FreeMarker

On Sat, May 29, 2010 at 3:47 PM, Jolse Maginnis <jaymagz@...> wrote:
> Hi All,
>
> First of all, I'd like to say that FreeMarker is a great templating engine
> and I thank the developers for their great work.
>
> I understand that there's some debate as to whether or not the 2.4 version
> (or 3.0 now?) will ever be released as stable.
> >From the looks of the change list there are some very good enhancements and
> a few that I actually need to use, such as the ability for
> TemplateDirectiveModel's to receive parameters that were passed by position
> with the @Parameters annotation.
> Can somebody explain to me why this version isn't likely to become a stable
> release?

Because the spirit is willing but the flesh is weak.

IOW, there is (justifiable) doubt about the gumption of key developers
(primarily me, but also A. Szegedi) to continue working on this
codebase. In that case, the only further development activity will be
incremental tweaks on the stable (2.3.x) codebase.

The main thing happening now is that Angelo Zerr is working on a
completely new eclipse plugin that uses the DLTK (dynamic languages
toolkit) project. I encouraged him to develop against the trunk (3.0)
but Daniel Dekany disagrees with that. My view is that independently
of my own motivation to put in the work on the cutting edge codebase,
there is the bootstrap problem that, if we say that we are not
confident in the stability of the code and don't use it (even
internally in the project to develop new tools) then it's a vicious
cycle, since the only way to develop confidence in the robustness of
the codebase is in fact to use it.

That is the main part of my reasoning. Also, there is the fact that
there are more degrees of freedom to muck about with the trunk without
breaking anybody's existing projects. So I think that writing new
tools against it, where you may well need to add hooks and expose
capabilities in the core library to support the tooling being
developed, it is actually better to use the most recent codebase.

That is basically what the whole conversation is about. I hope that
answers your question, and I appreciate that you are interested enough
to ask!

Regards,

JR



>
> Regards,
> Jolse
>
> ------------------------------------------------------------------------------
>
>
> _______________________________________________
> FreeMarker-devel mailing list
> FreeMarker-devel@...
> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>
>

On Mon, May 10, 2010 at 9:56 AM, Angelo zerr <angelo.zerr@...> wrote:
> Hi,
>
> 2010/5/7 Guo Du <mrduguo@...>
>>
>> On Fri, May 7, 2010 at 9:05 AM, angelozerr <angelo.zerr@...> wrote:
>> > 2. Are you interested with this plugin?
>> Interested to have a more decent/native eclipse plugin. Will check it
>> out once available.
>
> Thank a lot Guo for your interesting.
>
> I have continue my work and resolve my problem with start/end char (When you
> type for instance <#if > into the editor, it display a red round error into
> the line (like JBoss Feemarker editor), but it
> highlight too the <#if > (JBoss doesn't manage that). So to manage start/end
> char I have
>
> 1. getted Freemarker source from trunk SVN and I have noticed that Token was
> public (great!)
> 2. I have created freemarker .jar and I have noticed that FMParser parser
> packages have changed (now it's freemarker.core.parser, before it was
> freemarker.core)

We have a lot of flexibility developing against the FreeMarker trunk,
since that is FreeMarker 2.4.x and there has only been a preview
release, so basically nobody is using it. We can move packages around
and change things if necessary without much of a problem.

> 3. I have computed start/end char by using beforeColumn, endColumn,
> beforeLine,/endLine information about Token, NUT I iterate each character
> from the editor to compute start/end char, because I don't know size of each
> line. I think it should be better if FMParser could computed directly
> start/end char. Jonathan could you see that when I will commit my work?

Yes, I think so. I guess you mean the offset of the character in the
whole file? I guess I never anticipated that anybody would want that!
But it is probably pretty easy to put that in.

I would encourage you to check in your work as soon as possible.
(Though it is up to you, of course...)

JR

> Thank a lot.
>
> Before commiting my work, I would like clean my code. I would like know if
> you are OK with me :
>
> 1. I have 3 plugins, which have named like this :
>   - org.eclipse.dltk.freemarker.core : core of DLTK Freemarker (DLTK
> Freemarker parser...)
>   - org.eclipse.dltk.freemarker.ui : UI of DLTK Freemarker (Editor...).
> org.eclipse.dltk.freemarker.core and org.eclipse.dltk.freemarker.core follow
> the same name like another DLTK implementation (ex :
> org.eclipse.dltk.ruby.core, org.eclipse.dltk.javscript.core...)
>   - org.freemarker : this plugin contains only the freemarker.jar and export
> packages (see MANIFEST.MF). IHMO, I think it should be better if MANIFEST.MF
> of freemarker will contains directly OSGi information like this :
>
> ---------
> Manifest-Version: 1.0
> Bundle-ManifestVersion: 2
> Bundle-Name: Freemarker
> Bundle-SymbolicName: org.freemarker
> Bundle-Version: 2.3.98
> Bundle-RequiredExecutionEnvironment: J2SE-1.5
> Export-Package: freemarker.cache,
>  freemarker.core,
>  freemarker.core.ast,
>  freemarker.core.builtins,
>  freemarker.core.helpers,
>  freemarker.core.ooparam,
>  freemarker.core.parser,
>  freemarker.debug,
>  freemarker.debug.impl,
>  freemarker.ext.ant,
>  freemarker.ext.beans,
>  freemarker.ext.dom,
>  freemarker.ext.jruby,
>  freemarker.ext.jsp,
>  freemarker.ext.jython,
>  freemarker.ext.rhino,
>  freemarker.ext.script,
>  freemarker.ext.servlet,
>  freemarker.ext.util,
>  freemarker.log,
>  freemarker.template,
>  freemarker.template.utility
> ---------
>
> Freemarker.jar would be an OSGi bundle (you can notice, there are not
> dependance of OSGi API).
>
> 2. I have started to add EPL license into Java header file, but I don't know
> if you are OK with that?
>
> As soon as, my clean is done and you will answer me, I will commit my
> (basic) work.
>
> Regards Angelo
>
>>
>> -Guo
>>
>>
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> FreeMarker-devel mailing list
>> FreeMarker-devel@...
>> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>
>
> ------------------------------------------------------------------------------
>
>
> _______________________________________________
> FreeMarker-devel mailing list
> FreeMarker-devel@...
> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>
>

Ok,

I have commited my work on
https://freemarker.svn.sourceforge.net/svnroot/freemarker/sandbox/org.eclipse.dltk.freemarker/SVN.
You can get the 3 plugins that I have explained before.

I have developped that with Eclipse Galileo and you must install DLTK
(version 1.0.2) http://download.eclipse.org/technology/dltk/updates-dev/1.0

Jonathan, please see FreemarkerSourceParser class, I have write some
questions (search token @Jonathan).
I think Token should manage startOffset/endOffset to improve performance.

Regards Angelo

2010/5/10 Jonathan Revusky <revusky@...>

> On Mon, May 10, 2010 at 9:56 AM, Angelo zerr <angelo.zerr@...>
> wrote:
> > Hi,
> >
> > 2010/5/7 Guo Du <mrduguo@...>
> >>
> >> On Fri, May 7, 2010 at 9:05 AM, angelozerr <angelo.zerr@...>
> wrote:
> >> > 2. Are you interested with this plugin?
> >> Interested to have a more decent/native eclipse plugin. Will check it
> >> out once available.
> >
> > Thank a lot Guo for your interesting.
> >
> > I have continue my work and resolve my problem with start/end char (When
> you
> > type for instance <#if > into the editor, it display a red round error
> into
> > the line (like JBoss Feemarker editor), but it
> > highlight too the <#if > (JBoss doesn't manage that). So to manage
> start/end
> > char I have
> >
> > 1. getted Freemarker source from trunk SVN and I have noticed that Token
> was
> > public (great!)
> > 2. I have created freemarker .jar and I have noticed that FMParser parser
> > packages have changed (now it's freemarker.core.parser, before it was
> > freemarker.core)
>
> We have a lot of flexibility developing against the FreeMarker trunk,
> since that is FreeMarker 2.4.x and there has only been a preview
> release, so basically nobody is using it. We can move packages around
> and change things if necessary without much of a problem.
>
> > 3. I have computed start/end char by using beforeColumn, endColumn,
> > beforeLine,/endLine information about Token, NUT I iterate each character
> > from the editor to compute start/end char, because I don't know size of
> each
> > line. I think it should be better if FMParser could computed directly
> > start/end char. Jonathan could you see that when I will commit my work?
>
> Yes, I think so. I guess you mean the offset of the character in the
> whole file? I guess I never anticipated that anybody would want that!
> But it is probably pretty easy to put that in.
>
> I would encourage you to check in your work as soon as possible.
> (Though it is up to you, of course...)
>
> JR
>
> > Thank a lot.
> >
> > Before commiting my work, I would like clean my code. I would like know
> if
> > you are OK with me :
> >
> > 1. I have 3 plugins, which have named like this :
> >   - org.eclipse.dltk.freemarker.core : core of DLTK Freemarker (DLTK
> > Freemarker parser...)
> >   - org.eclipse.dltk.freemarker.ui : UI of DLTK Freemarker (Editor...).
> > org.eclipse.dltk.freemarker.core and org.eclipse.dltk.freemarker.core
> follow
> > the same name like another DLTK implementation (ex :
> > org.eclipse.dltk.ruby.core, org.eclipse.dltk.javscript.core...)
> >   - org.freemarker : this plugin contains only the freemarker.jar and
> export
> > packages (see MANIFEST.MF). IHMO, I think it should be better if
> MANIFEST.MF
> > of freemarker will contains directly OSGi information like this :
> >
> > ---------
> > Manifest-Version: 1.0
> > Bundle-ManifestVersion: 2
> > Bundle-Name: Freemarker
> > Bundle-SymbolicName: org.freemarker
> > Bundle-Version: 2.3.98
> > Bundle-RequiredExecutionEnvironment: J2SE-1.5
> > Export-Package: freemarker.cache,
> >  freemarker.core,
> >  freemarker.core.ast,
> >  freemarker.core.builtins,
> >  freemarker.core.helpers,
> >  freemarker.core.ooparam,
> >  freemarker.core.parser,
> >  freemarker.debug,
> >  freemarker.debug.impl,
> >  freemarker.ext.ant,
> >  freemarker.ext.beans,
> >  freemarker.ext.dom,
> >  freemarker.ext.jruby,
> >  freemarker.ext.jsp,
> >  freemarker.ext.jython,
> >  freemarker.ext.rhino,
> >  freemarker.ext.script,
> >  freemarker.ext.servlet,
> >  freemarker.ext.util,
> >  freemarker.log,
> >  freemarker.template,
> >  freemarker.template.utility
> > ---------
> >
> > Freemarker.jar would be an OSGi bundle (you can notice, there are not
> > dependance of OSGi API).
> >
> > 2. I have started to add EPL license into Java header file, but I don't
> know
> > if you are OK with that?
> >
> > As soon as, my clean is done and you will answer me, I will commit my
> > (basic) work.
> >
> > Regards Angelo
> >
> >>
> >> -Guo
> >>
> >>
> >>
> ------------------------------------------------------------------------------
> >>
> >> _______________________________________________
> >> FreeMarker-devel mailing list
> >> FreeMarker-devel@...
> >> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
> >
> >
> >
> ------------------------------------------------------------------------------
> >
> >
> > _______________________________________________
> > FreeMarker-devel mailing list
> > FreeMarker-devel@...
> > https://lists.sourceforge.net/lists/listinfo/freemarker-devel
> >
> >
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> FreeMarker-devel mailing list
> FreeMarker-devel@...
> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>

