On 24 February 2012 12:11, Andy Seaborne <a...@apache.org> wrote: > >>> We distribute through maven but also many of our users are not >>> experienced >>> developers but students, including ones new to java. >>> >>> To make it as easy as possible for this category of user, we ship a >>> distribution which is the collection of jars needed for use without a >>> maven/ivy infrastructure. >> >> >> That's fine, so long as the N&L files agree with the contents. > > > They should do. > > >>> This is created in the maven target/ area - and it is then renamed into >>> the >>> dist/ area as apache-jena* by the distribution build script. Maven >>> forces >>> the name to be jena-tdb-.... on staging whatever assembly root name is >>> given. >> >> >> Again that's OK, but the layout of the dist area - and the naming >> convention - is very confusing. >> >> There is no direct correspondance between the binary and source archives. >> It should be immediately obvious how to find the source and >> corresponding binary, but that's not the case at present. > > > The directory structures align: > > /download/jena-tdb-0.9.0-incubating > /source-release/jena-tdb-0.9.0-incubating
I see the following files (amongst others) in the RC area http://people.apache.org/~andy/dist-jena-tdb-0.9.0-incubating-RC-4/ KEYS download/apache-jena-tdb-0.9.0-incubating.zip download//jena-tdb-0.9.0-incubating/jena-tdb-0.9.0-incubating.jar source-release/jena-tdb-0.9.0-incubating/jena-tdb-0.9.0-incubating-source-release.zip Note that the source release is in source-release/jena-tdb-0.9.0-incubating/ whereas the corresponding binary release is in download/ i.e. different depths - and they have different names. [I'm discounting the jar, since that is Maven-specific and is not needed on the dist/ mirror site.] > but since this isn't "/binaries" and "/sources" we'll build about RC - we'll > be back in a week to 10 day I hope. Does not have to be binaries and source (sic), so long as the source and binary archives are directly related. They have different names at present, which makes it difficult to tie them together. Also, the convention is to have the source and binary archives: - in the same directory, or - in parallel directories (binaries/source) In the case of OS-specific binaries, these could be in folders such as binaries/plan9 binaries/windows etc. In such cases, the binary archives would be one level lower than the source archive, but that is easy to follow. > >> The scripts can all have AL headers; there are some XML files as well >> that could have them. > > > I will put either short or full AL headers on the scripts. > > find . -name \*xml | xargs grep -L "Licensed to the Apache Software > Foundation" | xargs wc -l > > 39 ./testing/Deployment/pom.xml > 8 ./src/main/java/com/hp/hpl/jena/tdb/tdb-properties.xml > 8 ./src/main/resources/org/apache/jena/tdb/tdb-properties.xml > 55 total > > and tdb-properies.xml has 2 Java properties. > > Andy > > >> There is one Java file that has no AL header. >> >>> 12 are NOTICE/LICENSE/DEPENDENCIES etc. >> >> >> These are OK >> >>> and then there setting and eclipse files that you have commented on >>> before. >>> >>> (aside: we tried autogenerating Eclipse files following previous comments >>> but encountered problems with eclipse:eclipse which we are investigating) >> >> >> No need to autogenerate; just don't store them with the default names or >> paths. >> >> Perhaps store them under a resources/eclipse directory. >> >>> >>> Andy > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org > For additional commands, e-mail: general-h...@incubator.apache.org > --------------------------------------------------------------------- To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org For additional commands, e-mail: general-h...@incubator.apache.org
