There seems to be a build issue after the reorg into common.

I can find e.g. net.liftweb.common.Box.class in both

lift-common-1.1-SNAPSHOT
lift-webkit-1.1-SNAPSHOT

This causes confusion in Eclipse when you try to find the source of a class.

Opening an issue at Indrajit's request to update JPADemo/spa/pom.xml to use a newer version of hibernate-entitymanager 3.4.0.GA and resolve slf4j-1.4.2 impl dependencies.

The hibernate-entitymanager 3.4.0.GA version update causes an update to hibernate-core to 3.3.0.SP1 which fixes auto-increment column problem with MySQL and H2 databases, which was present in the older hibernate 3.2.4 (which dated back to before they called it hibernate-core) upon which hibernate-entitymanager 3.3.1 was depending.

Below is the configuration that worked for me.

<dependency>
 <groupId>org.hibernate</groupId>
 <artifactId>hibernate-entitymanager</artifactId>
  <version>3.4.0.GA</version>
  <exclusions>
   <exclusion>
    <groupId>javax.transaction</groupId>
    <artifactId>jta</artifactId>
   </exclusion>
  </exclusions>
</dependency>
<dependency>
 <groupId>org.slf4j</groupId>
 <artifactId>slf4j-simple</artifactId>
 <version>1.4.2</version>
</dependency>

Had to manually install slf4j-simple/1.4.2 into to my local maven repo to avoid runtime exception for missing org.slf4j.impl.StaticLoggerBinder. By putting the slf4j-simple dependency in the spa/pom.xml this allows users to swap out different impl providers (for example slf4j-log4j12) as they see fit, in fact it might not hurt to have an example of switching this over to slf4j-log4j12 or some other provider, perhaps even commented out just below the other one right in the pom.xml such as:

<!--
 <dependency>
  <groupId>org.slf4j</groupId>
  <artifactId>slf4j-log4j12</artifactId>
  <version>1.4.2</version>
 </dependency>
-->

Thanks! Troy

Here is a post in the groups regarding the issue.

http://groups.google.com/group/liftweb/browse_thread/thread/fc7244ce54f1768e

I need to parse a path in the following format.
http://domain.com/api/package/my.company.package%3Akey%3Dvalue

The following should extract "my.company.package:key=value" into my packageName var

case Req(List("api","package", packageName), _, GetRequest)

It doesn't; only "my" is extracted.

-Andy

It is spelled reset.pasword.confirmarion

Let's say you have a bit of a page like so:

<lift:SomeSnippet.section>
<foo:name/> <lift:SomeSnippet.section>

Now, let's assume that computing whatever is bound to <foo:name/>
takes a long time (maybe it takes a network call, or a long database
query, or whatever). You don't want to hold up your entire page view
on that so.....

<lift:Util.lazyLoad>
<lift:SomeSnippet.section> <foo:name/> <lift:SomeSnippet.section> </lift:Util.lazyLoad>

Then in your Util snippet:

def lazyLoad(xhtml: NodeSeq) = {
val id = "lazy"+System.currentTimeMillis() val (name, exp) = ajaxInvoke(() => { SetHtml(id, xhtml) })

{Script(OnLoad(exp.cmd))} }

Feel free to critique if there is a better way of doing this. It's
working pretty good for me so far though.

Update Lift and JPA page on Wiki

Lighthouse #23: http://liftweb.lighthouseapp.com/projects/26102/tickets/23

Support for changing the context root for requests so the root can be blank even if the actual context path is not or that a context path can be added (and all requests will be rewritten removing that context path from the head of the path) or subtracted.

MappedField's apply methods take view bounds. This causes a nasty interaction with Scala's mechanism for converting to Unit. So:

val u: Unit = myModel.myForiegnKeyField(1L) // causes problems

Fix: concrete apply method without view bound conversion.

Fix varchar and text types in SqlServerDriver

Lighthouse #22: http://liftweb.lighthouseapp.com/projects/26102/tickets/22

Add Twitter IDs to the team page.

Lighthouse #3: http://liftweb.lighthouseapp.com/projects/26102/tickets/3

There should be support for image maps in SHtml. See http://groups.google.com/group/liftweb/browse_frm/thread/84fee9386108a922?hl=en#

SQL Server Alter Table Fail

Lighthouse #25: http://liftweb.lighthouseapp.com/projects/26102/tickets/25

http://groups.google.com/group/liftweb/browse_thread/thread/39d38f2124b4b2c7

Problems with MappedStringIndex (Derby / H2)

Lighthouse #26: http://liftweb.lighthouseapp.com/projects/26102/tickets/26

The Liftweb installer for Mac OS X generates a .bash_profile file whether .profile exists already or not. As a result of this, the shell will ignore the preexisting environment variables defined by .profile. The installer should detect the presence of .profile, and perform an append operation instead.

the default autocomplete options i.e.

val autocompleteOptions = JsRaw(...

need to be more easily set. I had to override and copy-paste the render function to change these. Making them settable without extending the class would be ideal. making the autocompleteOptions val an overridable outer function would be fine though.

