From: thomas at tada.se (Thomas Hallgren) Date: Thu, 07 Feb 2013 23:31:58 +0100 Subject: [Pljava-dev] Building pljava.so with maven-nar-plugin? In-Reply-To: <20130207170645.080294f4@james.intern> References: <20130111204013.00e19939@james.intern> <20130112234738.03b606b1@james.intern> <50F29266.5070408@tada.se> <50F29372.7080808@tada.se> <20130206023102.1b94cc01@james.intern> <51120BE7.6020109@tada.se> <20130207170645.080294f4@james.intern> Message-ID: <51142B5E.2080906@tada.se> On 2013-02-07 17:06, Christian Hammers wrote: > Am Wed, 06 Feb 2013 08:53:11 +0100 > schrieb Thomas Hallgren : > >> On 2013-02-06 02:31, Christian Hammers wrote: >>> Hello >>> >>> I played around with the maven-nar-plugin that was recommended on >>> this list. The below pom.xml does produce .h files from the java >>> classes and also compiles the C files. What's missing though is a >>> way to get a .h from "java.sql.Types" as this class is in the JRE's >>> rt.jar and not in the specified . Any ideas? >> Not sure why that matters. Executing: >> >> javah java.sql.Types >> >> without giving it a classpath does produce the java_sql_Types.h file >> without problems. Is the current javah configuration deliberately >> stripping off the java runtime? >> >> - thomas > The javah target of the maven-nar-plugin only scans a single > directory for actual *.class files, reads their > method signatures and only adds the class if there is at least one > "native" method. The parameter does not help here. > > I added a parameter to it that just adds its values > to the end of the javah call and with that it would work. > My pull requests for the maven-nar-plugin are here: > * https://github.com/maven-nar/maven-nar-plugin/pull/22 > * https://github.com/maven-nar/maven-nar-plugin/pull/21 > > bye, Cool. Let's hope you get those accepted. And thanks for your effort of making this happen. - thomas