Monday, May 10, 2010, 9:56:20 AM, Angelo zerr wrote:

> Hi,
>
> 2010/5/7 Guo Du <mrduguo@...>
> On Fri, May 7, 2010 at 9:05 AM, angelozerr <angelo.zerr@...> wrote:
>> 2. Are you interested with this plugin?
> Interested to have a more decent/native eclipse plugin. Will check it
> out once available.
> Thank a lot Guo for your interesting.
>
> I have continue my work and resolve my problem with start/end char
> (When you type for instance <#if > into the editor, it display a red
> round error into the line (like JBoss Feemarker editor), but it 
>
> highlight too the <#if > (JBoss doesn't manage that). So to manage start/end char I have
>
> 1. getted Freemarker source from trunk SVN and I have noticed that Token was public (great!).

The SNV trunk? Watch out... The trunk is the 3.0, and it possibly will
not not be ever released, but surely not in the foreseeable future.
You should work against branches/2.3. (Note that any changes in
branches/2.3 should be forward-ported into branches/2.3-gae and
branches/2.4.) You will notice that 2.4 and earlier has an inferior
API compared to 3.0 when it comes examining parsed templates. Of
course, as far as backward compatibility and stability is kept,
improvements can be done in 2.3 too. If a such change is non-BC but
with minor impact on the causal users, then it still can be done in
2.4.

Committing changes into FreeMarker requires a signed paper CLA sent to
us, because corporate lawyers have problems with using FreeMarker
otherwise. I suggest that the same should stand for the sandbox. If
this plugin later becomes part of JBoss IDE or like, the unclear legal
status can become a problem.

> 2. I have created freemarker .jar and I have noticed that FMParser
> parser packages have changed (now it's freemarker.core.parser, before it was freemarker.core)
> 3. I have computed start/end char by using beforeColumn, endColumn,
> beforeLine,/endLine information about Token, NUT I iterate each
> character from the editor to compute start/end char, because I don't
> know size of each line. I think it should be better if FMParser
> could computed directly start/end char. Jonathan could you see that
> when I will commit my work? Thank a lot.
>
> Before commiting my work, I would like clean my code. I would like know if you are OK with me :
>
> 1. I have 3 plugins, which have named like this : 
>   - org.eclipse.dltk.freemarker.core : core of DLTK Freemarker (DLTK Freemarker parser...)
>   - org.eclipse.dltk.freemarker.ui : UI of DLTK Freemarker
> (Editor...). org.eclipse.dltk.freemarker.core and
> org.eclipse.dltk.freemarker.core follow the same name like another
> DLTK implementation (ex : org.eclipse.dltk.ruby.core,
> org.eclipse.dltk.javscript.core...)
>   - org.freemarker : this plugin contains only the freemarker.jar
> and export packages (see MANIFEST.MF). IHMO, I think it should be
> better if MANIFEST.MF of freemarker will contains directly OSGi information like this :
>
> ---------
> Manifest-Version: 1.0
> Bundle-ManifestVersion: 2
> Bundle-Name: Freemarker
> Bundle-SymbolicName: org.freemarker
> Bundle-Version: 2.3.98
> Bundle-RequiredExecutionEnvironment: J2SE-1.5
> Export-Package: freemarker.cache,
>  freemarker.core,
>  freemarker.core.ast,
>  freemarker.core.builtins,
>  freemarker.core.helpers,
>  freemarker.core.ooparam,
>  freemarker.core.parser,
>  freemarker.debug,
>  freemarker.debug.impl,
>  freemarker.ext.ant,
>  freemarker.ext.beans,
>  freemarker.ext.dom,
>  freemarker.ext.jruby,
>  freemarker.ext.jsp,
>  freemarker.ext.jython,
>  freemarker.ext.rhino,
>  freemarker.ext.script,
>  freemarker.ext.servlet,
>  freemarker.ext.util,
>  freemarker.log,
>  freemarker.template,
>  freemarker.template.utility
> ---------
>
> Freemarker.jar would be an OSGi bundle (you can notice, there are not dependance of OSGi API).
>
> 2. I have started to add EPL license into Java header file, but I
> don't know if you are OK with that? 
>
> As soon as, my clean is done and you will answer me, I will commit my (basic) work.
>
> Regards Angelo
>
>
> -Guo
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
>
> FreeMarker-devel mailing list
> FreeMarker-devel@...
> https://lists.sourceforge.net/lists/listinfo/freemarker-devel

-- 
Best regards,
 Daniel Dekany

Monday, May 10, 2010, 12:52:02 PM, Jonathan Revusky wrote:

> On Mon, May 10, 2010 at 9:56 AM, Angelo zerr <angelo.zerr@...> wrote:
>> Hi,
>>
>> 2010/5/7 Guo Du <mrduguo@...>
>>>
>>> On Fri, May 7, 2010 at 9:05 AM, angelozerr <angelo.zerr@...> wrote:
>>> > 2. Are you interested with this plugin?
>>> Interested to have a more decent/native eclipse plugin. Will check it
>>> out once available.
>>
>> Thank a lot Guo for your interesting.
>>
>> I have continue my work and resolve my problem with start/end char (When you
>> type for instance <#if > into the editor, it display a red round error into
>> the line (like JBoss Feemarker editor), but it
>> highlight too the <#if > (JBoss doesn't manage that). So to manage start/end
>> char I have
>>
>> 1. getted Freemarker source from trunk SVN and I have noticed that Token was
>> public (great!)
>> 2. I have created freemarker .jar and I have noticed that FMParser parser
>> packages have changed (now it's freemarker.core.parser, before it was
>> freemarker.core)
>
> We have a lot of flexibility developing against the FreeMarker trunk,
> since that is FreeMarker 2.4.x

No, it's FreeMarker 3.0.0 for a while.

> and there has only been a preview
> release, so basically nobody is using it. We can move packages around
> and change things if necessary without much of a problem.
>
>> 3. I have computed start/end char by using beforeColumn, endColumn,
>> beforeLine,/endLine information about Token, NUT I iterate each character
>> from the editor to compute start/end char, because I don't know size of each
>> line. I think it should be better if FMParser could computed directly
>> start/end char. Jonathan could you see that when I will commit my work?
>
> Yes, I think so. I guess you mean the offset of the character in the
> whole file? I guess I never anticipated that anybody would want that!
> But it is probably pretty easy to put that in.
>
> I would encourage you to check in your work as soon as possible.
> (Though it is up to you, of course...)
>
> JR
>
>> Thank a lot.
>>
>> Before commiting my work, I would like clean my code. I would like know if
>> you are OK with me :
>>
>> 1. I have 3 plugins, which have named like this :
>>   - org.eclipse.dltk.freemarker.core : core of DLTK Freemarker (DLTK
>> Freemarker parser...)
>>   - org.eclipse.dltk.freemarker.ui : UI of DLTK Freemarker (Editor...).
>> org.eclipse.dltk.freemarker.core and org.eclipse.dltk.freemarker.core follow
>> the same name like another DLTK implementation (ex :
>> org.eclipse.dltk.ruby.core, org.eclipse.dltk.javscript.core...)
>>   - org.freemarker : this plugin contains only the freemarker.jar and export
>> packages (see MANIFEST.MF). IHMO, I think it should be better if MANIFEST.MF
>> of freemarker will contains directly OSGi information like this :
>>
>> ---------
>> Manifest-Version: 1.0
>> Bundle-ManifestVersion: 2
>> Bundle-Name: Freemarker
>> Bundle-SymbolicName: org.freemarker
>> Bundle-Version: 2.3.98
>> Bundle-RequiredExecutionEnvironment: J2SE-1.5
>> Export-Package: freemarker.cache,
>>  freemarker.core,
>>  freemarker.core.ast,
>>  freemarker.core.builtins,
>>  freemarker.core.helpers,
>>  freemarker.core.ooparam,
>>  freemarker.core.parser,
>>  freemarker.debug,
>>  freemarker.debug.impl,
>>  freemarker.ext.ant,
>>  freemarker.ext.beans,
>>  freemarker.ext.dom,
>>  freemarker.ext.jruby,
>>  freemarker.ext.jsp,
>>  freemarker.ext.jython,
>>  freemarker.ext.rhino,
>>  freemarker.ext.script,
>>  freemarker.ext.servlet,
>>  freemarker.ext.util,
>>  freemarker.log,
>>  freemarker.template,
>>  freemarker.template.utility
>> ---------
>>
>> Freemarker.jar would be an OSGi bundle (you can notice, there are not
>> dependance of OSGi API).
>>
>> 2. I have started to add EPL license into Java header file, but I don't know
>> if you are OK with that?
>>
>> As soon as, my clean is done and you will answer me, I will commit my
>> (basic) work.
>>
>> Regards Angelo
>>
>>>
>>> -Guo
>>>
>>>
>>> ------------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> FreeMarker-devel mailing list
>>> FreeMarker-devel@...
>>> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>>
>>
>> ------------------------------------------------------------------------------
>>
>>
>> _______________________________________________
>> FreeMarker-devel mailing list
>> FreeMarker-devel@...
>> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>>
>>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> FreeMarker-devel mailing list
> FreeMarker-devel@...
> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>

-- 
Best regards,
 Daniel Dekany

Hi Daniel,

Ok, I have getted Freemarker 2.3 source from SVN and Token are not visible
-(

>Committing changes into FreeMarker requires a signed paper CLA sent to
>us, because corporate lawyers have problems with using FreeMarker
>otherwise. I suggest that the same should stand for the sandbox. If
>this plugin later becomes part of JBoss IDE or like, the unclear legal
>status can become a problem.

