On Mon, Mar 29, 2010 at 12:13 PM, Simon Nash <[email protected]> wrote: > Simon Nash wrote: >> >> Simon Nash wrote: >>> >>> Simon L made the following comment on TUSCANY-3508 (see [1]): >>>> >>>> Do we need the binary version of the travel sample distribution? It >>>> seems >>> >>> > that people interested in the sample would want to look at the source >>> code. >>> > Just running it is not very interesting as, as is often the way with >>> samples, >>> > it doesn't really do anything useful in its own right. >>> >>> I don't think we need the binary distribution as a separately released >>> artifact. >>> However I do think we need to retain the distribution directory within >>> the >>> sample source tree so that people can build the binary distribution for >>> themselves and see what the executable binary artifacts look like. >>> >>> This approach would suggest that using the name "distribution" for this >>> directory might not be ideal. Other possibilities for this directory >>> name >>> are "binaries", "executables", "runtime", ??? >>> >>> Of the above options I would have a slight preference for "binaries". >>> >>> Thoughts, other suggestions? >>> >>> Simon >>> >>> [1] >>> https://issues.apache.org/jira/browse/TUSCANY-3508?focusedCommentId=12847769&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#action_12847769 >> >> On reflection we can't rename the "distribution" directory because this >> is also used to produce the source distribution. We could do the >> following: >> >> 1. Split the current "distribution" directory into two directories at >> the same level, one called "distribution" which creates the source >> distribution and one called "binaries" which creates what we are >> currently calling the binary distribution. >> >> 2. Take "distribution" out of the top-level maven build, so that when >> users build the travel sample they don't create an extra copy of >> the source distribution. >> >> 3. Add "binaries" to the top-level maven build, so that when users build >> the travel sample they get a complete set of the binary artifacts >> needed to run it. >> >> I can make these changes if we agree that this is the right approach. >> >> Simon >> >> > Any thoughts or comments on the above proposal? > > Simon > >
+1 for 1 and 2. If we do step 3 and build binaries is the suggestion that users run the samples from the binaries? It may be good to have this as an option but I'd rather we stick with the source distro based ant scripts primarily. It seem more straightforward. Simon -- Apache Tuscany committer: tuscany.apache.org Co-author of a book about Tuscany and SCA: tuscanyinaction.com
