11:02 a.m.
Marcus, yep, this is basically exactly what I did. Now I need to spend some
more time with the API and testing, etc. And unfortunately, today is not a
good day... So thx for the zip offer but I don't think I'll need it for
now. Eventually I'd like to run some load tests to compare performance vs.
iBATIS and Spring JDBC equivalents. I'll need to come up with a
plan/roadmap for this sub-project first and then I'll post that on this
list.
On Thu, Nov 26, 2009 at 12:12 AM, Marcus Smedman <marcus(a)smedman.org> wrote:
Hi Arbi,
Not sure if this is what you where referring to, but I got it up and
running by doing this:
1. Create an Eclipse project and add the source from seam-sql.zip
2. Download weld from http://seamframework.org/Download
3. Unpack and add the following libs to the project:
* <weld-x.y.z>\artifacts\weld\weld-se.jar
* <weld-x.y.z>\artifacts\weld\weld-servlet.jar
4. Add a hsql jdbc driver to the project (from
http://sourceforge.net/projects/hsqldb/files/ or
<jboss-seam-x.y.z>\lib\hsqldb.jar)
5. Uncomment the eg.Main.main method and execute it. Generates the
following output:
create table users (name varchar not null, username varchar not null
primary key, password varchar not null)
insert into users (name, username, password) values (?, ?, ?)
update users set name=?, password=? where username=?
select u.name as n from users as u where (u.username=? and
u.password=?)
Gavin King
I have an Eclipse proj packed up in a zip if you’d like it (4MB).
Regards
Marcus
*From:* weld-dev-bounces(a)lists.jboss.org [mailto:
weld-dev-bounces(a)lists.jboss.org] *On Behalf Of *Arbi Sookazian
*Sent:* den 26 november 2009 08:41
*To:* Gavin King
*Cc:* Weld-Dev
*Subject:* Re: [weld-dev] SQL framework proof of concept
I'm interested. Plz provide instructions on how to build/deploy so I can
test it out (it's getting late so I haven't even opened the project in
Eclipse yet). I noticed that the public static void main() in Main class is
commented out. Is it safe to assume that this is basically a Weld SE app
(or soon to be Seam3) to test with? thx.
On Wed, Nov 25, 2009 at 11:02 PM, Gavin King <gavin.king(a)gmail.com> wrote:
Here's a proof-of concept for the typesafe SQL framework. There's lots
of stuff missing, but I can already do basic queries, inserts and
updates. Anyone want to take over from here?
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
_______________________________________________
weld-dev mailing list
weld-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/weld-dev
2:16 p.m.
New subject: SQL framework proof of concept
OK. I played around with it a bit, but it's all pretty new to me. I mean not
only the seam-sql, but the whole JEE stack and everything that comes with
it. I've monitored this mail list for a while though and it's very
interesting and I really learn a lot.
Regarding the seam-sql I added a Drop (to speed up my small tests), Alter
(to add foreign keys) and a Join (a bit messy when/if using aliases) to try
to understand how Gavin had did the other stuff. There's a lot of stuff I
don't fully understand. (is a join an Expression for example?).
Anyway, my idea was to make some tests using (some of) the tables added when
running add-identity-management by seam-gen (user_account, user_role and
user_account_role). I have tested it (DROP/ALTER/JOIN) on MySQL and it
seems to work, but I'm not sure I fully understand how to follow the
"pattern" that Gavin have put together.
Anyway, if I can help in any way please let me know, I might get some time
to help with tests or whatever.
Regards
Marcus
From: Arbi Sookazian [mailto:asookazian@gmail.com]
Sent: den 26 november 2009 18:03
To: Marcus Smedman
Cc: Weld-Dev
Subject: Re: [weld-dev] SQL framework proof of concept
Marcus, yep, this is basically exactly what I did. Now I need to spend some
more time with the API and testing, etc. And unfortunately, today is not a
good day... So thx for the zip offer but I don't think I'll need it for
now. Eventually I'd like to run some load tests to compare performance vs.
iBATIS and Spring JDBC equivalents. I'll need to come up with a
plan/roadmap for this sub-project first and then I'll post that on this
list.
On Thu, Nov 26, 2009 at 12:12 AM, Marcus Smedman <marcus(a)smedman.org> wrote:
Hi Arbi,
Not sure if this is what you where referring to, but I got it up and running
by doing this:
1. Create an Eclipse project and add the source from seam-sql.zip
2. Download weld from http://seamframework.org/Download
3. Unpack and add the following libs to the project:
* <weld-x.y.z>\artifacts\weld\weld-se.jar
* <weld-x.y.z>\artifacts\weld\weld-servlet.jar
4. Add a hsql jdbc driver to the project (from
http://sourceforge.net/projects/hsqldb/files/ or
<jboss-seam-x.y.z>\lib\hsqldb.jar)
5. Uncomment the eg.Main.main method and execute it. Generates the following
output:
create table users (name varchar not null, username varchar not null
primary key, password varchar not null)
insert into users (name, username, password) values (?, ?, ?)
update users set name=?, password=? where username=?
select u.name as n from users as u where (u.username=? and u.password=?)
Gavin King
I have an Eclipse proj packed up in a zip if you'd like it (4MB).
Regards
Marcus
From: weld-dev-bounces(a)lists.jboss.org
[mailto:weld-dev-bounces@lists.jboss.org] On Behalf Of Arbi Sookazian
Sent: den 26 november 2009 08:41
To: Gavin King
Cc: Weld-Dev
Subject: Re: [weld-dev] SQL framework proof of concept
I'm interested. Plz provide instructions on how to build/deploy so I can
test it out (it's getting late so I haven't even opened the project in
Eclipse yet). I noticed that the public static void main() in Main class is
commented out. Is it safe to assume that this is basically a Weld SE app
(or soon to be Seam3) to test with? thx.
On Wed, Nov 25, 2009 at 11:02 PM, Gavin King <gavin.king(a)gmail.com> wrote:
Here's a proof-of concept for the typesafe SQL framework. There's lots
of stuff missing, but I can already do basic queries, inserts and
updates. Anyone want to take over from here?
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
_______________________________________________
weld-dev mailing list
weld-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/weld-dev
2:36 p.m.
New subject: SQL framework proof of concept
On Thu, Nov 26, 2009 at 3:16 PM, Marcus Smedman <marcus(a)smedman.org> wrote:
Regarding the seam-sql I added a Drop (to speed up my small tests),
Alter
(to add foreign keys) and a Join (a bit messy when/if using aliases) to try
to understand how Gavin had did the other stuff. There’s a lot of stuff I
don’t fully understand… (is a join an Expression for example?).
Join is probably a subclass of From:
From<users> u = query.from(users.table);
Join<permissions> p = fromUser.join(permissions.table)
p.on( eq( u.get(users.id), p.get(permissions.userId) ) );
Does that make sense?
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
3:06 p.m.
New subject: SQL framework proof of concept
It does. I think. :) I'll give it some more thought and have a look at it
again, thanks for the pointer.
