[ https://issues.apache.org/jira/browse/SIS-337?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15840132#comment-15840132 ]
Martin Desruisseaux commented on SIS-337: ----------------------------------------- In order to try to address the need better, I would like to know: for applications that need an embedded EPSG database, is there a need for the full EPSG database or do such applications typically know in advance the EPSG codes they are interested on? I was wondering if it would be useful to have a mechanism like that (the {{@PrefetchCRS}} annotation does not exist yet - this part of the proposal for which I would like to evaluate the interest): {code:java} @PrefetchCRS("EPSG:3395") CoordinateReferenceSystem myCRS; {code} An annotation processing tool provided by Apache SIS would inject automatically the necessary code for initializing {{myCRS}} to the EPSG:3395 Coordinate Reference System. The developer would need the EPSG database at build time, but not at runtime; the CRS definition would be copied for example using Java object serialization. If there is any {{CoordinateOperation}} explicitly defined in the EPSG database for some pair of CRS declared in different {{@PrefetchCRS}} annotation, they would also be copied. Of course this approach would be useful only if an application is interest in a relatively small set of CRS known at build time. > Package EPSG Derby DB in sis-epsg jar, eliminating the need for external > SIS_DATA dir > -------------------------------------------------------------------------------------- > > Key: SIS-337 > URL: https://issues.apache.org/jira/browse/SIS-337 > Project: Spatial Information Systems > Issue Type: Wish > Components: Referencing > Affects Versions: 0.7 > Reporter: Kevin Mehall > > Since [Derby > supports|https://db.apache.org/derby/docs/10.7/devguide/cdevdeploy11201.html] > read-only databases in .jar files, is there a reason that the > `org.apache.sis.non-free:sis-epsg` package includes SQL source to create an > external database, rather than the database itself? > It would make deployment much easier if there were no need for a SIS_DATA > directory and environment variable, and adding the dependency were the only > thing necessary. -- This message was sent by Atlassian JIRA (v6.3.4#6332)