First of all Vladimir, please cut out the sarcasm. It's not constructive. We can't claim that some of the files in our release are not Calcite sources. There is no other definition of "Calcite source code" than what we include in a release. So, everything in that tar file is by definition "Calcite source code".
There are some kinds of files, such as .jpeg, which do not exist in any form other than binary. One creates them by editing them directly. Since we have created them, and they are available nowhere else, they belong in the source release. There is a case to be made that we should omit files like EMPS.csv.gz. If you think it's important, go ahead and make that case. Code is different. The textual source is editable, but the object files (in this case the .class files in the .jar) are not. The open source definition makes a big deal about distributing source code, because it is not practical to improve on someone's work if your only means to change it is to edit their object files. Apache release policy bans object files because they are difficult to audit (for mischief such as IP infringement or malware). It's not really possible to vote on a release that contains object files. Lastly, this .jar file is non-essential. The release builds just fine without it. Julian On Fri, Mar 15, 2019 at 7:24 AM Vladimir Sitnikov <[email protected]> wrote: > > Julian> The Release Policy [1] states that for a (compiled) artifact > the source code has to be released alongside. > > maven-wrapper.jar has nothing to do with Calcite's source codes, so > maven-wrapper.jar is not "a compiled version of the source", so [1] > does not apply here. > It is the very same thing as "binary font file". > > I don't see why the policy would allow *.ttf and forbid *.jar at the > same time. It is basically the same thing as long as the binary comes > from somewhere else. > Of course, we should not include calcite.jar into source distribution. > However, maven-wrapper.jar is not something that is produced by > Calcite's sources, so it should be fine. > > Vladimir