/Marcus
-----Original Message-----
From: Gavin King [mailto:gavin.king@gmail.com]
Sent: den 26 november 2009 21:37
To: Marcus Smedman
Cc: Arbi Sookazian; Weld-Dev
Subject: Re: [weld-dev] SQL framework proof of concept
On Thu, Nov 26, 2009 at 3:16 PM, Marcus Smedman <marcus(a)smedman.org>
wrote:
> Regarding the seam-sql I added a Drop (to speed up my small tests),
Alter
> (to add foreign keys) and a Join (a bit messy when/if using aliases)
to try
> to understand how Gavin had did the other stuff. There's a lot of
stuff I
> don't fully understand. (is a join an Expression for example?).
Join is probably a subclass of From:
From<users> u = query.from(users.table);
Join<permissions> p = fromUser.join(permissions.table)
p.on( eq( u.get(users.id), p.get(permissions.userId) ) );
Does that make sense?
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
3:15 p.m.
New subject: SQL framework proof of concept
Here, I implemented that (Join). WDYT?
On Thu, Nov 26, 2009 at 4:06 PM, Marcus Smedman <marcus(a)smedman.org> wrote:
It does. I think. :) I'll give it some more thought and have a
look at it
again, thanks for the pointer.
/Marcus
> -----Original Message-----
> From: Gavin King [mailto:gavin.king@gmail.com]
> Sent: den 26 november 2009 21:37
> To: Marcus Smedman
> Cc: Arbi Sookazian; Weld-Dev
> Subject: Re: [weld-dev] SQL framework proof of concept
>
> On Thu, Nov 26, 2009 at 3:16 PM, Marcus Smedman <marcus(a)smedman.org>
> wrote:
>
> > Regarding the seam-sql I added a Drop (to speed up my small tests),
> Alter
> > (to add foreign keys) and a Join (a bit messy when/if using aliases)
> to try
> > to understand how Gavin had did the other stuff. There's a lot of
> stuff I
> > don't fully understand. (is a join an Expression for example?).
>
> Join is probably a subclass of From:
>
> From<users> u = query.from(users.table);
> Join<permissions> p = fromUser.join(permissions.table)
> p.on( eq( u.get(users.id), p.get(permissions.userId) ) );
>
> Does that make sense?
>
>
>
> --
> Gavin King
> gavin.king(a)gmail.com
> http://in.relation.to/Bloggers/Gavin
> http://hibernate.org
> http://seamframework.org
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
3:47 p.m.
New subject: SQL framework proof of concept
Latest, with a new approach to specifying types.
5:16 p.m.
New subject: SQL framework proof of concept
New rev, with ability to pass SQL as strings.
On Fri, Nov 27, 2009 at 4:47 PM, Gavin King <gavin.king(a)gmail.com> wrote:
Latest, with a new approach to specifying types.
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
8:44 p.m.
New subject: SQL framework proof of concept
Following comments apply to seam-sql5 rev...
1) In Tuple class:
public <T> T get(int column, Type<T> type) {
return (T) get(column);
}
The type param is not referenced in the body of the method. Why?
2) So I added this method in the Tuple class:
@SuppressWarnings("unchecked")
public <T> T testGet(int column) {
return (T) get(column);
}
And changed testQuery3() to this:
public static void testQuery3(Session session) {
System.out.println("begin testQuery3()");
Iterable<Tuple> result = session.prepareQuery("select username,
name, password, age from users where username=?")
.set(0, "gavin", VARCHAR)
.result();
for (Tuple tuple: result) {
System.out.println( tuple.testGet(0));
System.out.println( tuple.testGet(1));
System.out.println( tuple.testGet(2));
System.out.println( tuple.testGet(3));
}
}
And got the same results in the console. I don't like to be forced to pass
the datatype in the get() method for Tuple, that should not be requierd.
3) This is looking pretty good to me now in terms of conciseness and
clarity:
@Inject LoginQuery loginQuery;
public void testLoginQuery(Session session) {
Iterable<Tuple> result =
session.prepare(loginQuery).set(loginQuery.username,
"gavin").set(loginQuery.password, "foobax").result();
for (Tuple tuple: result) {
System.out.println( tuple.get(loginQuery.name) + ' ' +
tuple.get(loginQuery.age)
+ ' ' + tuple.get(loginQuery.permissionId) );
}
}
I need to take the time (when I have some!) to look at this API in detail
but the latest rev's addition of sql string as param to query exec is a nice
addition. Good job!
On Fri, Nov 27, 2009 at 3:16 PM, Gavin King <gavin.king(a)gmail.com> wrote:
New rev, with ability to pass SQL as strings.
On Fri, Nov 27, 2009 at 4:47 PM, Gavin King <gavin.king(a)gmail.com> wrote:
> Latest, with a new approach to specifying types.
>
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
_______________________________________________
weld-dev mailing list
weld-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/weld-dev
6:51 a.m.
New subject: SQL framework proof of concept
On 28 Nov 2009, at 02:44, Arbi Sookazian wrote:
Following comments apply to seam-sql5 rev...
1) In Tuple class:
public <T> T get(int column, Type<T> type) {
return (T) get(column);
}
The type param is not referenced in the body of the method. Why?
Right, you can rewrite this as
return type.cast(get(column));
which does a runtime cast and doesn't generate compiler warnings.
Really, the reason for passing in the "type" parameter is because type inference
in Java isn't that great and the compiler get's it wrong sometimes (interestingly
JDT is better at this than javac, so when you recompile on the command line you get
additional problems, ouch!). Without that type parameter you will end up having to write
tuple.<Foo>get(1);
occasionally in client code, which is just ugly.
Passing in the type parameter is the best of two bad options IMO.
12:25 a.m.
New subject: SQL framework proof of concept
Looks like there are two tables (users and permissions) but testQuery2() and
testQuery3() both select only from one table - users. So now we need a
small db schema (i.e. data model). The typical customers and orders would
suffice to demo FK and potential outer joins. How can I define customerId
as a FK on orders table? Ideally, it would be nice to be able to do all
this via tooling (i.e. exec a wizard that would generate the code for us).
But we need to be able to do all this via the API for now...
On Fri, Nov 27, 2009 at 3:16 PM, Gavin King <gavin.king(a)gmail.com> wrote:
New rev, with ability to pass SQL as strings.
On Fri, Nov 27, 2009 at 4:47 PM, Gavin King <gavin.king(a)gmail.com> wrote:
> Latest, with a new approach to specifying types.
>
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
_______________________________________________
weld-dev mailing list
weld-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/weld-dev
1:10 a.m.
