Re: (FREEMARKER-55) FM3 freemarker-spring module, Web MVC support

[Daniel Dekany]

Tuesday, August 29, 2017, 8:48:34 PM, Woonsan Ko wrote:

> On Tue, Aug 29, 2017 at 10:40 AM, Daniel Dekany  wrote:
>> Tuesday, August 29, 2017, 3:03:02 PM, Woonsan Ko wrote:
>>
>>> On Tue, Aug 29, 2017 at 5:00 AM, Daniel Dekany  wrote:
 Tuesday, August 29, 2017, 5:25:31 AM, Woonsan Ko wrote:

> Hi Daniel,
>
> Thanks for the remarks and hints! Please see my question inline.
>
> On Sat, Aug 26, 2017 at 4:50 AM, Daniel Dekany  
> wrote:
>> Saturday, August 26, 2017, 10:44:03 AM, Daniel Dekany wrote:
>>
>> [snip]
>>>   <#macro foo>
>>> <#local status = 'blah'>
>>> <@spring.bind "user.name">
>>>   ${status.value} <#-- Means 'blah'.value, won't work -->
>>
>> To clarify, here I have assumed that spring.bind calls
>> env.setVariable("status", ...) internally. Then the local with the
>> same name shadows that tempolate-namespace scoped variable.
>
> It seems working with env.setVariable("status", ...):
>
> // Excerpt from
> https://github.com/woonsan/incubator-freemarker/blob/feature/FREEMARKER-55-2/freemarker-spring/src/main/java/org/apache/freemarker/spring/model/BindDirective.java

 Looks good. But what are your thoughts on using a nested content
 parameter instead?
>>>
>>> I'm sure I'm missing something or in misunderstandings. I want to try
>>> the option using a nested content parameter instead, but I don't know
>>> how to do that. ;-)
>>> Ideally, <@spring.bind ... /> directive needs to set a local variable
>>> called "status" or whatever (if the variable name can be given by a
>>> nested content parameter) and developers should be able to use the
>>> status variable ("status" or whatever) somehow inside the directive.
>>> So, my assumption was, if the parameter is given to the directive, I
>>> can probably fill in a variable somehow by resolving the bind status
>>> object. I guess I'm totally ignorant on how nested content parameters
>>> work with directives...
>>
>> A directive can't create a nested content variable. The caller can
>> declare a nested content parameter, like `<@spring.bind ...; status>`,
>> and then the directive can specify the actual value of that parameter
>> when it calls the nested content with
>> `callPlace.executeNestedContent`. (Of course it's likely that the
>> caller will chose a shorter name, like `s`; see earlier mails.)
>
> Ah, I think I now understand it!
> So, for example, if I do this:
>
>   callPlace.executeNestedContent(new TemplateModel [] { status }, out, env);
>
> , then the nested template may read/use the passed status object by
> referring to by the name (e.g, 's') given by themselves, right?

Right.

> I totally overlooked the fact that I can pass arguments to the nested
> content when invoking #executeNestedContent().
> I'll try it out again. I personally it is better to use nested content
> parameters instead of #setVariable() as they are local.
> The following example seems a lot better than the JSP taglib to me:
>
> <@spring.bind "user.email"; status>
>   
>  Thanks again for the advice!
>
> Cheers,
>
> Woonsan
>
>>
>> You may wonder if we should allow the directive to create a nested
>> content variable. I think it's a probably a quite bad library design
>> if you need that. Creating variables without the caller seeing that
>> makes understanding what's going on significantly harder. Worse, that
>> magically appearing variable can hide a variable from a higher scope.
>> If the caller doesn't ask for the creation of the variable, then that
>> variable should be under the `spring` namespace. Of course that
>> (`spring.status`) would be more verbose than using a nested content
>> parameter, so it doesn't make sense in this case.
>>
>>> Regards,
>>>
>>> Woonsan
>>>

> Another question: is it okay to assume that I have a
> DefaultObjectWrapper when this directive is executed and throw an
> exception otherwise?

 At the moment it's enough to expect it to be an
 ObjectWrapperAndUnwrapper. FreeMarkerServlet also requires that BTW.
 I'm not sure if in the future expecting DefaultObjectWrapper will be
 necessary, but hopefully not. (Custom JSP tag support expects a
 DefaultObjectWrapper at the moment, for wrapping JSP functions.)

