Niall Pemberton wrote: > On Wed, Nov 25, 2009 at 11:39 PM, Niall Pemberton > <niall.pember...@gmail.com> wrote: >> On Wed, Nov 25, 2009 at 11:31 PM, Phil Steitz <phil.ste...@gmail.com> wrote: >>> Niall Pemberton wrote: >>>> On Wed, Nov 25, 2009 at 10:23 PM, Paul Benedict <pbened...@apache.org> >>>> wrote: >>>>> Phil, >>>>> >>>>> I don't think you should be modifying the version (and groups, really) >>>>> here. All the artifacts belong to version 1.3. >>>>> >>>>> Maven does have a concept of a qualifier, but according to Sonatype, >>>>> it's only to capture milestone builds: >>>>> http://www.sonatype.com/books/maven-book/reference/pom-relationships-sect-pom-syntax.html >>>> I don't think this is true maven has used "classifier" to distribute >>>> various artifacts that are attached to the project - such as >>>> "sources", "javadocs", test jar and it talks about them here in the >>>> same book >>>> >>>> http://www.sonatype.com/books/maven-book/reference/assemblies-sect-output-algorithm.html#assemblies-sect-transitive >>>> >>>> Also its been a fairly common pratice with many projects using a maven >>>> build to provide JDK 1.4 compatible jars after the project moved to >>>> JDK 1.5 using some kind of classifier - this is pretty much the same >>>> situation. >>>> >>>> If you use a different artifactId for the different jars then its >>>> going to be a bigger PITA for the release - since you'll need a pom >>>> and have to update maven-metadata.xml - probably anually. This is what >>>> happened in BeanUtils and doing a release is much more painful and >>>> prone to errors. >>> Stupid question. Assuming we go the classifier route, how can I use >>> just one pom? I was assuming I would have to hack a second pom in >>> either case. >> AFAIK you don't have to do anything - just produce the additional jars >> with the classifier in the name - its people who consume it who >> specifiy the classifier - for example say you produce an additional >> jar called commons-dbcp-1.3-jdbc3.jar then if someone wanted to use >> that rather than the standard commons-dbcp-1.3.jar then they would >> specify the dependency as follows: >> >> <dependency> >> <groupId>commons-dbcp</groupId> >> <artifactId>commons-dbcp</artifactId> >> <version>1.3</version> >> <classifier>jdbc3</classifier> >> </dependency> >> >> Haven't read it, but also found this: >> >> http://www.sonatype.com/books/maven-book/reference/profiles-sect-platform-classifier.html > > Found an example subethasmtp-smtp has a JDK 1.4 jar: > > http://repo2.maven.org/maven2/org/subethamail/subethasmtp-smtp/1.2/ > > And Commons Email 1.2 depends on the JDK 1.4 jar: > > http://repo2.maven.org/maven2/org/apache/commons/commons-email/1.2/commons-email-1.2.pom
Thanks, Niall! > > Niall > >> Niall >> >> >> >>> Phil >>>> I would go down the classifer route. >>>> >>>> Niall >>>> >>>>> What you have, simply, is, different artifacts. Keep the same groupId >>>>> and version, just alter the artifact names. >>>>> >>>>> JDBC 4 version (JDK 1.6) >>>>> groupId = org.apache.commons >>>>> artifactId = commons-dbcp >>>>> version = 1.3 >>>>> >>>>> JDBC 3 version (JDK 1.4-1.5) >>>>> groupId = org.apache.commons >>>>> artifactId = commons-dbcp-jdbc3 >>>>> version = 1.3 >>>>> >>>>> Paul >>>>> >>>>> On Wed, Nov 25, 2009 at 4:14 PM, Phil Steitz <phil.ste...@gmail.com> >>>>> wrote: >>>>>> Phil Steitz wrote: >>>>>>> I am about to roll an RC and I need to make sure all are OK with the >>>>>>> artifact names and repo placement >>>>>>> >>>>>>> JDBC 4 version (JDK 1.6) >>>>>>> groupId org.apache.maven >>>>>> Oops! I obviously mean commons above :) >>>>>>> artifactID commons-dbcp >>>>>>> version 1.3 >>>>>>> >>>>>>> JDBC 3 version (JDK 1.4-1.5) >>>>>>> groupId commons-dbcp >>>>>>> artifactId commons-dbcp >>>>>>> version 1.3-jdbc3 >>>>>>> >>>>>>> Giving the 1.3 name to the 1.6 version makes sense as this is the >>>>>>> main development version. Moving it gets it into compliance with >>>>>>> the maven standard and avoids unintended consequences of upgrading >>>>>>> for 1.4-1.5 users by requiring a bigger change. >>>>>>> >>>>>>> Alternatively, we could put descriptors on both and leave placement >>>>>>> as is. Opinions please. >>>>>>> >>>>>>> Phil >>>>>> --------------------------------------------------------------------- >>>>>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>>>>> For additional commands, e-mail: dev-h...@commons.apache.org >>>>>> >>>>>> >>>>> --------------------------------------------------------------------- >>>>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>>>> For additional commands, e-mail: dev-h...@commons.apache.org >>>>> >>>>> >>>> --------------------------------------------------------------------- >>>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>>> For additional commands, e-mail: dev-h...@commons.apache.org >>>> >>> >>> --------------------------------------------------------------------- >>> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org >>> For additional commands, e-mail: dev-h...@commons.apache.org >>> >>> > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