New subject: SQL framework proof of concept
Ok, my attempt at rev 6. introduced orders and customers tables. compiles
fine. throws exception here in createTable():
Create<orders> create4 = new Create<orders>(orders.table);
create4.add(orders.id, "varchar(40) not null primary key");
create4.add(orders.customersid, "varchar(40) not null foreign
key(customersid) references customers(id)");
session.prepare(create4).execute();
console output:
create table users (name varchar(40) not null, username varchar(10) not null
primary key, password varchar(10) not null, age smallint not null)
create table permissions (username varchar(10) not null, id varchar(20) not
null primary key)
create table customers (id varchar(40) not null primary key)
create table orders (id varchar(40) not null primary key, customersid
varchar(40) not null foreign key(customersid) references customers(id))
Exception in thread "main" java.lang.RuntimeException:
java.sql.SQLException: Unexpected token: FOREIGN in statement [create table
orders (id varchar(40) not null primary key, customersid varchar(40) not
null foreign]
at
org.seamframework.sql.ExecutableStatement.execute(ExecutableStatement.java:53)
at eg.Main.createTable(Main.java:87)
at eg.Main.main(Main.java:52)
Caused by: java.sql.SQLException: Unexpected token: FOREIGN in statement
[create table orders (id varchar(40) not null primary key, customersid
varchar(40) not null foreign]
at org.hsqldb.jdbc.Util.throwError(Unknown Source)
at org.hsqldb.jdbc.jdbcPreparedStatement.execute(Unknown Source)
at
org.seamframework.sql.ExecutableStatement.execute(ExecutableStatement.java:50)
... 2 more
We'll need to be able to handle this in a typesafe manner (I'm talking about
DDL stuff like defining not null and PK/FK, etc.) as well and also specify
rdbms vendor in a props file, etc. b/c I noticed that the syntax for sql
server and oracle are different for defining FK's:
http://www.1keydata.com/sql/sql-foreign-key.html
Also need to be able to handle identity columns for surrogate keys as well
as sequences.
thx.
On Fri, Nov 27, 2009 at 10:25 PM, Arbi Sookazian <asookazian(a)gmail.com>wrote:
Looks like there are two tables (users and permissions) but
testQuery2()
and testQuery3() both select only from one table - users. So now we need a
small db schema (i.e. data model). The typical customers and orders would
suffice to demo FK and potential outer joins. How can I define customerId
as a FK on orders table? Ideally, it would be nice to be able to do all
this via tooling (i.e. exec a wizard that would generate the code for us).
But we need to be able to do all this via the API for now...
On Fri, Nov 27, 2009 at 3:16 PM, Gavin King <gavin.king(a)gmail.com> wrote:
> New rev, with ability to pass SQL as strings.
>
> On Fri, Nov 27, 2009 at 4:47 PM, Gavin King <gavin.king(a)gmail.com> wrote:
> > Latest, with a new approach to specifying types.
> >
>
>
>
> --
> Gavin King
> gavin.king(a)gmail.com
> http://in.relation.to/Bloggers/Gavin
> http://hibernate.org
> http://seamframework.org
>
> _______________________________________________
> weld-dev mailing list
> weld-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/weld-dev
>
4:41 p.m.
New subject: SQL framework proof of concept
Would it be doable to use e.g. Javassist to transform a class into one
of those static typesafe "entity classes" at runtime?
On Sat, Nov 28, 2009 at 8:25 AM, Arbi Sookazian <asookazian(a)gmail.com> wrote:
Looks like there are two tables (users and permissions) but
testQuery2() and
testQuery3() both select only from one table - users. So now we need a
small db schema (i.e. data model). The typical customers and orders would
suffice to demo FK and potential outer joins. How can I define customerId
as a FK on orders table? Ideally, it would be nice to be able to do all
this via tooling (i.e. exec a wizard that would generate the code for us).
But we need to be able to do all this via the API for now...
On Fri, Nov 27, 2009 at 3:16 PM, Gavin King <gavin.king(a)gmail.com> wrote:
>
> New rev, with ability to pass SQL as strings.
>
> On Fri, Nov 27, 2009 at 4:47 PM, Gavin King <gavin.king(a)gmail.com> wrote:
> > Latest, with a new approach to specifying types.
> >
>
>
>
> --
> Gavin King
> gavin.king(a)gmail.com
> http://in.relation.to/Bloggers/Gavin
> http://hibernate.org
> http://seamframework.org
>
> _______________________________________________
> weld-dev mailing list
> weld-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/weld-dev
_______________________________________________
weld-dev mailing list
weld-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/weld-dev
--
---
Nik
4:50 p.m.
New subject: SQL framework proof of concept
Eh? By definition you can't do anything with the type system at *runtime*!
On Sat, Nov 28, 2009 at 5:41 PM, Nicklas Karlsson <nickarls(a)gmail.com> wrote:
Would it be doable to use e.g. Javassist to transform a class into
one
of those static typesafe "entity classes" at runtime?
On Sat, Nov 28, 2009 at 8:25 AM, Arbi Sookazian <asookazian(a)gmail.com> wrote:
> Looks like there are two tables (users and permissions) but testQuery2() and
> testQuery3() both select only from one table - users. So now we need a
> small db schema (i.e. data model). The typical customers and orders would
> suffice to demo FK and potential outer joins. How can I define customerId
> as a FK on orders table? Ideally, it would be nice to be able to do all
> this via tooling (i.e. exec a wizard that would generate the code for us).
> But we need to be able to do all this via the API for now...
>
> On Fri, Nov 27, 2009 at 3:16 PM, Gavin King <gavin.king(a)gmail.com> wrote:
>>
>> New rev, with ability to pass SQL as strings.
>>
>> On Fri, Nov 27, 2009 at 4:47 PM, Gavin King <gavin.king(a)gmail.com> wrote:
>> > Latest, with a new approach to specifying types.
>> >
>>
>>
>>
>> --
>> Gavin King
>> gavin.king(a)gmail.com
>> http://in.relation.to/Bloggers/Gavin
>> http://hibernate.org
>> http://seamframework.org
>>
>> _______________________________________________
>> weld-dev mailing list
>> weld-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/weld-dev
>
>
> _______________________________________________
> weld-dev mailing list
> weld-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/weld-dev
>
--
---
Nik
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
4:58 p.m.
New subject: SQL framework proof of concept
don't you hate the rules? it was an interesting question... break the
rules.
On Sat, Nov 28, 2009 at 2:50 PM, Gavin King <gavin.king(a)gmail.com> wrote:
Eh? By definition you can't do anything with the type system at
*runtime*!
On Sat, Nov 28, 2009 at 5:41 PM, Nicklas Karlsson <nickarls(a)gmail.com>
wrote:
> Would it be doable to use e.g. Javassist to transform a class into one
> of those static typesafe "entity classes" at runtime?
>
> On Sat, Nov 28, 2009 at 8:25 AM, Arbi Sookazian <asookazian(a)gmail.com>
wrote:
>> Looks like there are two tables (users and permissions) but testQuery2()
and
>> testQuery3() both select only from one table - users. So now we need a
>> small db schema (i.e. data model). The typical customers and orders
would
>> suffice to demo FK and potential outer joins. How can I define
customerId
>> as a FK on orders table? Ideally, it would be nice to be able to do all
>> this via tooling (i.e. exec a wizard that would generate the code for
us).
>> But we need to be able to do all this via the API for now...
>>
>> On Fri, Nov 27, 2009 at 3:16 PM, Gavin King <gavin.king(a)gmail.com>
wrote:
>>>
>>> New rev, with ability to pass SQL as strings.
>>>
>>> On Fri, Nov 27, 2009 at 4:47 PM, Gavin King <gavin.king(a)gmail.com>
wrote:
>>> > Latest, with a new approach to specifying types.
