Thanks Julian, after this refactor, the source code release package reduce to 32MB:) I think it will be a good candidate to release.
Will refer to calcite one to setup Kylin part, your experience is really great reference for us. I'm sending another vote thread and please help to verify again for rc2. Thank you very much. Best Regards! --------------------- Luke Han 2015-05-27 4:10 GMT+08:00 Julian Hyde <[email protected]>: > I think that is wise. > > Also consider taking binary test data out of the release. You could put it > in a different tar file, and ask people to unzip that tar file at a > particular place in your source tree. > > The goal, I think, is that a source release should consist only of source > code. Easier said than done, I know. > > In Calcite, we put some (smallish) test data into jar files managed by > maven (e.g. > http://search.maven.org/#search%7Cga%7C1%7Ca%3A%22foodmart-data-json%22 ). > And we have a separate project to build a VM populated with test data (see > https://github.com/vlsi/calcite-test-dataset). > > Also consider removing all generated code from the git repo. (This includes > test data and the web site.) This is somewhat orthogonal issue, and can be > solved longer term, definitely not required for this release. But if your > git repo consisted only of source code then it could be ~~20MB rather than > 100MB, and auditing the IP would be much simpler. > > Julian > > > On May 25, 2015, at 9:14 AM, Luke Han <[email protected]> wrote: > > As double check, I would like to recall this vote release, the majority > reason is the source shouldn't contains binary file like Tomcat. > > I will re-generate package and raise vote again. > > Thanks. > > > Best Regards! > --------------------- > > Luke Han > > 2015-05-24 23:16 GMT+08:00 Luke Han <[email protected]>: > > Thanks Julian. I wil double check the RAT result. > > Some binary files are required for testing and UI, like the mini-cluster > file for unit test, if there's no conflict with Apache Policy, I would like > to continue to include them in the release, so that everyone could download > and run unit test successfully, otherwise they have to download from > somewhere first. > > @All, please help to verify and check, and leave your comments here. > > I can generate RC2 if some changes should to be applied before release. > > Thank you very much. > > > > > Best Regards! > --------------------- > > Luke Han > > 2015-05-23 2:53 GMT+08:00 Julian Hyde <[email protected]>: > > +0 There are issues with rat and binary files. I would like to hear from > others whether these can be ignored in this first release. > > Now the details... > > The email has all of the necessary information, and the structure of the > zip looks good. > > If I understand the maven-release-plugin correctly, the hash of the > release > is actually b6af86cb047c8a160981b017dae2e57a8b923122, which is the parent > of 56608ec69bc15a1467ffe689bd613d8daad9f02c. > > Checked LICENSE and NOTICE are present in jars. > > Checked LICENSE, DISCLAIMER and NOTICE are present in source distribution. > > Verified GPG key and hashes. > > Builds under JDK 1.7 on ubuntu-15.04. I didn’t try running the test suite, > or JDK 1.8. > > Some files have spaces in their names. Not high priority, but consider > removing the spaces. They create havoc for utilities such as “find”. > > There are lots of binary files (.snapshot, hbase-export.tar.gz, .svg, …). > And you are distributing tomcat with all its jars (both zipped and > unzipped). And there is kylin.war (i.e. your own binaries). > > As a result of those binaries, it is a much larger download than I’d > expect > for a source release. > > I ran “mvn apache-rat:rat” and it is still running after 10 minutes. Maybe > it is choking on one of those binary files. I ran on the modules > individually: > > for i in atopcalcite common metadata dictionary cube job storage query > server jdbc invertedindex; do (cd $i && echo && echo $i && mvn > apache-rat:check); done > > and it completed, but there were errors in most modules. > > Julian > > > On May 22, 2015, at 9:31 AM, Luke Han <[email protected]> wrote: > > Hi all, > > I have created a build for Apache Kylin 0.7.1-incubating, release > candidate > 1. > > Thanks to everyone who has contributed to this release. > This is Apache Kylin first release attempt, please help to verify and > let’s > know if there’s any issue. > Here’s release notes: > > > https://github.com/apache/incubator-kylin/blob/0.7.1-staging/docs/release_notes.md > > The commit to be voted upon: > > > https://github.com/apache/incubator-kylin/commit/56608ec69bc15a1467ffe689bd613d8daad9f02c > > > Its hash is 56608ec69bc15a1467ffe689bd613d8daad9f02c > < > > > https://github.com/apache/incubator-kylin/commit/56608ec69bc15a1467ffe689bd613d8daad9f02c > > > . > > The artifacts to be voted on are located here: > > > https://dist.apache.org/repos/dist/dev/incubator/kylin/apache-kylin-0.7.1-incubating-rc1/ > > The hashes of the artifacts are as follows: > src.zip.md5 3995197634625eb6aae35926e9216376 > src.zip.sha1 148b95f047d7f3134f6bc68df78d2e1b4a3688dd > > A staged Maven repository is available for review at: > https://repository.apache.org/content/repositories/orgapachekylin-1002/ > > Release artifacts are signed with the following key: > https://people.apache.org/keys/committer/lukehan.asc > > Please vote on releasing this package as Apache Kylin 0.7.1-incubating. > > The vote is open for the next 72 hours and passes if a majority of > at least three +1 PPMC votes are cast. > > [ ] +1 Release this package as Apache Kylin 0.7.1-incubating > [ ] 0 I don't feel strongly about it, but I'm okay with the release > [ ] -1 Do not release this package because... > > > Here is my vote: > > +1 (binding) >
