Would it be possible to do a full Ratis snap-shot release? So, we can just consume all the changes with a single update on ozone side. Sorry, I am just being a little selfish here, but it makes ozone’s life little easier. Otherwise we will have to follow up with another release.
--Anu On 10/11/18, 11:44 AM, "Josh Elser" <[email protected]> wrote: Hey Anu, No need for me to wait around. This is just for the new thirdparty repo -- not a release of ratis itself :) On 10/11/18 2:31 PM, Anu Engineer wrote: > Hi Josh, > > Can you please include Ratis-348?, it fixes a critical issue for Ozone. You might have to wait until the end of day to roll the build. > > Thanks > Anu > > > On 10/11/18, 11:26 AM, "Josh Elser" <[email protected]> wrote: > > The only thing that the ASF releases is source code, therefore the > policy you're quoting isn't relevant. There is no requirement to vote on > binary artifacts that are created from that source release. The > obligation on us is to verify that anything else the source release > creates follows ASF licensing like the source release does (e.g. our > JARs contain appropriate L&N files). > > You are right about incubating in the filename though -- totally forgot > about that requirement. Let me roll an rc1 with that. > > Thanks for catching that! > > On 10/11/18 11:23 AM, Tsz Wo Sze wrote: > >> I don't see any value for that, tbh. ... > > > > I agree that no one is going to download and use the binary. However, > > it is an artifact which we can vote for. It seems ASF requires us to > > put this artifact in the distribution directory, which is a > > subdirectory of www.apache.org/dist/ according to > > http://www.apache.org/legal/release-policy.html#where-do-releases-go > > > > BTW, just found the following from > > https://incubator.apache.org/policy/incubation.html#releases > > - the release archive MUST contain the word "incubating" in the filename; and > > - the release archive MUST contain an Incubation disclaimer (as > > described in the previous section), clearly visible in the main > > documentation or README file. > > > > We don't have "incubating" in the rc0 filename and DISCLAIMER seems > > missing in the binary jars in > > https://repository.apache.org/content/repositories/orgapacheratis-1007/ > > > > I guess we need a rc1? > > > > Tsz-Wo > > On Thu, Oct 11, 2018 at 10:43 PM Josh Elser <[email protected]> wrote: > >> > >> Thanks for the vote, Nicholas! > >> > >> On 10/10/18 10:43 PM, Tsz Wo Sze wrote: > >> > - untar and then "mvn install" does work for me. It won't work if we > >> > run a second "mvn install" without clean. "mvn install" works again > >> > after "mvn clean". It seems not a problem. > >> > >> Will have to investigate what's going on. > >> > >>> Questions: > >>> - Should we post a rc for the binary? > >> > >> I don't see any value for that, tbh. 99% of people are not going to know > >> this even exists and will get it via Maven. In another line of thinking, > >> the Maven repository I sent out "is" the binary release :) > >> > >>> - Now the project name becomes "Apache Ratis Thirdparty Parent" (and > >>> the gz file name) instead of "Apache Ratis Thirdparty". It is a > >>> little odd. How about using "Apache Ratis Thirdparty" for the root > >>> module and "Apache Ratis Thirdparty Shaded" for the sub-module? I am > >>> fine if we do the rename later. > >> > >> More than happy to revisit naming later on :). I wasn't able to come up > >> with a good name for our general Ratis dependencies module. "Apache > >> Ratis Thirdparty Shaded" is probably the forerunner, but I don't feel > >> like it's very descriptive. Need to think about that some more :) > > >