>>> >
>>>
>>>
>>>
>>> --
>>> Gavin King
>>> gavin.king(a)gmail.com
>>> http://in.relation.to/Bloggers/Gavin
>>> http://hibernate.org
>>> http://seamframework.org
>>>
>>> _______________________________________________
>>> weld-dev mailing list
>>> weld-dev(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/weld-dev
>>
>>
>> _______________________________________________
>> weld-dev mailing list
>> weld-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/weld-dev
>>
>
>
>
> --
> ---
> Nik
>
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
_______________________________________________
weld-dev mailing list
weld-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/weld-dev
5:08 p.m.
New subject: SQL framework proof of concept
OK, great. I've just decided to break the rules and say that up is
down. You should all be lying on the ceiling by now.
On Sat, Nov 28, 2009 at 5:58 PM, Arbi Sookazian <asookazian(a)gmail.com> wrote:
don't you hate the rules? it was an interesting question...
break the
rules.
On Sat, Nov 28, 2009 at 2:50 PM, Gavin King <gavin.king(a)gmail.com> wrote:
>
> Eh? By definition you can't do anything with the type system at *runtime*!
>
> On Sat, Nov 28, 2009 at 5:41 PM, Nicklas Karlsson <nickarls(a)gmail.com>
> wrote:
> > Would it be doable to use e.g. Javassist to transform a class into one
> > of those static typesafe "entity classes" at runtime?
> >
> > On Sat, Nov 28, 2009 at 8:25 AM, Arbi Sookazian <asookazian(a)gmail.com>
> > wrote:
> >> Looks like there are two tables (users and permissions) but
> >> testQuery2() and
> >> testQuery3() both select only from one table - users. So now we need a
> >> small db schema (i.e. data model). The typical customers and orders
> >> would
> >> suffice to demo FK and potential outer joins. How can I define
> >> customerId
> >> as a FK on orders table? Ideally, it would be nice to be able to do
> >> all
> >> this via tooling (i.e. exec a wizard that would generate the code for
> >> us).
> >> But we need to be able to do all this via the API for now...
> >>
> >> On Fri, Nov 27, 2009 at 3:16 PM, Gavin King <gavin.king(a)gmail.com>
> >> wrote:
> >>>
> >>> New rev, with ability to pass SQL as strings.
> >>>
> >>> On Fri, Nov 27, 2009 at 4:47 PM, Gavin King
<gavin.king(a)gmail.com>
> >>> wrote:
> >>> > Latest, with a new approach to specifying types.
> >>> >
> >>>
> >>>
> >>>
> >>> --
> >>> Gavin King
> >>> gavin.king(a)gmail.com
> >>> http://in.relation.to/Bloggers/Gavin
> >>> http://hibernate.org
> >>> http://seamframework.org
> >>>
> >>> _______________________________________________
> >>> weld-dev mailing list
> >>> weld-dev(a)lists.jboss.org
> >>> https://lists.jboss.org/mailman/listinfo/weld-dev
> >>
> >>
> >> _______________________________________________
> >> weld-dev mailing list
> >> weld-dev(a)lists.jboss.org
> >> https://lists.jboss.org/mailman/listinfo/weld-dev
> >>
> >
> >
> >
> > --
> > ---
> > Nik
> >
>
>
>
> --
> Gavin King
> gavin.king(a)gmail.com
> http://in.relation.to/Bloggers/Gavin
> http://hibernate.org
> http://seamframework.org
>
> _______________________________________________
> weld-dev mailing list
> weld-dev(a)lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/weld-dev
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
5:17 p.m.
New subject: SQL framework proof of concept
OH come on. Can't you be more creative than that, it's the weekend after
all? It's like being trapped in a jail cell, the JVM, and you'll never
break free until you break the rules somehow. We need some kind of
advantage over the other groups: http://deoxy.org/hs_cehn.htm
what kind of an elevation is OOP or AOP really when you consider invisible
dimensions? there's gotta be something else lurking invisibly somewhere...
On Sat, Nov 28, 2009 at 3:08 PM, Gavin King <gavin.king(a)gmail.com> wrote:
OK, great. I've just decided to break the rules and say that up
is
down. You should all be lying on the ceiling by now.
On Sat, Nov 28, 2009 at 5:58 PM, Arbi Sookazian <asookazian(a)gmail.com>
wrote:
> don't you hate the rules? it was an interesting question... break the
> rules.
>
> On Sat, Nov 28, 2009 at 2:50 PM, Gavin King <gavin.king(a)gmail.com>
wrote:
>>
>> Eh? By definition you can't do anything with the type system at
*runtime*!
>>
>> On Sat, Nov 28, 2009 at 5:41 PM, Nicklas Karlsson <nickarls(a)gmail.com>
>> wrote:
>> > Would it be doable to use e.g. Javassist to transform a class into one
>> > of those static typesafe "entity classes" at runtime?
>> >
>> > On Sat, Nov 28, 2009 at 8:25 AM, Arbi Sookazian <asookazian(a)gmail.com
>
>> > wrote:
>> >> Looks like there are two tables (users and permissions) but
>> >> testQuery2() and
>> >> testQuery3() both select only from one table - users. So now we need
a
>> >> small db schema (i.e. data model). The typical customers and orders
>> >> would
>> >> suffice to demo FK and potential outer joins. How can I define
>> >> customerId
>> >> as a FK on orders table? Ideally, it would be nice to be able to do
>> >> all
>> >> this via tooling (i.e. exec a wizard that would generate the code for
>> >> us).
>> >> But we need to be able to do all this via the API for now...
>> >>
>> >> On Fri, Nov 27, 2009 at 3:16 PM, Gavin King
<gavin.king(a)gmail.com>
>> >> wrote:
>> >>>
>> >>> New rev, with ability to pass SQL as strings.
>> >>>
>> >>> On Fri, Nov 27, 2009 at 4:47 PM, Gavin King
<gavin.king(a)gmail.com>
>> >>> wrote:
>> >>> > Latest, with a new approach to specifying types.
>> >>> >
>> >>>
>> >>>
>> >>>
>> >>> --
>> >>> Gavin King
>> >>> gavin.king(a)gmail.com
>> >>> http://in.relation.to/Bloggers/Gavin
>> >>> http://hibernate.org
>> >>> http://seamframework.org
>> >>>
>> >>> _______________________________________________
>> >>> weld-dev mailing list
>> >>> weld-dev(a)lists.jboss.org
>> >>> https://lists.jboss.org/mailman/listinfo/weld-dev
>> >>
>> >>
>> >> _______________________________________________
>> >> weld-dev mailing list
>> >> weld-dev(a)lists.jboss.org
>> >> https://lists.jboss.org/mailman/listinfo/weld-dev
>> >>
>> >
>> >
>> >
>> > --
>> > ---
>> > Nik
>> >
>>
>>
>>
>> --
>> Gavin King
>> gavin.king(a)gmail.com
>> http://in.relation.to/Bloggers/Gavin
>> http://hibernate.org
>> http://seamframework.org
>>
>> _______________________________________________
>> weld-dev mailing list
>> weld-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/weld-dev
>
>
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
5:21 p.m.
New subject: SQL framework proof of concept
and what about tx demarcation? let's say we have a use case that requires
two updates in a single tx and would like to use seam-sql. How can I
implement this using your new API? Is there a @Transactional or similar
(BMT?) I can use or must I use Seam's @Transactional (w/o the REQUIRES_NEW
or NOT_SUPPORTED)...