>> [snip]
>>> With nested content parameter it's also more obvious what's going on,
>>> and if you chose a shorter name it's not that verbose either:
>>>
>>>   <@spring.bind "user.name"; s>
>>> ${s.value}
>
> If I choose to use nested content parameter instead of having
> env.setVariable() calls, how can I get access to the parameter, s?
> Can I access it through any argument of the #execute() method in the
> Java code? Any example java code using nested content parameter?

 I'm not sure what you try to achieve. As the nested content parameter
 is local to 

Re: (FREEMARKER-55) FM3 freemarker-spring module, Web MVC support

[Woonsan Ko]

On Tue, Aug 29, 2017 at 2:48 PM, Woonsan Ko  wrote:
> On Tue, Aug 29, 2017 at 10:40 AM, Daniel Dekany  wrote:
>> Tuesday, August 29, 2017, 3:03:02 PM, Woonsan Ko wrote:
>>
>>> On Tue, Aug 29, 2017 at 5:00 AM, Daniel Dekany  wrote:
 Tuesday, August 29, 2017, 5:25:31 AM, Woonsan Ko wrote:

> Hi Daniel,
>
> Thanks for the remarks and hints! Please see my question inline.
>
> On Sat, Aug 26, 2017 at 4:50 AM, Daniel Dekany  
> wrote:
>> Saturday, August 26, 2017, 10:44:03 AM, Daniel Dekany wrote:
>>
>> [snip]
>>>   <#macro foo>
>>> <#local status = 'blah'>
>>> <@spring.bind "user.name">
>>>   ${status.value} <#-- Means 'blah'.value, won't work -->
>>
>> To clarify, here I have assumed that spring.bind calls
>> env.setVariable("status", ...) internally. Then the local with the
>> same name shadows that tempolate-namespace scoped variable.
>
> It seems working with env.setVariable("status", ...):
>
> // Excerpt from
> https://github.com/woonsan/incubator-freemarker/blob/feature/FREEMARKER-55-2/freemarker-spring/src/main/java/org/apache/freemarker/spring/model/BindDirective.java

 Looks good. But what are your thoughts on using a nested content
 parameter instead?
>>>
>>> I'm sure I'm missing something or in misunderstandings. I want to try
>>> the option using a nested content parameter instead, but I don't know
>>> how to do that. ;-)
>>> Ideally, <@spring.bind ... /> directive needs to set a local variable
>>> called "status" or whatever (if the variable name can be given by a
>>> nested content parameter) and developers should be able to use the
>>> status variable ("status" or whatever) somehow inside the directive.
>>> So, my assumption was, if the parameter is given to the directive, I
>>> can probably fill in a variable somehow by resolving the bind status
>>> object. I guess I'm totally ignorant on how nested content parameters
>>> work with directives...
>>
>> A directive can't create a nested content variable. The caller can
>> declare a nested content parameter, like `<@spring.bind ...; status>`,
>> and then the directive can specify the actual value of that parameter
>> when it calls the nested content with
>> `callPlace.executeNestedContent`. (Of course it's likely that the
>> caller will chose a shorter name, like `s`; see earlier mails.)
>
> Ah, I think I now understand it!
> So, for example, if I do this:
>
>   callPlace.executeNestedContent(new TemplateModel [] { status }, out, env);
>
> , then the nested template may read/use the passed status object by
> referring to by the name (e.g, 's') given by themselves, right?
>
> I totally overlooked the fact that I can pass arguments to the nested
> content when invoking #executeNestedContent().
> I'll try it out again. I personally it is better to use nested content
> parameters instead of #setVariable() as they are local.
> The following example seems a lot better than the JSP taglib to me:
>
> <@spring.bind "user.email"; status>
>   
> https://github.com/woonsan/incubator-freemarker/commit/b656f1e1f7c1b0875526f2224db1fc793e1172d2
- 
https://github.com/woonsan/spring-mvc-freemarker3-demo/commit/e2530f62e1d8b1266f9b426dd3779bc26b1283bd
  (I'll move this temporary demo project to unit tests later when making PRs.)

Regards,

Woonsan

>
> Thanks again for the advice!
>
> Cheers,
>
> Woonsan
>
>>
>> You may wonder if we should allow the directive to create a nested
>> content variable. I think it's a probably a quite bad library design
>> if you need that. Creating variables without the caller seeing that
>> makes understanding what's going on significantly harder. Worse, that
>> magically appearing variable can hide a variable from a higher scope.
>> If the caller doesn't ask for the creation of the variable, then that
>> variable should be under the `spring` namespace. Of course that
>> (`spring.status`) would be more verbose than using a nested content
>> parameter, so it doesn't make sense in this case.
>>
>>> Regards,
>>>
>>> Woonsan
>>>

> Another question: is it okay to assume that I have a
> DefaultObjectWrapper when this directive is executed and throw an
> exception otherwise?