Ok Daniel, I don't have distrurb with you. You wish I stop commit my work?
Could you give me the link of the paper CLA please? If Token from Freemarker
2.3 can not be change (public class + manage startOffset/endOffset), it will
be difficult for me to continue to develop DLTK Freemarker. Please tell me
if change could be done?

Thank a lot.

Regards Angelo

2010/5/10 Daniel Dekany <ddekany@...>

> Monday, May 10, 2010, 12:52:02 PM, Jonathan Revusky wrote:
>
> > On Mon, May 10, 2010 at 9:56 AM, Angelo zerr <angelo.zerr@...>
> wrote:
> >> Hi,
> >>
> >> 2010/5/7 Guo Du <mrduguo@...>
> >>>
> >>> On Fri, May 7, 2010 at 9:05 AM, angelozerr <angelo.zerr@...>
> wrote:
> >>> > 2. Are you interested with this plugin?
> >>> Interested to have a more decent/native eclipse plugin. Will check it
> >>> out once available.
> >>
> >> Thank a lot Guo for your interesting.
> >>
> >> I have continue my work and resolve my problem with start/end char (When
> you
> >> type for instance <#if > into the editor, it display a red round error
> into
> >> the line (like JBoss Feemarker editor), but it
> >> highlight too the <#if > (JBoss doesn't manage that). So to manage
> start/end
> >> char I have
> >>
> >> 1. getted Freemarker source from trunk SVN and I have noticed that Token
> was
> >> public (great!)
> >> 2. I have created freemarker .jar and I have noticed that FMParser
> parser
> >> packages have changed (now it's freemarker.core.parser, before it was
> >> freemarker.core)
> >
> > We have a lot of flexibility developing against the FreeMarker trunk,
> > since that is FreeMarker 2.4.x
>
> No, it's FreeMarker 3.0.0 for a while.
>
> > and there has only been a preview
> > release, so basically nobody is using it. We can move packages around
> > and change things if necessary without much of a problem.
> >
> >> 3. I have computed start/end char by using beforeColumn, endColumn,
> >> beforeLine,/endLine information about Token, NUT I iterate each
> character
> >> from the editor to compute start/end char, because I don't know size of
> each
> >> line. I think it should be better if FMParser could computed directly
> >> start/end char. Jonathan could you see that when I will commit my work?
> >
> > Yes, I think so. I guess you mean the offset of the character in the
> > whole file? I guess I never anticipated that anybody would want that!
> > But it is probably pretty easy to put that in.
> >
> > I would encourage you to check in your work as soon as possible.
> > (Though it is up to you, of course...)
> >
> > JR
> >
> >> Thank a lot.
> >>
> >> Before commiting my work, I would like clean my code. I would like know
> if
> >> you are OK with me :
> >>
> >> 1. I have 3 plugins, which have named like this :
> >>   - org.eclipse.dltk.freemarker.core : core of DLTK Freemarker (DLTK
> >> Freemarker parser...)
> >>   - org.eclipse.dltk.freemarker.ui : UI of DLTK Freemarker (Editor...).
> >> org.eclipse.dltk.freemarker.core and org.eclipse.dltk.freemarker.core
> follow
> >> the same name like another DLTK implementation (ex :
> >> org.eclipse.dltk.ruby.core, org.eclipse.dltk.javscript.core...)
> >>   - org.freemarker : this plugin contains only the freemarker.jar and
> export
> >> packages (see MANIFEST.MF). IHMO, I think it should be better if
> MANIFEST.MF
> >> of freemarker will contains directly OSGi information like this :
> >>
> >> ---------
> >> Manifest-Version: 1.0
> >> Bundle-ManifestVersion: 2
> >> Bundle-Name: Freemarker
> >> Bundle-SymbolicName: org.freemarker
> >> Bundle-Version: 2.3.98
> >> Bundle-RequiredExecutionEnvironment: J2SE-1.5
> >> Export-Package: freemarker.cache,
> >>  freemarker.core,
> >>  freemarker.core.ast,
> >>  freemarker.core.builtins,
> >>  freemarker.core.helpers,
> >>  freemarker.core.ooparam,
> >>  freemarker.core.parser,
> >>  freemarker.debug,
> >>  freemarker.debug.impl,
> >>  freemarker.ext.ant,
> >>  freemarker.ext.beans,
> >>  freemarker.ext.dom,
> >>  freemarker.ext.jruby,
> >>  freemarker.ext.jsp,
> >>  freemarker.ext.jython,
> >>  freemarker.ext.rhino,
> >>  freemarker.ext.script,
> >>  freemarker.ext.servlet,
> >>  freemarker.ext.util,
> >>  freemarker.log,
> >>  freemarker.template,
> >>  freemarker.template.utility
> >> ---------
> >>
> >> Freemarker.jar would be an OSGi bundle (you can notice, there are not
> >> dependance of OSGi API).
> >>
> >> 2. I have started to add EPL license into Java header file, but I don't
> know
> >> if you are OK with that?
> >>
> >> As soon as, my clean is done and you will answer me, I will commit my
> >> (basic) work.
> >>
> >> Regards Angelo
> >>
> >>>
> >>> -Guo
> >>>
> >>>
> >>>
> ------------------------------------------------------------------------------
> >>>
> >>> _______________________________________________
> >>> FreeMarker-devel mailing list
> >>> FreeMarker-devel@...
> >>> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
> >>
> >>
> >>
> ------------------------------------------------------------------------------
> >>
> >>
> >> _______________________________________________
> >> FreeMarker-devel mailing list
> >> FreeMarker-devel@...
> >> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
> >>
> >>
> >
> >
> ------------------------------------------------------------------------------
> >
> > _______________________________________________
> > FreeMarker-devel mailing list
> > FreeMarker-devel@...
> > https://lists.sourceforge.net/lists/listinfo/freemarker-devel
> >
>
> --
> Best regards,
>  Daniel Dekany
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> FreeMarker-devel mailing list
> FreeMarker-devel@...
> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>

Tuesday, May 11, 2010, 1:24:11 AM, Angelo zerr wrote:

> Hi Daniel,
>
> Ok, I have getted Freemarker 2.3 source from SVN and Token are not visible -(
>
>>Committing changes into FreeMarker requires a signed paper CLA sent to
>>us, because corporate lawyers have problems with using FreeMarker
>>otherwise. I suggest that the same should stand for the sandbox. If
>>this plugin later becomes part of JBoss IDE or like, the unclear legal
>>status can become a problem.
>
> Ok Daniel, I don't have distrurb with you. You wish I stop commit
> my work?

I think that would be wise. However, it's not up to me. There was an
agreement regarding the code of FreeMarker itself (that CLA is
required), but not regarding the sandbox. So let's see what the other
two admins say.

Anyway, I think you better fill the CLA, and then at least you can
commit into FreeMarker as well. As of the stuff committed so far into
the sandbox... I'm not a lawyer, so I really don't know how to get out
of that now. But I guess we can't do more than deleting the current
commitment and then later when you have sent the CLA you just
re-commit it. /-: If that doesn't please the lawyers, well, too bad,
we won't lose a plugin because of that.

I'm sorry the other didn't warn you about the turn and the CLA.

> Could you give me the link of the paper CLA please?

I send it to you in private.

> If Token from Freemarker 2.3 can not be change (public class +
> manage startOffset/endOffset), it will be difficult for me to
> continue to develop DLTK Freemarker. Please tell me if change could
> be done?

Certainly... making things public that were not public should not
break backward compatibility after all. (I didn't look into the source
code however.)

> Thank a lot.
>
> Regards Angelo
>
> 2010/5/10 Daniel Dekany <ddekany@...>
> Monday, May 10, 2010, 12:52:02 PM, Jonathan Revusky wrote:
>
>> On Mon, May 10, 2010 at 9:56 AM, Angelo zerr <angelo.zerr@...> wrote:
>>> Hi,
>>>
>>> 2010/5/7 Guo Du <mrduguo@...>
>>>>
>>>> On Fri, May 7, 2010 at 9:05 AM, angelozerr <angelo.zerr@...> wrote:
>>>> > 2. Are you interested with this plugin?
>>>> Interested to have a more decent/native eclipse plugin. Will check it
>>>> out once available.
>>>
>>> Thank a lot Guo for your interesting.
>>>
>>> I have continue my work and resolve my problem with start/end char (When you
>>> type for instance <#if > into the editor, it display a red round error into
>>> the line (like JBoss Feemarker editor), but it
>>> highlight too the <#if > (JBoss doesn't manage that). So to manage start/end
>>> char I have
>>>
>>> 1. getted Freemarker source from trunk SVN and I have noticed that Token was
>>> public (great!)
>>> 2. I have created freemarker .jar and I have noticed that FMParser parser
>>> packages have changed (now it's freemarker.core.parser, before it was
>>> freemarker.core)
>>
>> We have a lot of flexibility developing against the FreeMarker trunk,
>> since that is FreeMarker 2.4.x
>
> No, it's FreeMarker 3.0.0 for a while.
>
>
>> and there has only been a preview
>> release, so basically nobody is using it. We can move packages around
>> and change things if necessary without much of a problem.
>>
>>> 3. I have computed start/end char by using beforeColumn, endColumn,
>>> beforeLine,/endLine information about Token, NUT I iterate each character
>>> from the editor to compute start/end char, because I don't know size of each
>>> line. I think it should be better if FMParser could computed directly
>>> start/end char. Jonathan could you see that when I will commit my work?
>>
>> Yes, I think so. I guess you mean the offset of the character in the
>> whole file? I guess I never anticipated that anybody would want that!
>> But it is probably pretty easy to put that in.
>>
>> I would encourage you to check in your work as soon as possible.
>> (Though it is up to you, of course...)
>>
>> JR
>>
>>> Thank a lot.
>>>
>>> Before commiting my work, I would like clean my code. I would like know if
>>> you are OK with me :
>>>
>>> 1. I have 3 plugins, which have named like this :
>>>   - org.eclipse.dltk.freemarker.core : core of DLTK Freemarker (DLTK
>>> Freemarker parser...)
>>>   - org.eclipse.dltk.freemarker.ui : UI of DLTK Freemarker (Editor...).
>>> org.eclipse.dltk.freemarker.core and org.eclipse.dltk.freemarker.core follow
>>> the same name like another DLTK implementation (ex :
>>> org.eclipse.dltk.ruby.core, org.eclipse.dltk.javscript.core...)
>>>   - org.freemarker : this plugin contains only the freemarker.jar and export
>>> packages (see MANIFEST.MF). IHMO, I think it should be better if MANIFEST.MF
>>> of freemarker will contains directly OSGi information like this :
>>>
>>> ---------
>>> Manifest-Version: 1.0
>>> Bundle-ManifestVersion: 2
>>> Bundle-Name: Freemarker
>>> Bundle-SymbolicName: org.freemarker
>>> Bundle-Version: 2.3.98
>>> Bundle-RequiredExecutionEnvironment: J2SE-1.5
>>> Export-Package: freemarker.cache,
>>>  freemarker.core,
>>>  freemarker.core.ast,
>>>  freemarker.core.builtins,
>>>  freemarker.core.helpers,
>>>  freemarker.core.ooparam,
>>>  freemarker.core.parser,
>>>  freemarker.debug,
>>>  freemarker.debug.impl,
>>>  freemarker.ext.ant,
>>>  freemarker.ext.beans,
>>>  freemarker.ext.dom,
>>>  freemarker.ext.jruby,
>>>  freemarker.ext.jsp,
>>>  freemarker.ext.jython,
>>>  freemarker.ext.rhino,
>>>  freemarker.ext.script,
>>>  freemarker.ext.servlet,
>>>  freemarker.ext.util,
>>>  freemarker.log,
>>>  freemarker.template,
>>>  freemarker.template.utility
>>> ---------
>>>
>>> Freemarker.jar would be an OSGi bundle (you can notice, there are not
>>> dependance of OSGi API).
>>>
>>> 2. I have started to add EPL license into Java header file, but I don't know
>>> if you are OK with that?
>>>
>>> As soon as, my clean is done and you will answer me, I will commit my
>>> (basic) work.
>>>
>>> Regards Angelo
>>>
>>>>
>>>> -Guo
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>>
>>>> _______________________________________________
>>>> FreeMarker-devel mailing list
>>>> FreeMarker-devel@...
>>>> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>>>
>>>
>>> ------------------------------------------------------------------------------
>>>
>>>
>>> _______________________________________________
>>> FreeMarker-devel mailing list
>>> FreeMarker-devel@...
>>> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>>>
>>>
>>
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> FreeMarker-devel mailing list
>> FreeMarker-devel@...
>> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>>
>
> --
> Best regards,
>  Daniel Dekany
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> FreeMarker-devel mailing list
> FreeMarker-devel@...
> https://lists.sourceforge.net/lists/listinfo/freemarker-devel