On Sat, Nov 28, 2009 at 3:17 PM, Arbi Sookazian <asookazian(a)gmail.com>wrote:
OH come on. Can't you be more creative than that, it's the
weekend after
all? It's like being trapped in a jail cell, the JVM, and you'll never
break free until you break the rules somehow. We need some kind of
advantage over the other groups: http://deoxy.org/hs_cehn.htm
what kind of an elevation is OOP or AOP really when you consider invisible
dimensions? there's gotta be something else lurking invisibly somewhere...
On Sat, Nov 28, 2009 at 3:08 PM, Gavin King <gavin.king(a)gmail.com> wrote:
> OK, great. I've just decided to break the rules and say that up is
> down. You should all be lying on the ceiling by now.
>
> On Sat, Nov 28, 2009 at 5:58 PM, Arbi Sookazian <asookazian(a)gmail.com>
> wrote:
> > don't you hate the rules? it was an interesting question... break the
> > rules.
> >
> > On Sat, Nov 28, 2009 at 2:50 PM, Gavin King <gavin.king(a)gmail.com>
> wrote:
> >>
> >> Eh? By definition you can't do anything with the type system at
> *runtime*!
> >>
> >> On Sat, Nov 28, 2009 at 5:41 PM, Nicklas Karlsson
<nickarls(a)gmail.com>
> >> wrote:
> >> > Would it be doable to use e.g. Javassist to transform a class into
> one
> >> > of those static typesafe "entity classes" at runtime?
> >> >
> >> > On Sat, Nov 28, 2009 at 8:25 AM, Arbi Sookazian <
> asookazian(a)gmail.com>
> >> > wrote:
> >> >> Looks like there are two tables (users and permissions) but
> >> >> testQuery2() and
> >> >> testQuery3() both select only from one table - users. So now we
> need a
> >> >> small db schema (i.e. data model). The typical customers and
orders
> >> >> would
> >> >> suffice to demo FK and potential outer joins. How can I define
> >> >> customerId
> >> >> as a FK on orders table? Ideally, it would be nice to be able to
do
> >> >> all
> >> >> this via tooling (i.e. exec a wizard that would generate the code
> for
> >> >> us).
> >> >> But we need to be able to do all this via the API for now...
> >> >>
> >> >> On Fri, Nov 27, 2009 at 3:16 PM, Gavin King
<gavin.king(a)gmail.com>
> >> >> wrote:
> >> >>>
> >> >>> New rev, with ability to pass SQL as strings.
> >> >>>
> >> >>> On Fri, Nov 27, 2009 at 4:47 PM, Gavin King
<gavin.king(a)gmail.com>
> >> >>> wrote:
> >> >>> > Latest, with a new approach to specifying types.
> >> >>> >
> >> >>>
> >> >>>
> >> >>>
> >> >>> --
> >> >>> Gavin King
> >> >>> gavin.king(a)gmail.com
> >> >>> http://in.relation.to/Bloggers/Gavin
> >> >>> http://hibernate.org
> >> >>> http://seamframework.org
> >> >>>
> >> >>> _______________________________________________
> >> >>> weld-dev mailing list
> >> >>> weld-dev(a)lists.jboss.org
> >> >>> https://lists.jboss.org/mailman/listinfo/weld-dev
> >> >>
> >> >>
> >> >> _______________________________________________
> >> >> weld-dev mailing list
> >> >> weld-dev(a)lists.jboss.org
> >> >> https://lists.jboss.org/mailman/listinfo/weld-dev
> >> >>
> >> >
> >> >
> >> >
> >> > --
> >> > ---
> >> > Nik
> >> >
> >>
> >>
> >>
> >> --
> >> Gavin King
> >> gavin.king(a)gmail.com
> >> http://in.relation.to/Bloggers/Gavin
> >> http://hibernate.org
> >> http://seamframework.org
> >>
> >> _______________________________________________
> >> weld-dev mailing list
> >> weld-dev(a)lists.jboss.org
> >> https://lists.jboss.org/mailman/listinfo/weld-dev
> >
> >
>
>
>
> --
> Gavin King
> gavin.king(a)gmail.com
> http://in.relation.to/Bloggers/Gavin
> http://hibernate.org
> http://seamframework.org
>
5:24 p.m.
New subject: SQL framework proof of concept
yes, and this is off topic but it is Saturday, so plz excuse me....
The finding of DMT in normal human body fluids opens up interesting legal
questions. Since DMT is illegal, as is "any material, compound, mixture or
preparation" containing DMT, it would seem we are all guilty of possession
of a controlled substance. *Jonathan Ott <http://deoxy.org/def/index.htm>
So the government is saying, indirectly, we're illegal... you're born
illegal. nice, very nice and weak. break the rules.
*
On Sat, Nov 28, 2009 at 3:21 PM, Arbi Sookazian <asookazian(a)gmail.com>wrote:
and what about tx demarcation? let's say we have a use case that
requires
two updates in a single tx and would like to use seam-sql. How can I
implement this using your new API? Is there a @Transactional or similar
(BMT?) I can use or must I use Seam's @Transactional (w/o the REQUIRES_NEW
or NOT_SUPPORTED)...
On Sat, Nov 28, 2009 at 3:17 PM, Arbi Sookazian <asookazian(a)gmail.com>wrote:
> OH come on. Can't you be more creative than that, it's the weekend after
> all? It's like being trapped in a jail cell, the JVM, and you'll never
> break free until you break the rules somehow. We need some kind of
> advantage over the other groups: http://deoxy.org/hs_cehn.htm
>
> what kind of an elevation is OOP or AOP really when you consider invisible
> dimensions? there's gotta be something else lurking invisibly somewhere...
>
>
> On Sat, Nov 28, 2009 at 3:08 PM, Gavin King <gavin.king(a)gmail.com> wrote:
>
>> OK, great. I've just decided to break the rules and say that up is
>> down. You should all be lying on the ceiling by now.
>>
>> On Sat, Nov 28, 2009 at 5:58 PM, Arbi Sookazian <asookazian(a)gmail.com>
>> wrote:
>> > don't you hate the rules? it was an interesting question... break the
>> > rules.
>> >
>> > On Sat, Nov 28, 2009 at 2:50 PM, Gavin King <gavin.king(a)gmail.com>
>> wrote:
>> >>
>> >> Eh? By definition you can't do anything with the type system at
>> *runtime*!
>> >>
>> >> On Sat, Nov 28, 2009 at 5:41 PM, Nicklas Karlsson
<nickarls(a)gmail.com
>> >
>> >> wrote:
>> >> > Would it be doable to use e.g. Javassist to transform a class into
>> one
>> >> > of those static typesafe "entity classes" at runtime?
>> >> >
>> >> > On Sat, Nov 28, 2009 at 8:25 AM, Arbi Sookazian <
>> asookazian(a)gmail.com>
>> >> > wrote:
>> >> >> Looks like there are two tables (users and permissions) but
>> >> >> testQuery2() and
>> >> >> testQuery3() both select only from one table - users. So now
we
>> need a
>> >> >> small db schema (i.e. data model). The typical customers and
>> orders
>> >> >> would
>> >> >> suffice to demo FK and potential outer joins. How can I
define
>> >> >> customerId
>> >> >> as a FK on orders table? Ideally, it would be nice to be able
to
>> do
>> >> >> all
>> >> >> this via tooling (i.e. exec a wizard that would generate the
code
>> for
>> >> >> us).
