[forge-dev] Fw: Re: Mojavelinux.com Feedback: seam-gen and cdi/weld
Max Andersen
manderse at redhat.com
Sat Apr 16 15:18:40 EDT 2011
> If persistence.xml only has a JNDI connection string, then you'll need to access that connection in order to connect to the database. Over-engineered? possibly. Alternatives to connecting w/JNDI?
>
Properties.
Like rails - prod, dev, test settings.
Like eclipse - dtp connections you can use too.
I.e. Your deployed datasource might nit even be able to allow reading metadata.
It's really not something that would work well without some explicit configuration.
> And what if your code is not compilable ?
>
> Your code is never deployed to the appserver. Only Forge-generated code would ever be deployed in this manner.
If you use custom naming strategies, usertypes And so forth the user classes will need to be deployed with it.
/max
> ~Lincoln
>
> On Sat, Apr 16, 2011 at 1:18 PM, Max Andersen <manderse at redhat.com> wrote:
> Why would you want to require a running and fully confugured Appserver for doing reverse engineering ?
>
> Sounds like a very overengineered approach ?
>
> And what if your code is not compilable ?
>
>
> /max (sent from my phone)
>
>
> On 16/04/2011, at 16.24, "Lincoln Baxter, III" <lincolnbaxter at gmail.com> wrote:
>
>> In theory, it should be possible to gain access to the AppServer JNDI connections by:
>>
>> Use Forge itself to inspect persistence.xml
>> If it detects JNDI connection settings, generate an EE webapp with the persistence.xml and hibernate-tools embedded inside.
>> Deploy the webapp, which then generates the entities for the project.
>> Undeploy and destroy the temp project.
>> Paul Bakker is using a similar technique to generate deployments using Arquillian:
>>
>> https://issues.jboss.org/browse/SEAMFORGE-109
>>
>> ~Lincoln
>>
>> On Sat, Apr 16, 2011 at 2:48 AM, Max Rydahl Andersen <max.andersen at redhat.com> wrote:
>> The persistence.xml just needs to be on the classpath at META-INF/persistence.xml as defined by the spec.
>> (assuming you use jpaconfiguration)
>>
>> But how often will a persistence.xml actually have the database connection info ?
>>
>> if it is the persistence.xml used for deploying then it will contain the datasource name and
>> nothing else aka. not enough info for hibernate tools nor hibernate it self to connect "outside" the server.
>>
>> This is why in hibernate tools (both ant and eclipse version) we allow you to specify a properties file
>> to use that will override anything found in persistence.xml on the classpath/default hibernate configuration semantics.
>>
>> Until Forge it self gets a notion of database connection info I would suggest using this separate properties file
>> for development connection info would be the simplest/easiest approach ? (we even have a "standard" name for
>> it with default content to get the best non-invasive behavior)
>>
>> here is what seam 2 projects made by seamgen and/or jboss tools looks like:
>> #File used by hibernate tools to override <datasource> and other container specific settings in persistence.xml
>> hibernate.connection.password=
>> hibernate.connection.username=sa
>> hibernate.connection.driver_class=org.hsqldb.jdbcDriver
>> hibernate.dialect=org.hibernate.dialect.HSQLDialect
>> hibernate.connection.url=jdbc:hsqldb:.
>>
>> hibernate.connection.provider_class=org.hibernate.connection.DriverManagerConnectionProvider
>> hibernate.datasource=
>> hibernate.transaction.manager_lookup_class=
>>
>> The default (bad?) name is hibernate-console.properties.
>>
>> Notice, the last three elements would probably be better off being defaulted in forge plugin to ensure
>> hibernate won't start doing any of its JEE integration code just because you have a peristence.xml with
>> datasource and jta tx info in it.
>>
>> /max
>>
>> On Apr 15, 2011, at 18:55, brian wrote:
>>
>>
>>
>> >
>> > hi L3 and john.
>> >
>> > I was thinking about the persistence.xml. Would using piping make sense
>> > to read/parse the fields from the persistence.xml and then pass them to
>> > GenerateEntities?
>> >
>> > so "read-persistence | generate-entities" ?
>> >
>> > ...I'm not sure if a plugin can read a file, but if it could, a little
>> > file-reader/parser might work?
>> >
>> >
>> > On Fri, 2011-04-15 at 12:04 -0400, Lincoln Baxter, III wrote:
>> >> Hi John!
>> >>
>> >> Thank you for your interest! I think it's fair to say that we are
>> >> "close" to having a fully-functional database reverse-engineering
>> >> plugin, but we are not there yet. Brian (ccd) has been doing some work
>> >> recently (when he has time) to get Max Andersen's prototype up and
>> >> running. The prototype is here, but has a few shortcomings:
>> >>
>> >> https://github.com/forge/plugin-hibernate-tools
>> >> 1. Connection settings are hardcoded
>> >> 2. JDBC drivers are not included with the plugin
>> >> 3. Does not take persistence.xml into account, if it exists
>> >> In short, it works, but is not usable yet; however, these are all
>> >> things that can be worked on if you are interested! And I would be
>> >> more than happy to help you get together with Brian and get two brains
>> >> working on this! Would you be intersted in that?