 At the moment it's enough to expect it to be an
 ObjectWrapperAndUnwrapper. FreeMarkerServlet also requires that BTW.
 I'm not sure if in the future expecting DefaultObjectWrapper will be
 necessary, but hopefully not. (Custom JSP tag support expects a
 DefaultObjectWrapper at the moment, for wrapping JSP functions.)

>> [snip]
>>> With nested content parameter it's also more obvious what's going on,
>>> and if you chose a shorter name it's not that verbose either:
>>>
>>>   <@spring.bind "user.name"; s>
>>> ${s.value}
>
> If I choose to use nested content parameter instead of having
> 

Re: (FREEMARKER-55) FM3 freemarker-spring module, Web MVC support

[Woonsan Ko]

On Tue, Aug 29, 2017 at 10:40 AM, Daniel Dekany  wrote:
> Tuesday, August 29, 2017, 3:03:02 PM, Woonsan Ko wrote:
>
>> On Tue, Aug 29, 2017 at 5:00 AM, Daniel Dekany  wrote:
>>> Tuesday, August 29, 2017, 5:25:31 AM, Woonsan Ko wrote:
>>>
 Hi Daniel,

 Thanks for the remarks and hints! Please see my question inline.

 On Sat, Aug 26, 2017 at 4:50 AM, Daniel Dekany  wrote:
> Saturday, August 26, 2017, 10:44:03 AM, Daniel Dekany wrote:
>
> [snip]
>>   <#macro foo>
>> <#local status = 'blah'>
>> <@spring.bind "user.name">
>>   ${status.value} <#-- Means 'blah'.value, won't work -->
>
> To clarify, here I have assumed that spring.bind calls
> env.setVariable("status", ...) internally. Then the local with the
> same name shadows that tempolate-namespace scoped variable.

 It seems working with env.setVariable("status", ...):

 // Excerpt from
 https://github.com/woonsan/incubator-freemarker/blob/feature/FREEMARKER-55-2/freemarker-spring/src/main/java/org/apache/freemarker/spring/model/BindDirective.java
>>>
>>> Looks good. But what are your thoughts on using a nested content
>>> parameter instead?
>>
>> I'm sure I'm missing something or in misunderstandings. I want to try
>> the option using a nested content parameter instead, but I don't know
>> how to do that. ;-)
>> Ideally, <@spring.bind ... /> directive needs to set a local variable
>> called "status" or whatever (if the variable name can be given by a
>> nested content parameter) and developers should be able to use the
>> status variable ("status" or whatever) somehow inside the directive.
>> So, my assumption was, if the parameter is given to the directive, I
>> can probably fill in a variable somehow by resolving the bind status
>> object. I guess I'm totally ignorant on how nested content parameters
>> work with directives...
>
> A directive can't create a nested content variable. The caller can
> declare a nested content parameter, like `<@spring.bind ...; status>`,
> and then the directive can specify the actual value of that parameter
> when it calls the nested content with
> `callPlace.executeNestedContent`. (Of course it's likely that the
> caller will chose a shorter name, like `s`; see earlier mails.)

Ah, I think I now understand it!
So, for example, if I do this:

  callPlace.executeNestedContent(new TemplateModel [] { status }, out, env);

, then the nested template may read/use the passed status object by
referring to by the name (e.g, 's') given by themselves, right?

I totally overlooked the fact that I can pass arguments to the nested
content when invoking #executeNestedContent().
I'll try it out again. I personally it is better to use nested content
parameters instead of #setVariable() as they are local.
The following example seems a lot better than the JSP taglib to me:

<@spring.bind "user.email"; status>
  

> You may wonder if we should allow the directive to create a nested
> content variable. I think it's a probably a quite bad library design
> if you need that. Creating variables without the caller seeing that
> makes understanding what's going on significantly harder. Worse, that
> magically appearing variable can hide a variable from a higher scope.
> If the caller doesn't ask for the creation of the variable, then that
> variable should be under the `spring` namespace. Of course that
> (`spring.status`) would be more verbose than using a nested content
> parameter, so it doesn't make sense in this case.
>
>> Regards,
>>
>> Woonsan
>>
>>>
 Another question: is it okay to assume that I have a
 DefaultObjectWrapper when this directive is executed and throw an
 exception otherwise?
>>>
>>> At the moment it's enough to expect it to be an
>>> ObjectWrapperAndUnwrapper. FreeMarkerServlet also requires that BTW.
>>> I'm not sure if in the future expecting DefaultObjectWrapper will be
>>> necessary, but hopefully not. (Custom JSP tag support expects a
>>> DefaultObjectWrapper at the moment, for wrapping JSP functions.)
>>>
> [snip]
>> With nested content parameter it's also more obvious what's going on,
>> and if you chose a shorter name it's not that verbose either:
>>
>>   <@spring.bind "user.name"; s>
>> ${s.value}