It should be reconfigured to use Joda Time

It would be nice if LIftRules has support for formatting/parsing of numbers in the same way as for dates. Both integers and decimals should be supported

When this is complete, Mapper should use this machinery in MappedInt (for thousand separators) and MappedDouble/Decimal

Problem:
On heavily JavaScript based forms (sections appearing/enabling depending on user input) it is good practice to outsource form creation code to widget/component classes (homegrown). It improves readability of Snippet code and reusability of widgets. For event registration and other client side code, each component class would generate a head/lift:tail and a script tag. While head/tail tags are merged by the Lift framework, the resulting HTML code still remains cluttered with script tags (example below).

Suggested solution:
Introduce a script merge that similarly to head/tail merge merges all script tags in head into one and all script tags in tail into a second one.

ECMA-262 (http://www.ecma-international.org/publications/files/ECMA-ST/ECMA-262.pdf) section
11.1.5 describes the object initializer syntax as being either empty braces or braces containing PropertyNameAndValueList. PropertyNameAndValueList is comma-separated sequence of PropertyName : AssignmentExpression. PropertyName is one of Identifier, StringLiteral, or NumericLiteral.

However, JE.JsObj.apply and JsObj both require the key to be a String, so what I would think the natural translation from the JavaScript:

{
    headers: {
        1: { sortable: false }
    }
}

To Lift:

JsObj("headers" -> JsObj(1 -> JsObj("sortable" -> false)))

Does not work:

<console>:7: error: type mismatch;
 found   : (Int, net.liftweb.http.js.JsObj)
 required: (String, net.liftweb.http.js.JsExp)
       JsObj("headers" -> JsObj(1 -> JsObj("sortable" -> false)))
                                  ^

as 1 is not a String. This is trivially worked around by either making ones own JsExp or by converting the Int to a String:

JsObj("headers" -> JsObj("1" -> JsObj("sortable" -> false)))

Apparently because JavaScript is entirely willing to treat strings as numbers and vice versa in property names, however it seems like it would be more natural.

This is low priority to me.

Related discussion here: http://groups.google.com/group/liftweb/browse_thread/thread/7de2d785a38bf27f

Br's,
Marius

A simple plain text mail should not be multipart. (Thread "URL in PlainMailBodyType to BlackBerry")

Hi, i newbie for lift and scala.. im using eclipse. i create project by maven and i try to run. thats all bu i get :

INFO - CREATE TABLE users (id BIGINT NOT NULL GENERATED ALWAYS AS IDENTITY , firstname VARCHAR(32) , lastname VARCHAR(32) , email VARCHAR(48) , locale VARCHAR(16) , timezone VARCHAR(32) , password_pw VARCHAR(48) , password_slt VARCHAR(20) , textarea VARCHAR(2048) , superuser SMALLINT , validated SMALLINT , uniqueıd VARCHAR(32))
INFO - ALTER TABLE users ADD CONSTRAINT users_PK PRIMARY KEY(id)
INFO - ALTER TABLE users ADD COLUMN id BIGINT NOT NULL GENERATED ALWAYS AS IDENTITY
ERROR - Failed to Boot
java.sql.SQLSyntaxErrorException: ALTER TABLE statement cannot add an IDENTITY column to a table.
at org.apache.derby.impl.jdbc.SQLExceptionFactory40.getSQLException(Unknown Source)

at org.apache.derby.impl.jdbc.Util.generateCsSQLException(Unknown Source)
at org.apache.derby.impl.jdbc.TransactionResourceImpl.wrapInSQLException(Unknown Source)
at org.apache.derby.impl.jdbc.TransactionResourceImpl.handleException(Unknown Source)
at org.apache.derby.impl.jdbc.EmbedConnection.handleException(Unknown Source)
at org.apache.derby.impl.jdbc.ConnectionChild.handleException(Unknown Source)
at org.apache.derby.impl.jdbc.EmbedStatement.execute(Unknown Source)

Add core nl_NL language file. Will also add as default nl language file since nl_BE isn't available.

Add Ext Core support to Lift

As discussed on the mailing list, it would be advantageous to:

• Add a type parameter to the base LocParam trait to allow for safety in the association between typeful LocParam instances such as Title and the type parameter of the Loc.
• Remove the overloading of "Param" that currently refers both to LocParam and the value associated with a Loc (potentially provided by a default or rewrite rule)
• Add data-aware instances of If, Unless, and Template LocParams
• Make LocParam a sealed trait with a well-defined extension point to help the compiler catch possible incomplete match statements.