>> >> >> But we need to be able to do all this via the API for now...
>> >> >>
>> >> >> On Fri, Nov 27, 2009 at 3:16 PM, Gavin King
<gavin.king(a)gmail.com>
>> >> >> wrote:
>> >> >>>
>> >> >>> New rev, with ability to pass SQL as strings.
>> >> >>>
>> >> >>> On Fri, Nov 27, 2009 at 4:47 PM, Gavin King
<gavin.king(a)gmail.com
>> >
>> >> >>> wrote:
>> >> >>> > Latest, with a new approach to specifying types.
>> >> >>> >
>> >> >>>
>> >> >>>
>> >> >>>
>> >> >>> --
>> >> >>> Gavin King
>> >> >>> gavin.king(a)gmail.com
>> >> >>> http://in.relation.to/Bloggers/Gavin
>> >> >>> http://hibernate.org
>> >> >>> http://seamframework.org
>> >> >>>
>> >> >>> _______________________________________________
>> >> >>> weld-dev mailing list
>> >> >>> weld-dev(a)lists.jboss.org
>> >> >>> https://lists.jboss.org/mailman/listinfo/weld-dev
>> >> >>
>> >> >>
>> >> >> _______________________________________________
>> >> >> weld-dev mailing list
>> >> >> weld-dev(a)lists.jboss.org
>> >> >> https://lists.jboss.org/mailman/listinfo/weld-dev
>> >> >>
>> >> >
>> >> >
>> >> >
>> >> > --
>> >> > ---
>> >> > Nik
>> >> >
>> >>
>> >>
>> >>
>> >> --
>> >> Gavin King
>> >> gavin.king(a)gmail.com
>> >> http://in.relation.to/Bloggers/Gavin
>> >> http://hibernate.org
>> >> http://seamframework.org
>> >>
>> >> _______________________________________________
>> >> weld-dev mailing list
>> >> weld-dev(a)lists.jboss.org
>> >> https://lists.jboss.org/mailman/listinfo/weld-dev
>> >
>> >
>>
>>
>>
>> --
>> Gavin King
>> gavin.king(a)gmail.com
>> http://in.relation.to/Bloggers/Gavin
>> http://hibernate.org
>> http://seamframework.org
>>
>
>
5:40 p.m.
New subject: SQL framework proof of concept
btw, I literally tried for at least an hour this morning (and finally gave
up) to create a JAR for the seam-sql project using Eclipse's export project
functionality. It would not add the lib directory with the weld and hsqldb
jar's to my JAR. Very embarassing and a PITA. In the wizard, it shows that
it will add the jar's but it doesn't.
I'd like to use this:
void start(@Observes ContainerInitialized ci, Session session) {
and this from the StartMain API doc:
java -jar MyApp.jar org.jboss.weld.environment.se.StartMain arguments
rather than public static void main()...
On Sat, Nov 28, 2009 at 3:24 PM, Arbi Sookazian <asookazian(a)gmail.com>wrote:
yes, and this is off topic but it is Saturday, so plz excuse me....
The finding of DMT in normal human body fluids opens up interesting legal
questions. Since DMT is illegal, as is "any material, compound, mixture or
preparation" containing DMT, it would seem we are all guilty of possession
of a controlled substance. *Jonathan Ott <http://deoxy.org/def/index.htm>
So the government is saying, indirectly, we're illegal... you're born
illegal. nice, very nice and weak. break the rules.
*
On Sat, Nov 28, 2009 at 3:21 PM, Arbi Sookazian <asookazian(a)gmail.com>wrote:
> and what about tx demarcation? let's say we have a use case that requires
> two updates in a single tx and would like to use seam-sql. How can I
> implement this using your new API? Is there a @Transactional or similar
> (BMT?) I can use or must I use Seam's @Transactional (w/o the REQUIRES_NEW
> or NOT_SUPPORTED)...
>
>
> On Sat, Nov 28, 2009 at 3:17 PM, Arbi Sookazian <asookazian(a)gmail.com>wrote:
>
>> OH come on. Can't you be more creative than that, it's the weekend
after
>> all? It's like being trapped in a jail cell, the JVM, and you'll never
>> break free until you break the rules somehow. We need some kind of
>> advantage over the other groups: http://deoxy.org/hs_cehn.htm
>>
>> what kind of an elevation is OOP or AOP really when you consider
>> invisible dimensions? there's gotta be something else lurking invisibly
>> somewhere...
>>
>>
>> On Sat, Nov 28, 2009 at 3:08 PM, Gavin King <gavin.king(a)gmail.com>wrote:
>>
>>> OK, great. I've just decided to break the rules and say that up is
>>> down. You should all be lying on the ceiling by now.
>>>
>>> On Sat, Nov 28, 2009 at 5:58 PM, Arbi Sookazian <asookazian(a)gmail.com>
>>> wrote:
>>> > don't you hate the rules? it was an interesting question... break
the
>>> > rules.
>>> >
>>> > On Sat, Nov 28, 2009 at 2:50 PM, Gavin King
<gavin.king(a)gmail.com>
>>> wrote:
>>> >>
>>> >> Eh? By definition you can't do anything with the type system at
>>> *runtime*!
>>> >>
>>> >> On Sat, Nov 28, 2009 at 5:41 PM, Nicklas Karlsson <
>>> nickarls(a)gmail.com>
>>> >> wrote:
>>> >> > Would it be doable to use e.g. Javassist to transform a class
into
>>> one
>>> >> > of those static typesafe "entity classes" at
runtime?
>>> >> >
>>> >> > On Sat, Nov 28, 2009 at 8:25 AM, Arbi Sookazian <
>>> asookazian(a)gmail.com>
>>> >> > wrote:
>>> >> >> Looks like there are two tables (users and permissions)
but
>>> >> >> testQuery2() and
>>> >> >> testQuery3() both select only from one table - users. So
now we
>>> need a
>>> >> >> small db schema (i.e. data model). The typical customers
and
>>> orders
>>> >> >> would
>>> >> >> suffice to demo FK and potential outer joins. How can I
define
>>> >> >> customerId
>>> >> >> as a FK on orders table? Ideally, it would be nice to be
able to
>>> do
>>> >> >> all
>>> >> >> this via tooling (i.e. exec a wizard that would generate
the code
>>> for
>>> >> >> us).
>>> >> >> But we need to be able to do all this via the API for
now...
>>> >> >>
>>> >> >> On Fri, Nov 27, 2009 at 3:16 PM, Gavin King
<gavin.king(a)gmail.com
>>> >
>>> >> >> wrote:
>>> >> >>>
>>> >> >>> New rev, with ability to pass SQL as strings.
>>> >> >>>
>>> >> >>> On Fri, Nov 27, 2009 at 4:47 PM, Gavin King <
>>> gavin.king(a)gmail.com>
>>> >> >>> wrote:
>>> >> >>> > Latest, with a new approach to specifying types.