 If I choose to use nested content parameter instead of having
 env.setVariable() calls, how can I get access to the parameter, s?
 Can I access it through any argument of the #execute() method in the
 Java code? Any example java code using nested content parameter?
>>>
>>> I'm not sure what you try to achieve. As the nested content parameter
>>> is local to the nested content, similarly as a local variable is local
>>> to the method body in Java, where do you want to access `s` from, and
>>> why?
>>>
 Thanks in advance,

 Woonsan

> [snip]
>
> Eh... of course the comment is 

Re: (FREEMARKER-55) FM3 freemarker-spring module, Web MVC support

[Daniel Dekany]

Tuesday, August 29, 2017, 3:03:02 PM, Woonsan Ko wrote:

> On Tue, Aug 29, 2017 at 5:00 AM, Daniel Dekany  wrote:
>> Tuesday, August 29, 2017, 5:25:31 AM, Woonsan Ko wrote:
>>
>>> Hi Daniel,
>>>
>>> Thanks for the remarks and hints! Please see my question inline.
>>>
>>> On Sat, Aug 26, 2017 at 4:50 AM, Daniel Dekany  wrote:
 Saturday, August 26, 2017, 10:44:03 AM, Daniel Dekany wrote:

 [snip]
>   <#macro foo>
> <#local status = 'blah'>
> <@spring.bind "user.name">
>   ${status.value} <#-- Means 'blah'.value, won't work -->

 To clarify, here I have assumed that spring.bind calls
 env.setVariable("status", ...) internally. Then the local with the
 same name shadows that tempolate-namespace scoped variable.
>>>
>>> It seems working with env.setVariable("status", ...):
>>>
>>> // Excerpt from
>>> https://github.com/woonsan/incubator-freemarker/blob/feature/FREEMARKER-55-2/freemarker-spring/src/main/java/org/apache/freemarker/spring/model/BindDirective.java
>>
>> Looks good. But what are your thoughts on using a nested content
>> parameter instead?
>
> I'm sure I'm missing something or in misunderstandings. I want to try
> the option using a nested content parameter instead, but I don't know
> how to do that. ;-)
> Ideally, <@spring.bind ... /> directive needs to set a local variable
> called "status" or whatever (if the variable name can be given by a
> nested content parameter) and developers should be able to use the
> status variable ("status" or whatever) somehow inside the directive.
> So, my assumption was, if the parameter is given to the directive, I
> can probably fill in a variable somehow by resolving the bind status
> object. I guess I'm totally ignorant on how nested content parameters
> work with directives...

A directive can't create a nested content variable. The caller can
declare a nested content parameter, like `<@spring.bind ...; status>`,
and then the directive can specify the actual value of that parameter
when it calls the nested content with
`callPlace.executeNestedContent`. (Of course it's likely that the
caller will chose a shorter name, like `s`; see earlier mails.)

You may wonder if we should allow the directive to create a nested
content variable. I think it's a probably a quite bad library design
if you need that. Creating variables without the caller seeing that
makes understanding what's going on significantly harder. Worse, that
magically appearing variable can hide a variable from a higher scope.
If the caller doesn't ask for the creation of the variable, then that
variable should be under the `spring` namespace. Of course that
(`spring.status`) would be more verbose than using a nested content
parameter, so it doesn't make sense in this case.