>> >>
>> >> No matter your level of commitment, I would recommend joining the dev
>> >> and users lists:
>> >> * https://lists.jboss.org/mailman/listinfo/forge-dev
>> >> * https://lists.jboss.org/mailman/listinfo/forge-users
>> >> Just let me know and I'll get you all the info that you'll need :)
>> >> ~Lincoln
>> >>
>> >> On Thu, Mar 17, 2011 at 8:57 PM, John du Clos <duclosjj at yahoo.com>
>> >> wrote:
>> >> Hi Lincoln-
>> >>
>> >> I was chatting with Dan Allen and he mentioned you are working
>> >> on Forge and specifically the reverse engineering feature from
>> >> the database. I wanted to reach out to you to see if you have
>> >> any early versions you need testers to help you verify this
>> >> feature.
>> >>
>> >> I look forward to hearing from you!
>> >> John
>> >>
>> >> --- On Wed, 3/16/11, Dan Allen <dan.j.allen at gmail.com> wrote:
>> >>
>> >>
>> >> From: Dan Allen <dan.j.allen at gmail.com>
>> >> Subject: Re: Mojavelinux.com Feedback: seam-gen and
>> >> cdi/weld
>> >> To: duclosjj at yahoo.com
>> >> Date: Wednesday, March 16, 2011, 9:33 PM
>> >>
>> >> Hey John,
>> >> Thanks for reading the book. I'm glad it helped you
>> >> along. I know about those painful days of J2EE you are
>> >> talking about.
>> >> The best way to get started atm is using the Maven
>> >> archetype.
>> >> http://tinyurl.com/gojavaee
>> >> That gives you a starter project. We don't yet have
>> >> the reverse engineering yet for this environment. That
>> >> what Forge will offer soon. Check out the issue
>> >> tracker or mailing list for updates.
>> >> Hope that gets you started. Cheers,
>> >> - Dan Allen
>> >> Sent from my Android-powered phone:
>> >> An open platform for carriers, consumers and
>> >> developers
>> >> On Mar 16, 2011 9:04 PM, <duclosjj at yahoo.com> wrote:
>> >>> Submitted by: John duClos <duclosjj at yahoo.com> on
>> >> Wednesday, March 16th, 2011 @ 9:04:25 pm (-0400)
>> >>>
>> >>> Online Form Fields
>> >>> ------------------
>> >>>
>> >>> Url:
>> >>>
>> >>>
>> >>> Message:
>> >>> Dan-
>> >>>
>> >>> I purchased your Seam book when it was published and
>> >> am a supporter of
>> >>> the Seam framework since I lived the challenging
>> >> days of early J2EE
>> >>> development.
>> >>>
>> >>> I work with database centric solutions and seam-gen
>> >> allowed me to
>> >>> develop solutions in record time by getting my
>> >> project started very
>> >>> quickly. I'd like to start implementing JEE 6/CDI
>> >> with JBoss AS 6, but
>> >>> I can't seem to find a way to issue the seam-gen
>> >> command.
>> >>>
>> >>> I looked at Forge but that is not where seam-gem
>> >> was. Do you know of a
>> >>> valid seam-gen solution for JEE 6/CDI based projects
>> >> (Seam 3).
>> >>>
>> >>> Thanks,
>> >>> John
>> >>>
>> >>> Client Variables
>> >>> ----------------
>> >>>
>> >>> HTTP_USER_AGENT:
>> >>> Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.0;
>> >> Trident/4.0; SLCC1; .NET CLR 2.0.50727; .NET CLR
>> >> 3.5.21022; .NET CLR 3.5.30729; MDDS; .NET CLR
>> >> 3.0.30729; MS-RTC LM 8; AskTB5.4)
>> >>>
>> >>>
>> >>
>> >>
>> >>
>> >>
>> >> --
>> >> Lincoln Baxter, III
>> >> http://ocpsoft.com
>> >> http://scrumshark.com
>> >> "Keep it Simple"
>> >
>> >
>> > _______________________________________________
>> > forge-dev mailing list
>> > forge-dev at lists.jboss.org
>> > https://lists.jboss.org/mailman/listinfo/forge-dev
>>
>> /max
>> http://about.me/maxandersen
>>
>>
>>
>>
>>
>>
>> --
>> Lincoln Baxter, III
>> http://ocpsoft.com
>> http://scrumshark.com
>> "Keep it Simple"
>
>
>
> --
> Lincoln Baxter, III
> http://ocpsoft.com
> http://scrumshark.com
> "Keep it Simple"
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/forge-dev/attachments/20110416/6e103382/attachment-0001.html
More information about the forge-dev
mailing list