>>> >> >>> >
>>> >> >>>
>>> >> >>>
>>> >> >>>
>>> >> >>> --
>>> >> >>> Gavin King
>>> >> >>> gavin.king(a)gmail.com
>>> >> >>> http://in.relation.to/Bloggers/Gavin
>>> >> >>> http://hibernate.org
>>> >> >>> http://seamframework.org
>>> >> >>>
>>> >> >>> _______________________________________________
>>> >> >>> weld-dev mailing list
>>> >> >>> weld-dev(a)lists.jboss.org
>>> >> >>> https://lists.jboss.org/mailman/listinfo/weld-dev
>>> >> >>
>>> >> >>
>>> >> >> _______________________________________________
>>> >> >> weld-dev mailing list
>>> >> >> weld-dev(a)lists.jboss.org
>>> >> >> https://lists.jboss.org/mailman/listinfo/weld-dev
>>> >> >>
>>> >> >
>>> >> >
>>> >> >
>>> >> > --
>>> >> > ---
>>> >> > Nik
>>> >> >
>>> >>
>>> >>
>>> >>
>>> >> --
>>> >> Gavin King
>>> >> gavin.king(a)gmail.com
>>> >> http://in.relation.to/Bloggers/Gavin
>>> >> http://hibernate.org
>>> >> http://seamframework.org
>>> >>
>>> >> _______________________________________________
>>> >> weld-dev mailing list
>>> >> weld-dev(a)lists.jboss.org
>>> >> https://lists.jboss.org/mailman/listinfo/weld-dev
>>> >
>>> >
>>>
>>>
>>>
>>> --
>>> Gavin King
>>> gavin.king(a)gmail.com
>>> http://in.relation.to/Bloggers/Gavin
>>> http://hibernate.org
>>> http://seamframework.org
>>>
>>
>>
>
4:59 p.m.
New subject: SQL framework proof of concept
OK. I was just brainstorming and thought something in the lines of
Starting with a class Foo
Use javassist magic to create a new class names e.g. FooEntity
Sneak in into the classloader somehow and use it dynamically (like
Class.forName) hiding details from the user
But you know more of this stuff than I...
On Sun, Nov 29, 2009 at 12:50 AM, Gavin King <gavin.king(a)gmail.com> wrote:
Eh? By definition you can't do anything with the type system at
*runtime*!
On Sat, Nov 28, 2009 at 5:41 PM, Nicklas Karlsson <nickarls(a)gmail.com> wrote:
> Would it be doable to use e.g. Javassist to transform a class into one
> of those static typesafe "entity classes" at runtime?
>
> On Sat, Nov 28, 2009 at 8:25 AM, Arbi Sookazian <asookazian(a)gmail.com> wrote:
>> Looks like there are two tables (users and permissions) but testQuery2() and
>> testQuery3() both select only from one table - users. So now we need a
>> small db schema (i.e. data model). The typical customers and orders would
>> suffice to demo FK and potential outer joins. How can I define customerId
>> as a FK on orders table? Ideally, it would be nice to be able to do all
>> this via tooling (i.e. exec a wizard that would generate the code for us).
>> But we need to be able to do all this via the API for now...
>>
>> On Fri, Nov 27, 2009 at 3:16 PM, Gavin King <gavin.king(a)gmail.com> wrote:
>>>
>>> New rev, with ability to pass SQL as strings.
>>>
>>> On Fri, Nov 27, 2009 at 4:47 PM, Gavin King <gavin.king(a)gmail.com>
wrote:
>>> > Latest, with a new approach to specifying types.
>>> >
>>>
>>>
>>>
>>> --
>>> Gavin King
>>> gavin.king(a)gmail.com
>>> http://in.relation.to/Bloggers/Gavin
>>> http://hibernate.org
>>> http://seamframework.org
>>>
>>> _______________________________________________
>>> weld-dev mailing list
>>> weld-dev(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/weld-dev
>>
>>
>> _______________________________________________
>> weld-dev mailing list
>> weld-dev(a)lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/weld-dev
>>
>
>
>
> --
> ---
> Nik
>
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
--
---
Nik
5:06 p.m.
New subject: SQL framework proof of concept
Nick, javassist works at runtime, the generated classes are not
available to the compiler. So all code that uses FooEntity is
uncompilable. Javaassist is useful for creating new implementations of
types that already exist, not for creating new types.
What you're describing can be achieved using a Java6 processor (a
compiler plugin), which is what we are doing in Hibernate for
generating the X_ classes for entity classes X.
But in this case, we don't have any entities. These classes exist to
represent database tables. I don't love the idea of going off to the
database and reading the db metadata every time I compile my code ;-)
On Sat, Nov 28, 2009 at 5:59 PM, Nicklas Karlsson <nickarls(a)gmail.com> wrote:
OK. I was just brainstorming and thought something in the lines of
Starting with a class Foo
Use javassist magic to create a new class names e.g. FooEntity
Sneak in into the classloader somehow and use it dynamically (like
Class.forName) hiding details from the user
But you know more of this stuff than I...
On Sun, Nov 29, 2009 at 12:50 AM, Gavin King <gavin.king(a)gmail.com> wrote:
> Eh? By definition you can't do anything with the type system at *runtime*!
>
> On Sat, Nov 28, 2009 at 5:41 PM, Nicklas Karlsson <nickarls(a)gmail.com> wrote:
>> Would it be doable to use e.g. Javassist to transform a class into one
>> of those static typesafe "entity classes" at runtime?
>>
>> On Sat, Nov 28, 2009 at 8:25 AM, Arbi Sookazian <asookazian(a)gmail.com>
wrote:
>>> Looks like there are two tables (users and permissions) but testQuery2() and
>>> testQuery3() both select only from one table - users. So now we need a
>>> small db schema (i.e. data model). The typical customers and orders would
>>> suffice to demo FK and potential outer joins. How can I define customerId
>>> as a FK on orders table? Ideally, it would be nice to be able to do all
>>> this via tooling (i.e. exec a wizard that would generate the code for us).
>>> But we need to be able to do all this via the API for now...
>>>
>>> On Fri, Nov 27, 2009 at 3:16 PM, Gavin King <gavin.king(a)gmail.com>
wrote:
>>>>
>>>> New rev, with ability to pass SQL as strings.
>>>>
>>>> On Fri, Nov 27, 2009 at 4:47 PM, Gavin King <gavin.king(a)gmail.com>
wrote:
>>>> > Latest, with a new approach to specifying types.
>>>> >
>>>>
>>>>
>>>>
>>>> --
>>>> Gavin King
>>>> gavin.king(a)gmail.com
>>>> http://in.relation.to/Bloggers/Gavin
>>>> http://hibernate.org
>>>> http://seamframework.org
>>>>
>>>> _______________________________________________
>>>> weld-dev mailing list
>>>> weld-dev(a)lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/weld-dev
>>>
>>>
>>> _______________________________________________
>>> weld-dev mailing list
>>> weld-dev(a)lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/weld-dev
>>>
>>
>>
>>
>> --
>> ---
>> Nik
>>
>
>
>
> --
> Gavin King
> gavin.king(a)gmail.com
> http://in.relation.to/Bloggers/Gavin
> http://hibernate.org
> http://seamframework.org
>
--
---
Nik
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
2:09 a.m.
