Mason

On Mon, 30 Sep 2002, Ken Williams wrote:

> Seems like it would be nice to allow it to be in another
> package, so people could keep a set of local escapes without
> adding them to the HTML::Mason::Escapes namespace.
>
> How about the following patch?

Sure, that looks fine.  Check it in with a doc patch too ;)

A test'd be nice as well.


-dave

/*==================
 http://www.urth.org
we await the New Sun
==================*/

On Monday, September 30, 2002, at 01:45  PM, Dave Rolsky wrote:

> On Mon, 30 Sep 2002, Ken Williams wrote:
>
>> Seems like it would be nice to allow it to be in another
>> package, so people could keep a set of local escapes without
>> adding them to the HTML::Mason::Escapes namespace.
>>
>> How about the following patch?
>
> Sure, that looks fine.  Check it in with a doc patch too ;)
>

Sure.  Actually, I think most of this code should be in 
Interp.pm, not in ApacheHandler.

> A test'd be nice as well.

Yeah.

  -Ken

On Mon, 30 Sep 2002, Ken Williams wrote:

> Sure.  Actually, I think most of this code should be in
> Interp.pm, not in ApacheHandler.

Hmm, I suppose it could be.  I'm not sure if we want to encourage people
to say "foo => bar" when they could just type "foo =>
\&HTML::Mason::Escapes::bar".  The Apache config is a special case because
it's not Perl.


-dave

/*==================
 http://www.urth.org
we await the New Sun
==================*/

> > Sure.  Actually, I think most of this code should be in
> > Interp.pm, not in ApacheHandler.

I agree. I hadn't noticed that this was placed into ApacheHandler, or I
would have made this comment as well.

>
> Hmm, I suppose it could be.  I'm not sure if we want to encourage people
> to say "foo => bar" when they could just type "foo =>
> \&HTML::Mason::Escapes::bar".

Yes, we do (I do anyway). The whole idea of the HTML::Mason::Escapes
shortcut is to be able to conveniently specify escapes that are built into,
and supported by, Mason. There's no reason to treat the ApacheHandler
parameter differently than the regular parameter. More generally...

>  The Apache config is a special case because it's not Perl.

we should strive the Apache config as little of a special case as possible.
First, because special cases cause user confusion and make the docs more
difficult to write. Second, because at some point there may be a third to
configure Mason, and the only way it's going to be scalable is if all these
configuration methods are just thin layers on top of a standard set of
parameters.

Jon

> Seems like it would be nice to allow it to be in another
> package, so people could keep a set of local escapes without
> adding them to the HTML::Mason::Escapes namespace.
>
> How about the following patch?

The original idea was that, if the value was a string, it would be searched
for in HTML::Mason::Escapes; otherwise, it had to be a code value. The code
value could be something in the user's own package, and AFAICT it's
perfectly possible to specify a code value from the Apache config. What
would we gain from this patch?

Jon

On Mon, 30 Sep 2002, Jonathan Swartz wrote:

> > Hmm, I suppose it could be.  I'm not sure if we want to encourage people
> > to say "foo => bar" when they could just type "foo =>
> > \&HTML::Mason::Escapes::bar".
>
> Yes, we do (I do anyway). The whole idea of the HTML::Mason::Escapes
> shortcut is to be able to conveniently specify escapes that are built into,
> and supported by, Mason. There's no reason to treat the ApacheHandler
> parameter differently than the regular parameter. More generally...

Ok.  This is certainly easy enough to move over to Interp.pm.



-dave

/*==================
 http://www.urth.org
we await the New Sun
==================*/

> > Yes, we do (I do anyway). The whole idea of the HTML::Mason::Escapes
> > shortcut is to be able to conveniently specify escapes that are 
> built into,
> > and supported by, Mason. There's no reason to treat the ApacheHandler
> > parameter differently than the regular parameter. More generally...
> 
> Ok.  This is certainly easy enough to move over to Interp.pm.

Ok. Also note that the parameter name is missing from this doc:

  PerlSetVar "flag  => \&subroutine"
  PerlSetVar "uc    => sub { ${$_[0]} = uc ${$_[0]}; }"
  PerlAddVar "thing => other_thing"

On Mon, 30 Sep 2002, Jonathan Swartz wrote:

> > > Yes, we do (I do anyway). The whole idea of the HTML::Mason::Escapes
> > > shortcut is to be able to conveniently specify escapes that are
> > built into,
> > > and supported by, Mason. There's no reason to treat the ApacheHandler
> > > parameter differently than the regular parameter. More generally...
> >
> > Ok.  This is certainly easy enough to move over to Interp.pm.
>
> Ok. Also note that the parameter name is missing from this doc:
>
>   PerlSetVar "flag  => \&subroutine"
>   PerlSetVar "uc    => sub { ${$_[0]} = uc ${$_[0]}; }"
>   PerlAddVar "thing => other_thing"

Will be fixed in a moment.


/*==================
 http://www.urth.org
we await the New Sun
==================*/

On Tuesday, October 1, 2002, at 03:25  AM, Jonathan Swartz wrote:

>> Seems like it would be nice to allow it to be in another
>> package, so people could keep a set of local escapes without
>> adding them to the HTML::Mason::Escapes namespace.
>>
>> How about the following patch?
>
> The original idea was that, if the value was a string, it would 
> be searched
> for in HTML::Mason::Escapes; otherwise, it had to be a code value.

Yes, I know.  I didn't like that.  Since it went a little 
against the grain, I posted my patch rather than just applying 
it.


> The code value could be something in the user's own package, 
> and AFAICT it's
> perfectly possible to specify a code value from the Apache config. What
> would we gain from this patch?

I would be able to specify

   PerlSetVar MasonEscapeFlags "h => MyApp::Escapes::my_routine"


In general, people are going to want to write their own escaping 
routines, and we should neither (a) make them shove them into 
the HTML::Mason::Escapes namespace, nor (b) make it more 
difficult/ugly to use user-defined routines than the ones we 
provide.

  -Ken

Ok. I don't feel strongly about it either way.

> I would be able to specify
> 
>    PerlSetVar MasonEscapeFlags "h => MyApp::Escapes::my_routine"
> 
> 
> In general, people are going to want to write their own escaping 
> routines, and we should neither (a) make them shove them into 
> the HTML::Mason::Escapes namespace, nor (b) make it more 
> difficult/ugly to use user-defined routines than the ones we 
> provide.
> 
>   -Ken
> 
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Mason-devel mailing list
> Mason-devel@...
> https://lists.sourceforge.net/lists/listinfo/mason-devel
>