-- 
Best regards,
 Daniel Dekany

On Tue, May 11, 2010 at 2:11 AM, Daniel Dekany <ddekany@...> wrote:
> Tuesday, May 11, 2010, 1:24:11 AM, Angelo zerr wrote:
>
>> Hi Daniel,
>>
>> Ok, I have getted Freemarker 2.3 source from SVN and Token are not visible -(
>>
>>>Committing changes into FreeMarker requires a signed paper CLA sent to
>>>us, because corporate lawyers have problems with using FreeMarker
>>>otherwise. I suggest that the same should stand for the sandbox. If
>>>this plugin later becomes part of JBoss IDE or like, the unclear legal
>>>status can become a problem.
>>
>> Ok Daniel, I don't have distrurb with you. You wish I stop commit
>> my work?
>
> I think that would be wise. However, it's not up to me. There was an
> agreement regarding the code of FreeMarker itself (that CLA is
> required), but not regarding the sandbox. So let's see what the other
> two admins say.

I have no recollection of sending in the CLA. Did I? If not, I will, I
was barely paying attention to that whole conversation.

Anyway, I don't know whether it should really be necessary to sign the
CLA in order to commit code to /sandbox. Though Angelo might as well
send in the CLA anyway...


> Anyway, I think you better fill the CLA, and then at least you can
> commit into FreeMarker as well. As of the stuff committed so far into
> the sandbox... I'm not a lawyer, so I really don't know how to get out
> of that now. But I guess we can't do more than deleting the current
> commitment and then later when you have sent the CLA you just
> re-commit it. /-: If that doesn't please the lawyers, well, too bad,
> we won't lose a plugin because of that.

I think this gets silly. Just leave the code there. He's expressed
willingness to send in the CLA thing and he will. What difference does
it make?

JR

Tuesday, May 11, 2010, 2:26:16 PM, Jonathan Revusky wrote:

> On Tue, May 11, 2010 at 2:11 AM, Daniel Dekany <ddekany@...> wrote:
>> Tuesday, May 11, 2010, 1:24:11 AM, Angelo zerr wrote:
>>
>>> Hi Daniel,
>>>
>>> Ok, I have getted Freemarker 2.3 source from SVN and Token are not visible -(
>>>
>>>>Committing changes into FreeMarker requires a signed paper CLA sent to
>>>>us, because corporate lawyers have problems with using FreeMarker
>>>>otherwise. I suggest that the same should stand for the sandbox. If
>>>>this plugin later becomes part of JBoss IDE or like, the unclear legal
>>>>status can become a problem.
>>>
>>> Ok Daniel, I don't have distrurb with you. You wish I stop commit
>>> my work?
>>
>> I think that would be wise. However, it's not up to me. There was an
>> agreement regarding the code of FreeMarker itself (that CLA is
>> required), but not regarding the sandbox. So let's see what the other
>> two admins say.
>
> I have no recollection of sending in the CLA. Did I? If not, I will, I
> was barely paying attention to that whole conversation.

Well, I don't know how you bypassed this... but it doesn't mater
anyway.

> Anyway, I don't know whether it should really be necessary to sign the
> CLA in order to commit code to /sandbox. Though Angelo might as well
> send in the CLA anyway...
>
>
>> Anyway, I think you better fill the CLA, and then at least you can
>> commit into FreeMarker as well. As of the stuff committed so far into
>> the sandbox... I'm not a lawyer, so I really don't know how to get out
>> of that now. But I guess we can't do more than deleting the current
>> commitment and then later when you have sent the CLA you just
>> re-commit it. /-: If that doesn't please the lawyers, well, too bad,
>> we won't lose a plugin because of that.
>
> I think this gets silly. Just leave the code there. He's expressed
> willingness to send in the CLA thing and he will. What difference does
> it make?

For lawyers, I don't know... law is silly after all. But maybe it's
indeed useless. I'm just saying it's the most we can do now.

> JR

-- 
Best regards,
 Daniel Dekany

Hi,

At first, excuse me Daniel I have not sent my CLA (I will do that this
week). but I have commited my work in order to Jonathan can see the plugin
in action. I have created a documentation wich explain how install/test DLTK
Freemarker and which show you (with screenshot) some feature of DLTK
Freemarker. You can find this documentation on SVN
https://freemarker.svn.sourceforge.net/svnroot/freemarker/sandbox/org.eclipse.dltk.freemarker/docs/.
This folder contains DLTK-Freemarker
- Eclipse Plugin.odt<https://freemarker.svn.sourceforge.net/svnroot/freemarker/sandbox/org.eclipse.dltk.freemarker/docs/DLTK-Freemarker%20-%20Eclipse%20Plugin.odt>which
is an OpenOffice ODT (hope you have OppenOffice).

Don't hesitate to change the documentation to add new ideas about new
features.

Regards Angelo

2010/5/11 Daniel Dekany <ddekany@...>

> Tuesday, May 11, 2010, 2:26:16 PM, Jonathan Revusky wrote:
>
> > On Tue, May 11, 2010 at 2:11 AM, Daniel Dekany <ddekany@...>
> wrote:
> >> Tuesday, May 11, 2010, 1:24:11 AM, Angelo zerr wrote:
> >>
> >>> Hi Daniel,
> >>>
> >>> Ok, I have getted Freemarker 2.3 source from SVN and Token are not
> visible -(
> >>>
> >>>>Committing changes into FreeMarker requires a signed paper CLA sent to
> >>>>us, because corporate lawyers have problems with using FreeMarker
> >>>>otherwise. I suggest that the same should stand for the sandbox. If
> >>>>this plugin later becomes part of JBoss IDE or like, the unclear legal
> >>>>status can become a problem.
> >>>
> >>> Ok Daniel, I don't have distrurb with you. You wish I stop commit
> >>> my work?
> >>
> >> I think that would be wise. However, it's not up to me. There was an
> >> agreement regarding the code of FreeMarker itself (that CLA is
> >> required), but not regarding the sandbox. So let's see what the other
> >> two admins say.
> >
> > I have no recollection of sending in the CLA. Did I? If not, I will, I
> > was barely paying attention to that whole conversation.
>
> Well, I don't know how you bypassed this... but it doesn't mater
> anyway.
>
> > Anyway, I don't know whether it should really be necessary to sign the
> > CLA in order to commit code to /sandbox. Though Angelo might as well
> > send in the CLA anyway...
> >
> >
> >> Anyway, I think you better fill the CLA, and then at least you can
> >> commit into FreeMarker as well. As of the stuff committed so far into
> >> the sandbox... I'm not a lawyer, so I really don't know how to get out
> >> of that now. But I guess we can't do more than deleting the current
> >> commitment and then later when you have sent the CLA you just
> >> re-commit it. /-: If that doesn't please the lawyers, well, too bad,
> >> we won't lose a plugin because of that.
> >
> > I think this gets silly. Just leave the code there. He's expressed
> > willingness to send in the CLA thing and he will. What difference does
> > it make?
>
> For lawyers, I don't know... law is silly after all. But maybe it's
> indeed useless. I'm just saying it's the most we can do now.
>
> > JR
>
> --
> Best regards,
>  Daniel Dekany
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> FreeMarker-devel mailing list
> FreeMarker-devel@...
> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>

Hi,

Juste for information, I have commited my work on DLTK Freemarker and now
you can use Run As->Freemarker to launch your FTL template (like a Java
class). It display result of the merge into Eclipse Console. If you are
interested  , plese read section "Run As → Freemarker" into DLTK-Freemarker
- Eclipse Plugin.odt<https://freemarker.svn.sourceforge.net/svnroot/freemarker/sandbox/org.eclipse.dltk.freemarker/docs/DLTK-Freemarker%20-%20Eclipse%20Plugin.odt>
You will see screenshot.

Now I would like study how we could debug FTL with DBGp (it seems it's a big
work).