> Regards,
>
> Woonsan
>
>>
>>> Another question: is it okay to assume that I have a
>>> DefaultObjectWrapper when this directive is executed and throw an
>>> exception otherwise?
>>
>> At the moment it's enough to expect it to be an
>> ObjectWrapperAndUnwrapper. FreeMarkerServlet also requires that BTW.
>> I'm not sure if in the future expecting DefaultObjectWrapper will be
>> necessary, but hopefully not. (Custom JSP tag support expects a
>> DefaultObjectWrapper at the moment, for wrapping JSP functions.)
>>
 [snip]
> With nested content parameter it's also more obvious what's going on,
> and if you chose a shorter name it's not that verbose either:
>
>   <@spring.bind "user.name"; s>
> ${s.value}
>>>
>>> If I choose to use nested content parameter instead of having
>>> env.setVariable() calls, how can I get access to the parameter, s?
>>> Can I access it through any argument of the #execute() method in the
>>> Java code? Any example java code using nested content parameter?
>>
>> I'm not sure what you try to achieve. As the nested content parameter
>> is local to the nested content, similarly as a local variable is local
>> to the method body in Java, where do you want to access `s` from, and
>> why?
>>
>>> Thanks in advance,
>>>
>>> Woonsan
>>>
 [snip]

 Eh... of course the comment is outdated here. It will work in this
 case. `s` can't be hidden by anyone there, it wouldn't mater if we had
 <#local s = 'blah'> for example.

 --
 Thanks,
  Daniel Dekany

>>>
>>
>> --
>> Thanks,
>>  Daniel Dekany
>>
>

-- 
Thanks,
 Daniel Dekany

Re: (FREEMARKER-55) FM3 freemarker-spring module, Web MVC support

[Woonsan Ko]

On Tue, Aug 29, 2017 at 5:00 AM, Daniel Dekany  wrote:
> Tuesday, August 29, 2017, 5:25:31 AM, Woonsan Ko wrote:
>
>> Hi Daniel,
>>
>> Thanks for the remarks and hints! Please see my question inline.
>>
>> On Sat, Aug 26, 2017 at 4:50 AM, Daniel Dekany  wrote:
>>> Saturday, August 26, 2017, 10:44:03 AM, Daniel Dekany wrote:
>>>
>>> [snip]
   <#macro foo>
 <#local status = 'blah'>
 <@spring.bind "user.name">
   ${status.value} <#-- Means 'blah'.value, won't work -->
>>>
>>> To clarify, here I have assumed that spring.bind calls
>>> env.setVariable("status", ...) internally. Then the local with the
>>> same name shadows that tempolate-namespace scoped variable.
>>
>> It seems working with env.setVariable("status", ...):
>>
>> // Excerpt from
>> https://github.com/woonsan/incubator-freemarker/blob/feature/FREEMARKER-55-2/freemarker-spring/src/main/java/org/apache/freemarker/spring/model/BindDirective.java
>
> Looks good. But what are your thoughts on using a nested content
> parameter instead?

I'm sure I'm missing something or in misunderstandings. I want to try
the option using a nested content parameter instead, but I don't know
how to do that. ;-)
Ideally, <@spring.bind ... /> directive needs to set a local variable
called "status" or whatever (if the variable name can be given by a
nested content parameter) and developers should be able to use the
status variable ("status" or whatever) somehow inside the directive.
So, my assumption was, if the parameter is given to the directive, I
can probably fill in a variable somehow by resolving the bind status
object. I guess I'm totally ignorant on how nested content parameters
work with directives...

Regards,

Woonsan

>
>> Another question: is it okay to assume that I have a
>> DefaultObjectWrapper when this directive is executed and throw an
>> exception otherwise?
>
> At the moment it's enough to expect it to be an
> ObjectWrapperAndUnwrapper. FreeMarkerServlet also requires that BTW.
> I'm not sure if in the future expecting DefaultObjectWrapper will be
> necessary, but hopefully not. (Custom JSP tag support expects a
> DefaultObjectWrapper at the moment, for wrapping JSP functions.)
>
>>> [snip]
 With nested content parameter it's also more obvious what's going on,
 and if you chose a shorter name it's not that verbose either:

   <@spring.bind "user.name"; s>
 ${s.value}
>>
>> If I choose to use nested content parameter instead of having
>> env.setVariable() calls, how can I get access to the parameter, s?
>> Can I access it through any argument of the #execute() method in the
>> Java code? Any example java code using nested content parameter?
>
> I'm not sure what you try to achieve. As the nested content parameter
> is local to the nested content, similarly as a local variable is local
> to the method body in Java, where do you want to access `s` from, and
> why?
>
>> Thanks in advance,
>>
>> Woonsan
>>
>>> [snip]
>>>
>>> Eh... of course the comment is outdated here. It will work in this
>>> case. `s` can't be hidden by anyone there, it wouldn't mater if we had
>>> <#local s = 'blah'> for example.
>>>
>>> --
>>> Thanks,
>>>  Daniel Dekany
>>>
>>
>
> --
> Thanks,
>  Daniel Dekany
>