, so set me to test work early =) > > Sig > > On 21 February 2010 17:55, Ethan Jewett <[email protected]> wrote: >> Hi, >> >> Thanks for humoring my concerns, and again for all the work you're doing >> Dick. >> >> I've checked out the new tag and tested it. It passes all unit tests, >> search is working out of the box, and the security hole is closed. >> From my point of view, we're ready for a vote. >> >> Thanks, >> Ethan >> >> On Sun, Feb 21, 2010 at 10:08 AM, Richard Hirsch <[email protected]> >> wrote: >>> OK. >>> >>> I've tagged a new Release Candidate: >>> https://svn.apache.org/repos/asf/incubator/esme/tags/apache-esme-1.0-incubating/ >>> >>> This means that we have to vote again (sigh!) -. >>> >>> This time I suggest we test the tagged RC before we do a vote. >>> >>> D. >>> >>> On Sun, Feb 21, 2010 at 12:49 AM, Ethan Jewett <[email protected]> wrote: >>>> Hi all, >>>> >>>> For the original issue, where doing nothing simply results in a loss >>>> of functionality, I would agree. However, I think this is a major >>>> security hole that requires that the person deploying the software to >>>> take a specific action. If they don't take this action, then their >>>> deployment is vulnerable. I'm not comfortable putting the ESME stamp >>>> on a release that we know has this kind of issue. I think it's worth >>>> spending the extra time to address this issue and set the precedent >>>> that we don't release software with known security holes. >>>> >>>> I'm sticking with my -1 at this point. I'm not trying to veto (I don't >>>> even know if I can :-), so if a majority have voted for release after >>>> 72 hours (which I think is the case), then feel free to go ahead. >>>> >>>> Ethan >>>> >>>> On Sat, Feb 20, 2010 at 3:46 PM, Richard Hirsch <[email protected]> >>>> wrote: >>>>> I agree with Bertrand. >>>>> >>>>> I'd like to get this release out and then do a another release soon to >>>>> fix the errors. >>>>> >>>>> Right now, there are the two issues that have to be changed and Ethan >>>>> has already changed them in SVN. >>>>> >>>>>>I believe that this will happen on any system and I think the fact that >>>>>>search and the API2 doesn't work out of the box will really >>>>>>confuse people. >>>>> The fact that search doesn't work is IMHO the lesser of the two >>>>> errors. Does the API2 not work at all or is the problem more the >>>>> security one associated with the "role.api_test=integration-admin" >>>>> setting? >>>>> >>>>> I'm reluctant to cut a new release , because then we'd have to start >>>>> over again. Like I've said, I see this first release as a learning >>>>> experience. No release will be perfect and will always include a few >>>>> bugs. I'd rather get this release out and then do another release in 2 >>>>> weeks time with the bug fixes. Now that we know how to do create >>>>> releases it will be easier the next time. We should get used to >>>>> >>>>> I'd rather describe the two changes that have to made in the resource >>>>> files in a blog post or on the wiki and then push for a new release. >>>>> >>>>> Anyone else have thoughts on this >>>>> >>>>> D. >>>>> >>>>> >>>>> >>>>> On Sat, Feb 20, 2010 at 10:10 PM, Bertrand Delacretaz >>>>> <[email protected]> wrote: >>>>>> On Sat, Feb 20, 2010 at 9:03 PM, Ethan Jewett <[email protected]> wrote: >>>>>>> ...Maybe making this two-line change to one file is small enough that we >>>>>>> don't have to revote. I'm not sure. Maybe the mentors can weigh in.... >>>>>> >>>>>> Anything that changes the release artifacts needs a new vote. >>>>>> >>>>>> On the other hand, if there's a workaround (IIUC people can change >>>>>> something manually to get things to work?) I suggest releasing as is. >>>>>> >>>>>> Nothing prevents you from making another release soon, if needed. >>>>>> Getting used to releasing is good progress towards graduation ;-) >>>>>> >>>>>> -Bertrand >>>>>> >>>>> >>>> >>> >> >