Regards Angelo

2010/5/18 Angelo zerr <angelo.zerr@...>

> Hi,
>
> At first, excuse me Daniel I have not sent my CLA (I will do that this
> week). but I have commited my work in order to Jonathan can see the plugin
> in action. I have created a documentation wich explain how install/test DLTK
> Freemarker and which show you (with screenshot) some feature of DLTK
> Freemarker. You can find this documentation on SVN
> https://freemarker.svn.sourceforge.net/svnroot/freemarker/sandbox/org.eclipse.dltk.freemarker/docs/. This folder contains DLTK-Freemarker
> - Eclipse Plugin.odt<https://freemarker.svn.sourceforge.net/svnroot/freemarker/sandbox/org.eclipse.dltk.freemarker/docs/DLTK-Freemarker%20-%20Eclipse%20Plugin.odt>which is an OpenOffice ODT (hope you have OppenOffice).
>
> Don't hesitate to change the documentation to add new ideas about new
> features.
>
> Regards Angelo
>
> 2010/5/11 Daniel Dekany <ddekany@...>
>
> Tuesday, May 11, 2010, 2:26:16 PM, Jonathan Revusky wrote:
>>
>> > On Tue, May 11, 2010 at 2:11 AM, Daniel Dekany <ddekany@...>
>> wrote:
>> >> Tuesday, May 11, 2010, 1:24:11 AM, Angelo zerr wrote:
>> >>
>> >>> Hi Daniel,
>> >>>
>> >>> Ok, I have getted Freemarker 2.3 source from SVN and Token are not
>> visible -(
>> >>>
>> >>>>Committing changes into FreeMarker requires a signed paper CLA sent to
>> >>>>us, because corporate lawyers have problems with using FreeMarker
>> >>>>otherwise. I suggest that the same should stand for the sandbox. If
>> >>>>this plugin later becomes part of JBoss IDE or like, the unclear legal
>> >>>>status can become a problem.
>> >>>
>> >>> Ok Daniel, I don't have distrurb with you. You wish I stop commit
>> >>> my work?
>> >>
>> >> I think that would be wise. However, it's not up to me. There was an
>> >> agreement regarding the code of FreeMarker itself (that CLA is
>> >> required), but not regarding the sandbox. So let's see what the other
>> >> two admins say.
>> >
>> > I have no recollection of sending in the CLA. Did I? If not, I will, I
>> > was barely paying attention to that whole conversation.
>>
>> Well, I don't know how you bypassed this... but it doesn't mater
>> anyway.
>>
>> > Anyway, I don't know whether it should really be necessary to sign the
>> > CLA in order to commit code to /sandbox. Though Angelo might as well
>> > send in the CLA anyway...
>> >
>> >
>> >> Anyway, I think you better fill the CLA, and then at least you can
>> >> commit into FreeMarker as well. As of the stuff committed so far into
>> >> the sandbox... I'm not a lawyer, so I really don't know how to get out
>> >> of that now. But I guess we can't do more than deleting the current
>> >> commitment and then later when you have sent the CLA you just
>> >> re-commit it. /-: If that doesn't please the lawyers, well, too bad,
>> >> we won't lose a plugin because of that.
>> >
>> > I think this gets silly. Just leave the code there. He's expressed
>> > willingness to send in the CLA thing and he will. What difference does
>> > it make?
>>
>> For lawyers, I don't know... law is silly after all. But maybe it's
>> indeed useless. I'm just saying it's the most we can do now.
>>
>> > JR
>>
>> --
>> Best regards,
>>  Daniel Dekany
>>
>>
>>
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> FreeMarker-devel mailing list
>> FreeMarker-devel@...
>> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>>
>
>

Hi,

Just for information, I have commited my work on DLTK Freemarker and now you
can use Run As->Freemarker to launch your FTL template (like a Java class).
It display result of the merge into Eclipse Console. If you are interested
, plese read section "Run As → Freemarker" into DLTK-Freemarker - Eclipse
Plugin.odt<https://freemarker.svn.sourceforge.net/svnroot/freemarker/sandbox/org.eclipse.dltk.freemarker/docs/DLTK-Freemarker%20-%20Eclipse%20Plugin.odt>
You will see screenshot.

Now I would like study how we could debug FTL with DBGp (it seems it's a big
work).

Regards Angelo

Looks promising. It's also something that could help FreeMarker users
a lot when it's completed.

I notice you are still working against the FreeMarker trunk (which is
currently called FreeMarker 3.0). Just don't forget that that's a
trap, as I said earlier. You should work against 2.4 (branches/2.4). I
know that's a less IDE friendly version, but as far as backward
compatibility is kept (not evens strictly, but for the 99% of uses),
its API can be changed. Actually, it should be changed so it helps
tool developers, and since you are working on an actual plugin, you
know best what should be exposed and how.

Another minor thing... Any reason this plugin uses the
org.eclipse.dltk.freemarker package, which starts with
org.eclipse.dltk?


Friday, May 21, 2010, 2:12:10 PM, Angelo zerr wrote:

> Hi, 
>
> Just for information, I have commited my work on DLTK Freemarker
> and now you can use Run As->Freemarker to launch your FTL template
> (like a Java class). It display result of the merge into Eclipse
> Console. If you are interested  , plese read section "Run As →
> Freemarker" into DLTK-Freemarker - Eclipse Plugin.odt
> You will see screenshot.
>
> Now I would like study how we could debug FTL with DBGp (it seems it's a big work).
>
> Regards Angelo

-- 
Best regards,
 Daniel Dekany

Hi Daniel,

2010/5/24 Daniel Dekany <ddekany@...>

> Looks promising. It's also something that could help FreeMarker users
> a lot when it's completed.
>
That's cool, I hope I will finish it, because there is a lot of work.

>
> I notice you are still working against the FreeMarker trunk (which is
> currently called FreeMarker 3.0). Just don't forget that that's a
> trap, as I said earlier. You should work against 2.4 (branches/2.4). I
> know that's a less IDE friendly version, but as far as backward
> compatibility is kept (not evens strictly, but for the 99% of uses),
> its API can be changed. Actually, it should be changed so it helps
> tool developers, and since you are working on an actual plugin, you
> know best what should be exposed and how.
>

Yes I know that. Jonathan Revusky suggested me to use Freemarker trunk,
that's why I have used it. I'm using Freemarker FMParser to get the errors.
To show errors into Eclipse editor, we need :

* line error
* column error
* start/end char offset errors occured

It's very difficult to find that without Token which is private into
Freemarker 2.4. Token doesn't manage start/end char offset. But I have
noticed that Templatre into 2.4 wich use FMParser count the lines into
LineTableBuilder (I must study that). The problem is that LineTableBuilderi
s called into Template constructor. In my case I must create Template and
call the parse of FMParser to get the line (I must study that).

>
> Another minor thing... Any reason this plugin uses the
> org.eclipse.dltk.freemarker package, which starts with
> org.eclipse.dltk?
>

The reason is that I'm discovering DLTK and I'm using several examples with
DLTK (Javascript, Python, Ruby...)  and every DLTK implementation use this
convention org.eclipse.dltk.javascript.ui,
org.eclipse.dltk.javascript.debug.ui.... For me it help me to retrieve
classes, plugin.xml from another DLTK implementation.

Morever, perhaps DLTK will be interested with our Freemarker implementation?
Today I do DLTK Freemarker to learn DLTK, but I don't know if I will finish
it (if I'm alone to develop this plugin, I know it's difficult to keep the
motivation when we develop alone).

Anyway thank Daniel for your feedback.

Regards Angelo

>
>
> Friday, May 21, 2010, 2:12:10 PM, Angelo zerr wrote:
>
> > Hi,
> >
> > Just for information, I have commited my work on DLTK Freemarker
> > and now you can use Run As->Freemarker to launch your FTL template
> > (like a Java class). It display result of the merge into Eclipse
> > Console. If you are interested  , plese read section "Run As →
> > Freemarker" into DLTK-Freemarker - Eclipse Plugin.odt
> > You will see screenshot.
> >
> > Now I would like study how we could debug FTL with DBGp (it seems it's a
> big work).
> >
> > Regards Angelo
>
> --
> Best regards,
>  Daniel Dekany
>
>

Monday, May 24, 2010, 4:32:31 PM, Angelo zerr wrote:

[snip]
> I'm using Freemarker FMParser to get the errors. To show errors into
> Eclipse editor, we need :
>
> * line error
> * column error
> * start/end char offset errors occured
>  
> It's very difficult to find that without Token which is private into
> Freemarker 2.4. Token doesn't manage start/end char offset.

Does it not? I never fiddled with the parser of FM, but
freemarker.core.Token (and TemplateObject on the AST level) has:

  public int beginLine, beginColumn, endLine, endColumn;

Surely it's not public ATM, but it could be exposed if that's what is
needed. It's not necessarily a good idea to expose the Token class
directly though. It's maybe better to introduce some public API-s that
let you receive the stream of token data (with some kind of listener)
or traverse the AST. The advantage is that the internals of the
parsers can be changed without breaking the tools that need this kind
of information. But, again, I never dig into the parser either... So
far it was mostly the field of Jonathan.

> But I have noticed that Templatre into 2.4 wich use FMParser count
> the lines into LineTableBuilder (I must study that). The problem is
> that LineTableBuilderi s called into Template constructor. In my
> case I must create Template and call the parse of FMParser to get
> the line (I must study that).
>
> Another minor thing... Any reason this plugin uses the
> org.eclipse.dltk.freemarker package, which starts with
> org.eclipse.dltk?
>
> The reason is that I'm discovering DLTK and I'm using several
> examples with DLTK (Javascript, Python, Ruby...)  and every DLTK
> implementation use this convention org.eclipse.dltk.javascript.ui,
> org.eclipse.dltk.javascript.debug.ui.... For me it help me to
> retrieve classes, plugin.xml from another DLTK implementation.

Maybe the guys working on those are on some way part of the DLTK
project itself? Whatever...

