On 6 September 2018 at 17:24, Thomas Vandahl <t...@apache.org> wrote: > On 27.08.18 18:12, Benedikt Ritter wrote: >> ~/w/a/r/r/t/jcs git:(upstream ⚡ master) > diff -r commons-jcs-2.2.1-RC4/ >> commons-jcs-2.2.1-src >> Only in commons-jcs-2.2.1-RC4/: .gitignore >> Only in commons-jcs-2.2.1-RC4/: .svn >> Only in commons-jcs-2.2.1-RC4/: .travis.yml >> Only in commons-jcs-2.2.1-src: LICENSE >> Only in commons-jcs-2.2.1-RC4/: LICENSE.txt >> Only in commons-jcs-2.2.1-src: NOTICE >> Only in commons-jcs-2.2.1-RC4/: NOTICE.txt >> Only in commons-jcs-2.2.1-RC4/: auxiliary-builds >> Only in commons-jcs-2.2.1-RC4/: checkstyle.xml >> Only in commons-jcs-2.2.1-RC4/: commons-jcs-sandbox >> Only in commons-jcs-2.2.1-RC4/: init-git-svn.sh >> Only in commons-jcs-2.2.1-RC4/: jcache-fast.sh >> Only in commons-jcs-2.2.1-RC4/: maven-eclipse-codestyle.xml >> >> This should be fixed for the next release. > > Is it absolutely necessary that the tag contains the exact same files as > the distribution or would it be sufficient that the tag is able to > *build* the identical distribution? That has always been my interpretation.
Reviewers need to be able to account for any discrepancies between the contents to ensure that: - the dist does not contain spurious files - the dist is not lacking any required files. The source distribution also has to be able to build itself. If the tag contains files that are not in the distribution, why are they present at all? If they are needed, why are they not in the distribution? Lots of discrepancies means lots of checking. > Bye, Thomas > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
