Here are few thoughts based on experiences with past projects - 1. Headers for data files - We have had csv & txt files without headers for Drill as well. This is for the data files which do not understand headers.
3. README File - We can add deploy instructions on starting server and client cli etc. This would enable user to quickly play around with the project. Few projects also keep an INSTALL file along with README that has installation specific information. 4. MD5 Hash- Few ideas from Aurura https://issues.apache.org/jira/browse/AURORA-1069 5. Dev credits - Many projects give credits on the project's official website rather than pom file. Thanks, Yash On Wed, May 6, 2015 at 4:05 PM, amareshwarisr . <[email protected]> wrote: > Hi all, > > Here are some issues reported in release vote. Starting a discuss thread to > decide what needs to be done. > > > ---------- Forwarded message ---------- > From: Justin Mclean <[email protected]> > > >> - All source files have headers (there’s a few .data files that could > possibly have a header added) > > Data files are data that is loaded in to tables during tests/examples, > which may not understand headers. > Can anyone have any ideas on what we can do here? > > >> I did have issues getting the tests to run with this error, but I’d > assume it’s just my setup. > "There is no database named testqueryrewrite)” > > AFAIK, i think this is not a test failure issue, but the first time > creation of db might be giving this error. We can create tracking jira to > remove such exceptions from console. > > >> You may want to add a little more information the README on running the > tests and what is required. You might also want to suggest recommended > values for MAVEN_OPTS. > > Default MAVEN_OPTS should work for lens. The only thing i know of test > setup is having JAVA_HOME defined. Anyone else have more ideas here on what > can be improved? > > > >> MD5 content is oddly formatted (make comparison a little harder) > I was using 'gpg --print-md5' and 'gpg --print-sha512' for generating > hashcodes. Anyone know of other options ? > > >> Signed email has "(Works for Hadoop and ecosystem)” which is probably > not needed > I gave it in my key when i first created. Now that it is signed by many > people, Will it possible to modify it? > > >> Pom.xml lists developers and their companies, software is developed by > individuals at the the ASF > > Many projects have the affiliations mentioned, should we really remove > affiliations from developers? > > Thanks > Amareshwari >