> Morever, perhaps DLTK will be interested with our Freemarker
> implementation? Today I do DLTK Freemarker to learn DLTK, but I
> don't know if I will finish it (if I'm alone to develop this plugin,
> I know it's difficult to keep the motivation when we develop alone).
>
> Anyway thank Daniel for your feedback.
>
> Regards Angelo
>
>
>
> Friday, May 21, 2010, 2:12:10 PM, Angelo zerr wrote:
>
>> Hi,
>>
>> Just for information, I have commited my work on DLTK Freemarker
>> and now you can use Run As->Freemarker to launch your FTL template
>> (like a Java class). It display result of the merge into Eclipse
>> Console. If you are interested  , plese read section "Run As →
>> Freemarker" into DLTK-Freemarker - Eclipse Plugin.odt
>> You will see screenshot.
>>
>> Now I would like study how we could debug FTL with DBGp (it seems it's a big work).
>>
>> Regards Angelo
>
> --
> Best regards,
>  Daniel Dekany


-- 
Best regards,
 Daniel Dekany

On Mon, May 24, 2010 at 1:02 PM, Daniel Dekany <ddekany@...> wrote:
> Looks promising. It's also something that could help FreeMarker users
> a lot when it's completed.
>
> I notice you are still working against the FreeMarker trunk (which is
> currently called FreeMarker 3.0). Just don't forget that that's a
> trap, as I said earlier.

I'm pretty sure you're wrong about that. I told Angelo to go ahead and
develop against the trunk.

You seem to miss the basic bootstrap problem, which is that the only
way to get the trunk into a state where we are confident in it is to
actually USE it. If you won't use it because it is less tested, this
is just a vicious cycle sort of trap to fall into. That, and I also
think you overestimate just how bad the situation is with 3.0.

JR

> You should work against 2.4 (branches/2.4). I
> know that's a less IDE friendly version, but as far as backward
> compatibility is kept (not evens strictly, but for the 99% of uses),
> its API can be changed. Actually, it should be changed so it helps
> tool developers, and since you are working on an actual plugin, you
> know best what should be exposed and how.
>
> Another minor thing... Any reason this plugin uses the
> org.eclipse.dltk.freemarker package, which starts with
> org.eclipse.dltk?
>
>
> Friday, May 21, 2010, 2:12:10 PM, Angelo zerr wrote:
>
>> Hi,
>>
>> Just for information, I have commited my work on DLTK Freemarker
>> and now you can use Run As->Freemarker to launch your FTL template
>> (like a Java class). It display result of the merge into Eclipse
>> Console. If you are interested  , plese read section "Run As →
>> Freemarker" into DLTK-Freemarker - Eclipse Plugin.odt
>> You will see screenshot.
>>
>> Now I would like study how we could debug FTL with DBGp (it seems it's a big work).
>>
>> Regards Angelo
>
> --
> Best regards,
>  Daniel Dekany
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> FreeMarker-devel mailing list
> FreeMarker-devel@...
> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>

> It's very difficult to find that without Token which is private into
> > Freemarker 2.4. Token doesn't manage start/end char offset.
>
> Does it not? I never fiddled with the parser of FM, but
> freemarker.core.Token (and TemplateObject on the AST level) has:
>
>  public int beginLine, beginColumn, endLine, endColumn;
>
> Surely it's not public ATM, but it could be exposed if that's what is
> needed. It's not necessarily a good idea to expose the Token class
> directly though. It's maybe better to introduce some public API-s that
> let you receive the stream of token data (with some kind of listener)
> or traverse the AST. The advantage is that the internals of the
> parsers can be changed without breaking the tools that need this kind
> of information. But, again, I never dig into the parser either... So
> far it was mostly the field of Jonathan.
>

You are right, but my problem is to get start/end char. beginLine,
beginColumn, endLine, endColumn
can be used to compute start/end char BUT I must know the line number
(computed with '\n' and '\r' to compute start/end char. So I must when there
are erros, today I reparse the template to know line. IHMO I think it should
be very cool if  FMParser could manage that => start/end offset.

>
> >
> > The reason is that I'm discovering DLTK and I'm using several
> > examples with DLTK (Javascript, Python, Ruby...)  and every DLTK
> > implementation use this convention org.eclipse.dltk.javascript.ui,
> > org.eclipse.dltk.javascript.debug.ui.... For me it help me to
> > retrieve classes, plugin.xml from another DLTK implementation.
>
> Maybe the guys working on those are on some way part of the DLTK
> project itself? Whatever...
>
> For the moment, I prefer use this packages name because it's very helpfull
for me. Once pluginw will be stable we could change package if need. Are you
OK?

I'm pretty sure you're wrong about that. I told Angelo to go ahead and
> develop against the trunk.
>
> You seem to miss the basic bootstrap problem, which is that the only
> way to get the trunk into a state where we are confident in it is to
> actually USE it. If you won't use it because it is less tested, this
> is just a vicious cycle sort of trap to fall into. That, and I also
> think you overestimate just how bad the situation is with 3.0.
>

Today I don't use a lot of Freemarker API (like FMParser), because I'm
waiting your decision. Daniel/Jonathan could you take a decision about that
please.
I don't want change code of Freemarker, so I prefer that you change the
Freemarker API to manage start/end offset  Token.

Thank a lot for your help.

Regards Angelo


>
> JR
>
> > You should work against 2.4 (branches/2.4). I
> > know that's a less IDE friendly version, but as far as backward
> > compatibility is kept (not evens strictly, but for the 99% of uses),
> > its API can be changed. Actually, it should be changed so it helps
> > tool developers, and since you are working on an actual plugin, you
> > know best what should be exposed and how.
> >
> > Another minor thing... Any reason this plugin uses the
> > org.eclipse.dltk.freemarker package, which starts with
> > org.eclipse.dltk?
> >
> >
> > Friday, May 21, 2010, 2:12:10 PM, Angelo zerr wrote:
> >
> >> Hi,
> >>
> >> Just for information, I have commited my work on DLTK Freemarker
> >> and now you can use Run As->Freemarker to launch your FTL template
> >> (like a Java class). It display result of the merge into Eclipse
> >> Console. If you are interested  , plese read section "Run As →
> >> Freemarker" into DLTK-Freemarker - Eclipse Plugin.odt
> >> You will see screenshot.
> >>
> >> Now I would like study how we could debug FTL with DBGp (it seems it's a
> big work).
> >>
> >> Regards Angelo
> >
> > --
> > Best regards,
> >  Daniel Dekany
> >
> >
> >
> ------------------------------------------------------------------------------
> >
> > _______________________________________________
> > FreeMarker-devel mailing list
> > FreeMarker-devel@...
> > https://lists.sourceforge.net/lists/listinfo/freemarker-devel
> >
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> FreeMarker-devel mailing list
> FreeMarker-devel@...
> https://lists.sourceforge.net/lists/listinfo/freemarker-devel
>

Monday, May 24, 2010, 9:37:56 PM, Angelo zerr wrote:

[snip]
> You are right, but my problem is to get start/end char. beginLine, beginColumn, endLine, endColumn
> can be used to compute start/end char BUT I must know the line
> number (computed with '\n' and '\r' to compute start/end char. So I
> must when there are erros, today I reparse the template to know
> line.

You have to do that even when you are working against the trunk (3.0),
right?

> IHMO I think it should be very cool if  FMParser could manage that => start/end offset.

Since it already detects where the line begins, it should be easy...
unless JavaCC is on the way somehow.

[snip]
>> For the moment, I prefer use this packages name because it's very
>> helpfull for me. Once pluginw will be stable we could change
>> package if need. Are you OK?

I don't care that much about this detail... I just found it strange.

>>> I'm pretty sure you're wrong about that. I told Angelo to go ahead and
>>> develop against the trunk.
>>
>> You seem to miss the basic bootstrap problem, which is that the only
>> way to get the trunk into a state where we are confident in it is to
>> actually USE it. If you won't use it because it is less tested, this
>> is just a vicious cycle sort of trap to fall into. That, and I also
>> think you overestimate just how bad the situation is with 3.0.
>>
> Today I don't use a lot of Freemarker API (like FMParser), because
> I'm waiting your decision. Daniel/Jonathan could you take a decision
> about that please.

It would be easy to decide if we see into the future... my *guess* is
that 3.0 will not be released in the foreseeable future (or ever).
Jonathan's *guess* is the opposite. Anyway, there are 3 of us:
Jonathan, Attila and me. So, Jonathan votes for developing the plugin
against 3.0, and I vote for 2.4. So let's see what's the opinion of
Attila. (Note what a bastard I am: Now Attila's word will decide. ;)
Unless... he chooses to be silent. In that case I suppose it's
Jonathan who wins, since he did more work here than I did.)

>> I don't want change code of Freemarker, so I prefer that you change
>> the Freemarker API to manage start/end offset  Token.
>
> Thank a lot for your help.
>
> Regards Angelo
>  
>
> JR

-- 
Best regards,
 Daniel Dekany

It looks like you're mixing JSP and FreeMarker with JSP taglibs in the same project. This would ordinarily be fine, but it's failing for you... The reason it ordinarily works fine is that FreeMarker will only install FreeMarkerJspFactory as a fallback JspFactory, provided for taglibs in case there's no other JspFactory implementation in the servlet container. The initialization code in FreeMarkerPageContext is this:

    static {
        if(JspFactory.getDefaultFactory() == null) {
            JspFactory.setDefaultFactory(new FreeMarkerJspFactory21());
        }
        logger.debug("Using JspFactory implementation class " + 
                JspFactory.getDefaultFactory().getClass().getName());
    }

This means that FreeMarkerJspFactory won't even get installed if there's another JspFactory already installed. I wonder why is Jetty not installing its own JspFactory for JSP pages. Can you try triggering Jetty into installing its own JspFactory first? Like, ensuring a JSP page is rendered before FreeMarker JSP support is being used? Would that solve the problem? If yes, we can think of how to move forward.

FWIW, the definition for the getPageContext() and releasePageContext are indeed this:

    public PageContext getPageContext(Servlet servlet, ServletRequest request, 
            ServletResponse response, String errorPageURL, 
            boolean needsSession, int bufferSize, boolean autoFlush) {
        // This is never meant to be called. JSP pages compiled to Java 
        // bytecode use this API, but in FreeMarker, we're running templates,
        // and not JSP pages precompiled to bytecode, therefore we have no use
        // for this API.
        throw new UnsupportedOperationException();
    }

    public void releasePageContext(PageContext ctx) {
        // This is never meant to be called. JSP pages compiled to Java 
        // bytecode use this API, but in FreeMarker, we're running templates,
        // and not JSP pages precompiled to bytecode, therefore we have no use
        // for this API.
        throw new UnsupportedOperationException();
    }

Note the comments explaining the reason for UnsupportedOperationException() -- these are never meant to be called from templates. 
Now, what's *really* strange is that your precompiled JSP page named "org.apache.jsp.WEB_002dINF.decorators.default_002ddecorator_jsp._jspService" is not calling getPageContext() first and getting an exception on that...