This would allow client-side processing or updates after the form is submitted.

Tim ran into this crappy behavior when calling submit() function on a form the onsubmit event doesn't actually get called. Apparently this is a known browser behavior.

I propose to have alternative helpers in SHtml calls where we feed say the form ID and the actual Ajax call happens immediately outside of onsubmit context. For instance

1. Assume <form id="login"> ... </form>

2. Somewhere in a bind we have
   <button onclick={submitJsonForm(jsonHandler, "login")}>Press me </button>

def submitJsonForm(jsonHandler: JsonHandler, formId: String): JsCmd = {
jsonHandler.call("processForm", FormToJSON(formId)) }

a similar concept for ajaxForms ...

haven't compiled it but you get the idea ...

Br's,
Marius

Looking at: http://github.com/dpp/liftweb/blob/master/lift-base/lift-webkit/src/main/scala/net/liftweb/builtin/snippet/Msgs.scala

The documentation suggests that certain elements (like error_class) can be overridden. But I did it using:

<lift:Msgs>
    <lift:error_class>errorBox</lift:error_class>
</lift:Msgs>

I'm guessing that the <lift:snippet type="error_report"> is dated?

Please make the Scala identifier to database identifier translation convention dependent on whether the database supports case sensitive or case insensitive identifiers.

For example, Hibernate offers an Enhanced Naming Convention that translates Java camelCase to a more Oracle friendly snake_case. I am interested in doing the same thing with Lift Mapper/
Record. In the mean time, I need to override dbTableName and dbColumnName, but I'd like to avoid that and just agree on the convention.

I could imagine that the Database Provider would maintain a flag to say whether the database is case sensitive or not, and then based on that flag, would pass either the camelCase directly (for case sensitive), or snake_case(camelCase) (for case insensitive).

I am specifically using Lift Mapper with Oracle, but I think this issue relates to Lift Record too as well as any other case insensitive database.

See discussion: http://groups.google.com/group/liftweb/t/e7f85ede13b303da

Part of the build process should insure that archetypes are tested automatically

Please add valMinLen and valMaxLen validation functions to net.liftweb.record.field.StringField. If it's preferred to put them in a separate helper object that would be fine as well.

Also, a companion object for LocaleField that has a list of all possible locales (similar to TimeZoneField) would be nice.

Thanks

Module

Mapper

Request

When updating the referenced object in a MappedForeignKey column, the .is field returns the new value of the id, but the .obj field still returns the reference of the object originally loaded with the owner. The .is and .obj should be coherent.

Example

    val testUser : User = User.find(By(User.id, 1)).open_!
    val testAddress = testUser.homeAddress.obj;
    Log.debug("Current FK id is "+testUser.homeAddress.is) // returns 1

    testAddress match {
      case Full(addr) =>
         val anotherAddr : Address = Address.find(By(Address.id,2)).open_!
         Log.debug("There was an address: "+addr+". Replace it with addr id#2: "+anotherAddr);
          testUser.homeAddress(anotherAddr).save

      case _ =>
        [...]
    }
    val retestAddress = testUser.homeAddress.obj
    Log.debug("Now obj is "+retestAddress); // returns a Box with the old address (id == 1)
    Log.debug("And is is "+testUser.homeAddress.is) // returns 2 

See

http://groups.google.com/group/liftweb/browse_thread/thread/0041894ec33fc538/0bcb1a85658155af?hl=en#0bcb1a85658155af

Regarding to this thread

The *SNAPSHOT-sources.jar doesn't seem to contain any scala source files, only static resources

If javax.activation is merely a transitive dependency for javax.mail, then it should not be explicitly declared in http://scala-tools.org/repo-releases/net/liftweb/lift/1.1-M7/lift-1.1-M7.pom

Note that according to this: http://java.sun.com/products/javamail/
the Activation framework dependency is not necessary for Java 6 users because it is included.

See thread: http://groups.google.com/group/liftweb/browse_thread/thread/f435fd3a0ddbd08e

In Driver.scala there is no way to intercept singleton method calcDriver to allow user supported driver abstractions since in DB.scala DriverType object is used explicitly.

Missing a space after "URL" and before "was":

private[liftweb] def defaultCreateNotFound(in: Req) =
XhtmlResponse(( The Requested URL{in.contextPath + in.uri}was not found on this server ), ResponseInfo.docType(in), List("Content-Type" -> "text/html; charset=utf-8"), Nil, 404, S.ieMode)