On 5 Aug 2014 at 15:48:27, Jeremie BOUSQUET ([email protected](mailto:[email protected])) wrote:

Hello XWiki Community,

Got some issue with executing groovy in a page ...

I have a page with a groovy class, that needs some jars dependencies. The following "works":

{{groovy jars="attach:toto.jar,..."}}

import ...

public class MyClass { ... }

def myobj = new MyClass() // do something with myobj ... {{/groovy}}

Then I'd like to execute this code from another page. So I removed the {{groovy}} macro tags and the isolated code at the end, leaving only the class declaration and imports. From the other page I use the following version of parseGroovyFromPage: http://maven.xwiki.org/site/docs/xwiki-javadoc-4.1.x/com/xpn/xwiki/api/XWiki...)

Why not keep the {{groovy}} macro and instead use {{include reference=“…”/}} from the calling page? Using parseGroovyFromPage is the old way of doing it. Thanks -Vincent

I provide it with the name of the page containing my groovy class, twice (once for the script and once for the jars). It results in a page loading infinitely until it time-outs ... From the logs it seems my class never gets executed.

Any idea ?

Note: I think it's a "crappy" method, and I plan to put this as a java component, but in a "proof of concept" stage the groovy script is very useful.

Thanks, Jeremie

  On 16 Aug 2014 at 22:31:24, Jeremie BOUSQUET ([email protected](mailto:[email protected])) wrote:

Hi Vincent, Le 16 août 2014 21:42, "[email protected](mailto:[email protected])" a écrit :

...

On 5 Aug 2014 at 15:48:27, Jeremie BOUSQUET ([email protected](mailto:[email protected])(mailto:[email protected])) wrote:

...

Hello XWiki Community,

Got some issue with executing groovy in a page ...

I have a page with a groovy class, that needs some jars dependencies. The following "works":

{{groovy jars="attach:toto.jar,..."}}

import ...

public class MyClass { ... }

def myobj = new MyClass() // do something with myobj ... {{/groovy}}

Then I'd like to execute this code from another page. So I removed the {{groovy}} macro tags and the isolated code at the end, leaving only the class declaration and imports. From the other page I use the following version of parseGroovyFromPage: http://maven.xwiki.org/site/docs/xwiki-javadoc-4.1.x/com/xpn/xwiki/api/XWiki...)

Why not keep the {{groovy}} macro and instead use {{include reference=“…”/}} from the calling page? Using parseGroovyFromPage is the old way of doing it.

Right, but in a scheduler job I can only put groovy, not wiki syntax, isn't it ?

Indeed, something we really need to fix… You could still render a page using $doc.getRenderedContent() from a scheduler job and use wiki syntax 2.x in that page ;) Thanks -Vincent

...

Thanks -Vincent

...

I provide it with the name of the page containing my groovy class, twice (once for the script and once for the jars). It results in a page loading infinitely until it time-outs ... From the logs it seems my class never gets executed.

Any idea ?

Note: I think it's a "crappy" method, and I plan to put this as a java component, but in a "proof of concept" stage the groovy script is very useful.

Thanks, Jeremie

Hi Thomas, Le 27 août 2014 10:43, "Thomas Mortagne" <[email protected]> a écrit :

And you really don't want to install those jars as extensions ?

What I'm not fond of, is that they would be added to the classloader globally though their use will be very limited. Also those are not currently available in a maven repo because of licensing, and I'd have to manage this in my own repo. But, why not.

The jars parameter of the script macro is not really nice for performance as it create a new classloader and reload those jar each time you execute the macro.

An alternative is to use Groovy @Grab feature, see http://groovy.codehaus.org/Grape.

Yes I thought about it also. I'll see how I 'll manage this, thanks everyone !

On Sat, Aug 16, 2014 at 10:42 PM, [email protected] <[email protected]>

wrote:

...

On 16 Aug 2014 at 22:31:24, Jeremie BOUSQUET ([email protected]

(mailto:[email protected])) wrote:

...

...

Hi Vincent, Le 16 août 2014 21:42, "[email protected](mailto:[email protected])"

a écrit :

...

...

...

On 5 Aug 2014 at 15:48:27, Jeremie BOUSQUET (

[email protected](mailto:[email protected])(mailto: [email protected])) wrote:

...

...

Hello XWiki Community,

Got some issue with executing groovy in a page ...

I have a page with a groovy class, that needs some jars

dependencies.

...

...

The following "works":

{{groovy jars="attach:toto.jar,..."}}

import ...

public class MyClass { ... }

def myobj = new MyClass() // do something with myobj ... {{/groovy}}

Then I'd like to execute this code from another page. So I removed the {{groovy}} macro tags and the isolated code at the end, leaving only the class declaration and imports. From the other page I use the following version of parseGroovyFromPage:

http://maven.xwiki.org/site/docs/xwiki-javadoc-4.1.x/com/xpn/xwiki/api/XWiki...)

Why not keep the {{groovy}} macro and instead use {{include reference=“…”/}} from the calling page? Using parseGroovyFromPage is the old way of doing it.

Right, but in a scheduler job I can only put groovy, not wiki syntax, isn't it ?

Indeed, something we really need to fix… You could still render a page using $doc.getRenderedContent() from a scheduler job and use wiki syntax 2.x in that page ;)

Thanks -Vincent

...

...

Thanks -Vincent

...

I provide it with the name of the page containing my groovy class, twice (once for the script and once for the jars). It results in a page loading infinitely until it time-outs ... From the logs it seems my class never gets executed.

Any idea ?

Note: I think it's a "crappy" method, and I plan to put this as a java component, but in a "proof of concept" stage the groovy script is very useful.

Thanks, Jeremie

_______________________________________________ users mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/users

-- Thomas Mortagne