Attila.

On 2010.05.10., at 11:37, sudhir543-dev@... wrote:

> Found one more problem with GAE + sitemesh..
> 
> I use sitemesh to decorate pages, the decorator is a jsp file..the pages which are being decorated are freemarker templates.. 
> it's fine because sitemash just care about the HTML.
> 
> The setup runs fine on local environment.. where as when deployed to GAE it gives me this error..
> 
> javax.servlet.ServletContext log: Unhandled exception occurred whilst decorating page
> java.lang.UnsupportedOperationException
> 	at freemarker.ext.jsp.FreeMarkerJspFactory.releasePageContext(FreeMarkerJspFactory.java:41)
> 	at org.apache.jsp.WEB_002dINF.decorators.default_002ddecorator_jsp._jspService(default_002ddecorator_jsp.java:99)
> 	at
>  org.apache.jasper.runtime.HttpJspBase.service(HttpJspBase.java:94)
> 	at javax.servlet.http.HttpServlet.service(HttpServlet.java:806)
> 	at org.mortbay.jetty.servlet.ServletHolder.handle(ServletHolder.java:511)
> 	at org.mortbay.jetty.servlet.ServletHandler.handle(ServletHandler.java:390)
> 	at org.mortbay.jetty.security.SecurityHandler.handle(SecurityHandler.java:216)
> 	at org.mortbay.jetty.servlet.SessionHandler.handle(SessionHandler.java:182)
> 	at org.mortbay.jetty.handler.ContextHandler.handle(ContextHandler.java:765)
> 	at org.mortbay.jetty.webapp.WebAppContext.handle(WebAppContext.java:418)
> 	at org.mortbay.jetty.servlet.Dispatcher.include(Dispatcher.java:192)
> 	at com.opensymphony.sitemesh.compatability.OldDecorator2NewDecorator.render(OldDecorator2NewDecorator.java:46)
> 	at com.opensymphony.sitemesh.webapp.decorator.BaseWebAppDecorator.render(BaseWebAppDecorator.java:33)
> 	at
>  com.opensymphony.sitemesh.webapp.SiteMeshFilter.doFilter(SiteMeshFilter.java:84)
> 	at org.mortbay.jetty.servlet.ServletHandler$CachedChain.doFilter(ServletHandler.java:1157)
> 
>  Seems that app engine is doing some thing strange with FreeMarkerJspFactory or reusing it in someway which isn't expected.
>    
> Sudhir Nimavat
> Senior software engineer. 
> Quick start global PVT LTD.
> Baroda - 390007
> Gujarat, India
> http://www.ramanandi.org

Yeah, tried and it works.. I created a hello.jsp. After deploying the application, accessed hello.jsp before calling any other URL that would render ftl.
And yes, after that everything else works.. 

What could be the permanent solution of this ? May be this is a bug with jetty implementation on GAE.

 
    
Sudhir NimavatSenior software engineer. 
Quick start global PVT LTD.
Baroda - 390007
Gujarat, India
http://www.ramanandi.org
Warning: Spelling errors in this message are the product of a poor school system. Pay teachers more than  athletes..
  




________________________________
From: Attila Szegedi <szegedia@...>
To: FreeMarker-devel <freemarker-devel@...>
Cc: freemarker <freemarker-user@...>
Sent: Mon, 10 May, 2010 3:46:42 PM
Subject: Re: [FreeMarker-user] [Freemarker-devel] One more problem with GAE + Sitemesh

It looks like you're mixing JSP and FreeMarker with JSP taglibs in the same project. This would ordinarily be fine, but it's failing for you... The reason it ordinarily works fine is that FreeMarker will only install FreeMarkerJspFactory as a fallback JspFactory, provided for taglibs in case there's no other JspFactory implementation in the servlet container. The initialization code in FreeMarkerPageContext is this:


    static {
        if(JspFactory.getDefaultFactory() == null) {
            JspFactory.setDefaultFactory(new FreeMarkerJspFactory21());
        }
        logger.debug("Using JspFactory implementation class " + 
                JspFactory.getDefaultFactory().getClass().getName());
    }

This means that FreeMarkerJspFactory won't even get installed if there's another JspFactory already installed. I wonder why is Jetty not installing its own JspFactory for JSP pages. Can you try triggering Jetty into installing its own JspFactory first? Like, ensuring a JSP page is rendered before FreeMarker JSP support is being used? Would that solve the problem? If yes, we can think of how to move forward.

FWIW, the definition for the getPageContext() and releasePageContext are indeed this:

    public PageContext getPageContext(Servlet servlet, ServletRequest request, 
            ServletResponse response, String errorPageURL, 
            boolean needsSession, int bufferSize, boolean autoFlush) {
        // This is never meant to be called. JSP pages compiled to Java 
        // bytecode use this API, but in FreeMarker, we're running templates,
        // and not JSP pages precompiled to bytecode, therefore we have no use
        // for this API.
        throw new UnsupportedOperationException();
    }

    public void releasePageContext(PageContext ctx) {
        // This is never meant to be called. JSP pages compiled to Java 
        // bytecode use this API, but in FreeMarker, we're running templates,
        // and not JSP pages precompiled to bytecode, therefore we have no use
        // for this API.
        throw new UnsupportedOperationException();
    }

Note the comments explaining the reason for UnsupportedOperationException() -- these are never meant to be called from templates. 
Now, what's *really* strange is that your precompiled JSP page named "org.apache.jsp.WEB_002dINF.decorators.default_002ddecorator_jsp._jspService" is not calling getPageContext() first and getting an exception on that...

Attila.

On 2010.05.10., at 11:37, sudhir543-dev@... wrote:

Found one more problem with GAE + sitemesh..
>
>I use sitemesh to decorate pages, the decorator is a jsp file..the pages which are being decorated are freemarker templates.. 
>it's fine because sitemash just care about the HTML.
>
>The setup runs fine on local environment.. where as when deployed to GAE it gives me this error..
>
>
>javax.servlet.ServletContext log: Unhandled exception occurred whilst decorating page
>java.lang.UnsupportedOperationException
>	at freemarker.ext.jsp.FreeMarkerJspFactory.releasePageContext(FreeMarkerJspFactory.java:41)
>	at org.apache.jsp.WEB_002dINF.decorators.default_002ddecorator_jsp._jspService(default_002ddecorator_jsp.java:99)
>	at
> org.apache.jasper.runtime.HttpJspBase.service(HttpJspBase.java:94)
>	at javax.servlet.http.HttpServlet.service(HttpServlet.java:806)
>	at org.mortbay.jetty.servlet.ServletHolder.handle(ServletHolder.java:511)
>	at org.mortbay.jetty.servlet.ServletHandler.handle(ServletHandler.java:390)
>	at org.mortbay.jetty.security.SecurityHandler.handle(SecurityHandler.java:216)
>	at org.mortbay.jetty.servlet.SessionHandler.handle(SessionHandler.java:182)
>	at org.mortbay.jetty.handler.ContextHandler.handle(ContextHandler.java:765)
>	at org.mortbay.jetty.webapp.WebAppContext.handle(WebAppContext.java:418)
>	at org.mortbay.jetty.servlet.Dispatcher.include(Dispatcher.java:192)
>	at com.opensymphony.sitemesh.compatability.OldDecorator2NewDecorator.render(OldDecorator2NewDecorator.java:46)
>	at com.opensymphony.sitemesh.webapp.decorator.BaseWebAppDecorator.render(BaseWebAppDecorator.java:33)
>	at
> com.opensymphony.sitemesh.webapp.SiteMeshFilter.doFilter(SiteMeshFilter.java:84)
>	at org.mortbay.jetty.servlet.ServletHandler$CachedChain.doFilter(ServletHandler.java:1157)
>
>
> Seems that app engine is doing some thing strange with FreeMarkerJspFactory or reusing it in someway which isn't expected.
>
>   
>Sudhir NimavatSenior software engineer. 
>Quick start global PVT LTD.
>Baroda - 390007
>Gujarat, India
>http://www.ramanandi.org

Any comments on permanent solution !

 
    
Sudhir NimavatSenior software engineer. 
Quick start global PVT LTD.
Baroda - 390007
Gujarat, India
http://www.ramanandi.org
Warning: Spelling errors in this message are the product of a poor school system. Pay teachers more than  athletes..
  




________________________________
From: "sudhir543-dev@..." <sudhir543-dev@...>
To: FreeMarker-user <freemarker-user@...>
Cc: freemarker devel <freemarker-devel@...>
Sent: Mon, 10 May, 2010 4:44:39 PM
Subject: Re: [Freemarker-devel] One more problem with GAE + Sitemesh


Yeah, tried and it works.. I created a hello.jsp. After deploying the application, accessed hello.jsp before calling any other URL that would render ftl.
And yes, after that everything else works.. 

What could be the permanent solution of this ? May be this is a bug with jetty implementation on GAE.

 
    
Sudhir NimavatSenior software engineer. 
Quick start global PVT LTD.
Baroda - 390007
Gujarat, India
http://www.ramanandi.org
Warning: Spelling errors in this message are the product of a poor school system. Pay teachers more than  athletes..
  




________________________________
From: Attila Szegedi <szegedia@...>
To: FreeMarker-devel <freemarker-devel@...>
Cc: freemarker <freemarker-user@...>
Sent: Mon, 10 May, 2010 3:46:42 PM
Subject: Re: [FreeMarker-user] [Freemarker-devel] One more problem with GAE + Sitemesh

It looks like you're mixing JSP and FreeMarker with JSP taglibs in the same project. This would ordinarily be fine, but it's failing for you... The reason it ordinarily works fine is that FreeMarker will only install FreeMarkerJspFactory as a fallback JspFactory, provided for taglibs in case there's no other JspFactory implementation in the servlet container. The initialization code in FreeMarkerPageContext is this:


    static {
        if(JspFactory.getDefaultFactory() == null) {
            JspFactory.setDefaultFactory(new FreeMarkerJspFactory21());
        }
        logger.debug("Using JspFactory implementation class " + 
                JspFactory.getDefaultFactory().getClass().getName());
    }

This means that FreeMarkerJspFactory won't even get installed if there's another JspFactory already installed. I wonder why is Jetty not installing its own JspFactory for JSP pages. Can you try triggering Jetty into installing its own JspFactory first? Like, ensuring a JSP page is rendered before FreeMarker JSP support is being used? Would that solve the problem? If yes, we can think of how to move forward.