New subject: SQL framework proof of concept
Yeah, I noticed I stopped thinking before the line of code where I
actually used the generated class, well it was 1am ;-)
On Sun, Nov 29, 2009 at 1:06 AM, Gavin King <gavin.king(a)gmail.com> wrote:
Nick, javassist works at runtime, the generated classes are not
available to the compiler. So all code that uses FooEntity is
uncompilable. Javaassist is useful for creating new implementations of
types that already exist, not for creating new types.
What you're describing can be achieved using a Java6 processor (a
compiler plugin), which is what we are doing in Hibernate for
generating the X_ classes for entity classes X.
But in this case, we don't have any entities. These classes exist to
represent database tables. I don't love the idea of going off to the
database and reading the db metadata every time I compile my code ;-)
On Sat, Nov 28, 2009 at 5:59 PM, Nicklas Karlsson <nickarls(a)gmail.com> wrote:
> OK. I was just brainstorming and thought something in the lines of
>
> Starting with a class Foo
> Use javassist magic to create a new class names e.g. FooEntity
> Sneak in into the classloader somehow and use it dynamically (like
> Class.forName) hiding details from the user
>
> But you know more of this stuff than I...
>
>
>
> On Sun, Nov 29, 2009 at 12:50 AM, Gavin King <gavin.king(a)gmail.com> wrote:
>> Eh? By definition you can't do anything with the type system at *runtime*!
>>
>> On Sat, Nov 28, 2009 at 5:41 PM, Nicklas Karlsson <nickarls(a)gmail.com>
wrote:
>>> Would it be doable to use e.g. Javassist to transform a class into one
>>> of those static typesafe "entity classes" at runtime?
>>>
>>> On Sat, Nov 28, 2009 at 8:25 AM, Arbi Sookazian <asookazian(a)gmail.com>
wrote:
>>>> Looks like there are two tables (users and permissions) but testQuery2()
and
>>>> testQuery3() both select only from one table - users. So now we need a
>>>> small db schema (i.e. data model). The typical customers and orders
would
>>>> suffice to demo FK and potential outer joins. How can I define
customerId
>>>> as a FK on orders table? Ideally, it would be nice to be able to do all
>>>> this via tooling (i.e. exec a wizard that would generate the code for
us).
>>>> But we need to be able to do all this via the API for now...
>>>>
>>>> On Fri, Nov 27, 2009 at 3:16 PM, Gavin King <gavin.king(a)gmail.com>
wrote:
>>>>>
>>>>> New rev, with ability to pass SQL as strings.
>>>>>
>>>>> On Fri, Nov 27, 2009 at 4:47 PM, Gavin King
<gavin.king(a)gmail.com> wrote:
>>>>> > Latest, with a new approach to specifying types.
>>>>> >
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Gavin King
>>>>> gavin.king(a)gmail.com
>>>>> http://in.relation.to/Bloggers/Gavin
>>>>> http://hibernate.org
>>>>> http://seamframework.org
>>>>>
>>>>> _______________________________________________
>>>>> weld-dev mailing list
>>>>> weld-dev(a)lists.jboss.org
>>>>> https://lists.jboss.org/mailman/listinfo/weld-dev
>>>>
>>>>
>>>> _______________________________________________
>>>> weld-dev mailing list
>>>> weld-dev(a)lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/weld-dev
>>>>
>>>
>>>
>>>
>>> --
>>> ---
>>> Nik
>>>
>>
>>
>>
>> --
>> Gavin King
>> gavin.king(a)gmail.com
>> http://in.relation.to/Bloggers/Gavin
>> http://hibernate.org
>> http://seamframework.org
>>
>
>
>
> --
> ---
> Nik
>
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
--
---
Nik
6:05 p.m.
New subject: SQL framework proof of concept
Not too bad, but that makes sense if I have a tool that shows me the native
sql generated by the query. or better yet, I write a native sql statement
(say a query) and it reverse engineers it into the typesafe syntax you have
in your API.
honestly, this:
From<users> u = query.from(users.table);
Join<permissions> p = fromUser.join(permissions.
table)
p.on( eq( u.get(users.id), p.get(permissions.userId) ) );
seems a little busy and difficult to "swallow"...
On Thu, Nov 26, 2009 at 12:36 PM, Gavin King <gavin.king(a)gmail.com> wrote:
On Thu, Nov 26, 2009 at 3:16 PM, Marcus Smedman
<marcus(a)smedman.org>
wrote:
> Regarding the seam-sql I added a Drop (to speed up my small tests), Alter
> (to add foreign keys) and a Join (a bit messy when/if using aliases) to
try
> to understand how Gavin had did the other stuff. There’s a lot of stuff I
> don’t fully understand… (is a join an Expression for example?).
Join is probably a subclass of From:
From<users> u = query.from(users.table);
Join<permissions> p = fromUser.join(permissions.table)
p.on( eq( u.get(users.id), p.get(permissions.userId) ) );
Does that make sense?
--
Gavin King
gavin.king(a)gmail.com
http://in.relation.to/Bloggers/Gavin
http://hibernate.org
http://seamframework.org
6:13 p.m.
New subject: SQL framework proof of concept
query.where( and( eq( fromUsers.get(users.username), usernameParameter ),
eq( fromUsers.get(users.password),
passwordParameter ) ) );
what about left/right outer joins?
On Thu, Nov 26, 2009 at 4:05 PM, Arbi Sookazian <asookazian(a)gmail.com>wrote:
Not too bad, but that makes sense if I have a tool that shows me the
native
sql generated by the query. or better yet, I write a native sql statement
(say a query) and it reverse engineers it into the typesafe syntax you have
in your API.
honestly, this:
From<users> u = query.from(users.table);
Join<permissions> p = fromUser.join(permissions.
>
> table)
> p.on( eq( u.get(users.id), p.get(permissions.userId) ) );
seems a little busy and difficult to "swallow"...
On Thu, Nov 26, 2009 at 12:36 PM, Gavin King <gavin.king(a)gmail.com> wrote:
> On Thu, Nov 26, 2009 at 3:16 PM, Marcus Smedman <marcus(a)smedman.org>
> wrote:
>
> > Regarding the seam-sql I added a Drop (to speed up my small tests),
> Alter
> > (to add foreign keys) and a Join (a bit messy when/if using aliases) to
> try
> > to understand how Gavin had did the other stuff. There’s a lot of stuff
> I
> > don’t fully understand… (is a join an Expression for example?).
>
> Join is probably a subclass of From:
>
> From<users> u = query.from(users.table);
> Join<permissions> p = fromUser.join(permissions.table)
> p.on( eq( u.get(users.id), p.get(permissions.userId) ) );
>
> Does that make sense?
>
>
>
> --
> Gavin King
> gavin.king(a)gmail.com
> http://in.relation.to/Bloggers/Gavin
> http://hibernate.org
> http://seamframework.org
>
5475
days inactive
5477
days old
23 comments
5 participants
participants (5)
-
Arbi Sookazian -
Gavin King -
Marcus Smedman -
Nicklas Karlsson -
Pete Muir