FWIW, the definition for the getPageContext() and releasePageContext are indeed this:

    public PageContext getPageContext(Servlet servlet, ServletRequest request, 
            ServletResponse response, String errorPageURL, 
            boolean needsSession, int bufferSize, boolean autoFlush) {
        // This is never meant to be called. JSP pages compiled to Java 
        // bytecode use this API, but in FreeMarker, we're running templates,
        // and not JSP pages precompiled to bytecode, therefore we have no use
        // for this API.
        throw new UnsupportedOperationException();
    }

    public void releasePageContext(PageContext ctx) {
        // This is never meant to be called. JSP pages compiled to Java 
        // bytecode use this API, but in FreeMarker, we're running templates,
        // and not JSP pages precompiled to bytecode, therefore we have no use
        // for this API.
        throw new UnsupportedOperationException();
    }

Note the comments explaining the reason for UnsupportedOperationException() -- these are never meant to be called from templates. 
Now, what's *really* strange is that your precompiled JSP page named "org.apache.jsp.WEB_002dINF.decorators.default_002ddecorator_jsp._jspService" is not calling getPageContext() first and getting an exception on that...

Attila.

On 2010.05.10., at 11:37, sudhir543-dev@... wrote:

Found one more problem with GAE + sitemesh..
>
>I use sitemesh to decorate pages, the
> decorator is a jsp file..the pages which are being decorated are freemarker templates.. 
>it's fine because sitemash just care about the HTML.
>
>The setup runs fine on local environment.. where as when deployed to GAE it gives me this error..
>
>
>javax.servlet.ServletContext log: Unhandled exception occurred whilst decorating page
>java.lang.UnsupportedOperationException
>	at freemarker.ext.jsp.FreeMarkerJspFactory.releasePageContext(FreeMarkerJspFactory.java:41)
>	at org.apache.jsp.WEB_002dINF.decorators.default_002ddecorator_jsp._jspService(default_002ddecorator_jsp.java:99)
>	at
> org.apache.jasper.runtime.HttpJspBase.service(HttpJspBase.java:94)
>	at javax.servlet.http.HttpServlet.service(HttpServlet.java:806)
>	at org.mortbay.jetty.servlet.ServletHolder.handle(ServletHolder.java:511)
>	at
> org.mortbay.jetty.servlet.ServletHandler.handle(ServletHandler.java:390)
>	at org.mortbay.jetty.security.SecurityHandler.handle(SecurityHandler.java:216)
>	at org.mortbay.jetty.servlet.SessionHandler.handle(SessionHandler.java:182)
>	at org.mortbay.jetty.handler.ContextHandler.handle(ContextHandler.java:765)
>	at org.mortbay.jetty.webapp.WebAppContext.handle(WebAppContext.java:418)
>	at org.mortbay.jetty.servlet.Dispatcher.include(Dispatcher.java:192)
>	at com.opensymphony.sitemesh.compatability.OldDecorator2NewDecorator.render(OldDecorator2NewDecorator.java:46)
>	at com.opensymphony.sitemesh.webapp.decorator.BaseWebAppDecorator.render(BaseWebAppDecorator.java:33)
>	at
> com.opensymphony.sitemesh.webapp.SiteMeshFilter.doFilter(SiteMeshFilter.java:84)
>	at org.mortbay.jetty.servlet.ServletHandler$CachedChain.doFilter(ServletHandler.java:1157)
>
>
> Seems that app engine is doing some thing
> strange with FreeMarkerJspFactory or reusing it in someway which isn't expected.
>
>   
>Sudhir NimavatSenior software engineer. 
>Quick start global PVT LTD.
>Baroda - 390007
>Gujarat, India
>http://www.ramanandi.org

Filed a bug on gae 
http://code.google.com/p/googleappengine/issues/detail?id=3252

Attila, do you have any comments for solution from our side ?

 
    
Sudhir NimavatSenior software engineer. 
Quick start global PVT LTD.
Baroda - 390007
Gujarat, India
http://www.ramanandi.org
Warning: Spelling errors in this message are the product of a poor school system. Pay teachers more than  athletes..
  




________________________________
From: "sudhir543-dev@..." <sudhir543-dev@...>
To: FreeMarker-devel <freemarker-devel@...>
Cc: freemarker <freemarker-user@...>
Sent: Tue, 11 May, 2010 8:05:10 PM
Subject: [Freemarker-devel] RE One more problem with GAE + Sitemesh


Any comments on permanent solution !

 
    
Sudhir NimavatSenior software engineer. 
Quick start global PVT LTD.
Baroda - 390007
Gujarat, India
http://www.ramanandi.org
Warning: Spelling errors in this message are the product of a poor school system. Pay teachers more than  athletes..
  




________________________________
From: "sudhir543-dev@..." <sudhir543-dev@...>
To: FreeMarker-user <freemarker-user@...>
Cc: freemarker devel <freemarker-devel@...>
Sent: Mon, 10 May, 2010 4:44:39 PM
Subject: Re: [Freemarker-devel] One more problem with GAE + Sitemesh


Yeah, tried and it works.. I created a hello.jsp. After deploying the application, accessed hello.jsp before calling any other URL that would render ftl.
And yes, after that everything else works.. 

What could be the permanent solution of this ? May be this is a bug with jetty implementation on GAE.

 
    
Sudhir NimavatSenior software engineer. 
Quick start global PVT LTD.
Baroda - 390007
Gujarat, India
http://www.ramanandi.org
Warning: Spelling errors in this message are the product of a poor school system. Pay teachers more than  athletes..
  




________________________________
From: Attila Szegedi <szegedia@...>
To: FreeMarker-devel <freemarker-devel@...>
Cc: freemarker <freemarker-user@...>
Sent: Mon, 10 May, 2010 3:46:42 PM
Subject: Re: [FreeMarker-user] [Freemarker-devel] One more problem with GAE + Sitemesh

It looks like you're mixing JSP and FreeMarker with JSP taglibs in the same project. This would ordinarily be fine, but it's failing for you... The reason it ordinarily works fine is that FreeMarker will only install FreeMarkerJspFactory as a fallback JspFactory, provided for taglibs in case there's no other JspFactory implementation in the servlet container. The initialization code in FreeMarkerPageContext is this:


    static {
        if(JspFactory.getDefaultFactory() == null) {
            JspFactory.setDefaultFactory(new FreeMarkerJspFactory21());
        }
        logger.debug("Using JspFactory implementation class " + 
                JspFactory.getDefaultFactory().getClass().getName());
    }

This means that FreeMarkerJspFactory won't even get installed if there's another JspFactory already installed. I wonder why is Jetty not installing its own JspFactory for JSP pages. Can you try triggering Jetty into installing its own JspFactory first? Like, ensuring a JSP page is rendered before FreeMarker JSP support is being used? Would that solve the problem? If yes, we can think of how to move forward.

FWIW, the definition for the getPageContext() and releasePageContext are indeed this:

    public PageContext getPageContext(Servlet servlet, ServletRequest request, 
            ServletResponse response, String errorPageURL, 
            boolean needsSession, int bufferSize, boolean autoFlush) {
        // This is never meant to be called. JSP pages compiled to Java 
        // bytecode use this API, but in FreeMarker, we're running templates,
        // and not JSP pages precompiled to bytecode, therefore we have no use
        // for this API.
        throw new UnsupportedOperationException();
    }

    public void releasePageContext(PageContext ctx) {
        // This is never meant to be called. JSP pages compiled to Java 
        // bytecode use this API, but in FreeMarker, we're running templates,
        // and not JSP pages precompiled to bytecode, therefore we have no use
        // for this API.
        throw new UnsupportedOperationException();
    }

Note the comments explaining the reason for UnsupportedOperationException() -- these are never meant to be called from templates. 
Now, what's *really* strange is that your precompiled JSP page named "org.apache.jsp.WEB_002dINF.decorators.default_002ddecorator_jsp._jspService" is not calling getPageContext() first and getting an exception on that...

Attila.

On 2010.05.10., at 11:37, sudhir543-dev@... wrote:

Found one more problem with GAE + sitemesh..
>
>I use sitemesh to decorate pages, the
> decorator is a jsp file..the pages which are being decorated are freemarker templates.. 
>it's fine because sitemash just care about the HTML.
>
>The setup runs fine on local environment.. where as when deployed to GAE it gives me this error..
>
>
>javax.servlet.ServletContext log: Unhandled exception occurred whilst decorating page
>java.lang.UnsupportedOperationException
>	at freemarker.ext.jsp.FreeMarkerJspFactory.releasePageContext(FreeMarkerJspFactory.java:41)
>	at org.apache.jsp.WEB_002dINF.decorators.default_002ddecorator_jsp._jspService(default_002ddecorator_jsp.java:99)
>	at
> org.apache.jasper.runtime.HttpJspBase.service(HttpJspBase.java:94)
>	at javax.servlet.http.HttpServlet.service(HttpServlet.java:806)
>	at org.mortbay.jetty.servlet.ServletHolder.handle(ServletHolder.java:511)
>	at
>
> org.mortbay.jetty.servlet.ServletHandler.handle(ServletHandler.java:390)
>	at org.mortbay.jetty.security.SecurityHandler.handle(SecurityHandler.java:216)
>	at org.mortbay.jetty.servlet.SessionHandler.handle(SessionHandler.java:182)
>	at org.mortbay.jetty.handler.ContextHandler.handle(ContextHandler.java:765)
>	at org.mortbay.jetty.webapp.WebAppContext.handle(WebAppContext.java:418)
>	at org.mortbay.jetty.servlet.Dispatcher.include(Dispatcher.java:192)
>	at com.opensymphony.sitemesh.compatability.OldDecorator2NewDecorator.render(OldDecorator2NewDecorator.java:46)
>	at com.opensymphony.sitemesh.webapp.decorator.BaseWebAppDecorator.render(BaseWebAppDecorator.java:33)
>	at
> com.opensymphony.sitemesh.webapp.SiteMeshFilter.doFilter(SiteMeshFilter.java:84)
>	at org.mortbay.jetty.servlet.ServletHandler$CachedChain.doFilter(ServletHandler.java:1157)
>
>
> Seems that app engine is doing some thing
> strange with FreeMarkerJspFactory or reusing it in someway which isn't expected.
>
>   
>Sudhir NimavatSenior software engineer. 
>Quick start global PVT LTD.
>Baroda - 390007
>Gujarat, India
>http://www.ramanandi.org