From adamlounds at gmail.com  Fri Mar  3 15:54:11 2006
From: adamlounds at gmail.com (adam)
Date: Fri Mar  3 15:54:22 2006
Subject: [Petal] metal:use-macro should be TALES?
Message-ID: <6228ceab0603030754k260dfd19l70a91c4292d62934@mail.gmail.com>

Howdy.

I'm fairly new to petal, and have been battling through the standard
problems (overcoming my aversion to &amp;s in hrefs, javascript stuff etc
etc), but am stumped on the latest one.

Basically I want to choose which metal template to use at run-time, using
something along the lines of

metal:use-macro="dat/header" or
metal:use-macro="foo/${dat/bar}/baz.xml" or
metal:use-macro="string: dat/header"

Is this supported? Should I be doing something entirely different?

Thanks for your help,
--
Adam
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://lists.webarch.co.uk/pipermail/petal/attachments/20060303/393caa2d/attachment.html
From jon at 2xlp.com  Fri Mar  3 16:54:07 2006
From: jon at 2xlp.com (Jonathan Vanasco)
Date: Fri Mar  3 16:54:20 2006
Subject: [Petal] metal:use-macro should be TALES?
In-Reply-To: <6228ceab0603030754k260dfd19l70a91c4292d62934@mail.gmail.com>
References: <6228ceab0603030754k260dfd19l70a91c4292d62934@mail.gmail.com>
Message-ID: <D8AF5855-100F-407C-A4FF-AA35C5F81430@2xlp.com>

Usually for things like that, i do:

<tal tal:omit-tag="string:1" tal:define="dynVar foo/${dat/bar}/ 
baz.xml"/>
<span tal:attributes="class var"/>

I'm not sure if that will work in your example.

btw, you can just
	$processed =~ s/&amp;/&/g ;

i do that for a bunch of petal symbols when doing emails/etc


On Mar 3, 2006, at 10:54 AM, adam wrote:

> Howdy.
>
> I'm fairly new to petal, and have been battling through the  
> standard problems (overcoming my aversion to &amp;s in hrefs,  
> javascript stuff etc etc), but am stumped on the latest one.
>
> Basically I want to choose which metal template to use at run-time,  
> using something along the lines of
>
> metal:use-macro="dat/header" or
> metal:use-macro="foo/${dat/bar}/baz.xml" or
> metal:use-macro="string: dat/header"
>
> Is this supported? Should I be doing something entirely different?
>
> Thanks for your help,
> -- 
> Adam

From adamlounds at gmail.com  Mon Mar  6 14:15:25 2006
From: adamlounds at gmail.com (adam)
Date: Mon Mar  6 14:15:34 2006
Subject: [Petal] metal:use-macro should be TALES?
In-Reply-To: <6228ceab0603060356m7b4b1754nf14dc17d8179d0ce@mail.gmail.com>
References: <6228ceab0603030754k260dfd19l70a91c4292d62934@mail.gmail.com>
	<D8AF5855-100F-407C-A4FF-AA35C5F81430@2xlp.com>
	<6228ceab0603060356m7b4b1754nf14dc17d8179d0ce@mail.gmail.com>
Message-ID: <6228ceab0603060615p3a46c225l1db37e8e3bf4c3ae@mail.gmail.com>

Not quite, I want to choose the file to include, not just have a
string " foo/${dat/bar}/
baz.xml" in the output.

I'm very wary of replacing all &amp;s with & - I'm expecting something like

<a href="http://server/cgi?foo=bar&baz=boogie">http://server/cgi?foo=bar&amp;baz=boogie</a>

where the visible content has &amp; but the href has &, so replacing
all &amp;s would be Bad.

Patch:
--- lib/Petal/CodeGenerator.pm  2006-01-17 12:27:53.000000000 +0000
+++ libnew/Petal/CodeGenerator.pm       2006-03-06 11:40:46.000000000 +0000
@@ -233,15 +233,23 @@
     my $included_from = $petal_object->_file();
     $included_from =~ s/\#.*$//;

+    # allow ${dat/var} expansion in includes
+    if($path=~ /^\$\{(.*)\}$/ ) {
+       $path = $class->comp_expr_encoded ($1);
+    } else {
+       $path = "'$path'";
+    }
+
     $class->add_code ("do {");
     $class->indent_increment();

     $class->add_code ("my \$new_hash = \$hash->new();");
     $class->add_code ("\$new_hash->{__included_from__} = '$included_from';");
+    $class->add_code ("my \$path = $path;");

     (defined $lang and $lang) ?
-        $class->add_code ("my \$res = eval { Petal->new (file =>
'$path', lang => '$lang')->process (\$new_hash) };") :
-       $class->add_code ("my \$res = eval { Petal->new
('$path')->process (\$new_hash) };");
+        $class->add_code ("my \$res = eval { Petal->new (file =>
\$path, lang => '$lang')->process (\$new_hash) };") :
+       $class->add_code ("my \$res = eval { Petal->new
(\$path)->process (\$new_hash) };");
?????
     $class->add_code ("\$res = \"<!--\\n\$\@\\n-->\" if (defined \$\@
and \$\@);");
     $class->add_code ("\$res;");

A bit (OK, a lot) hacky, but it allows me to dynamically choose the
file to be included.

I think the underlying problem is that nested variables aren't
supported, as otherwise
<xi:include href="${dat/header}" />
 would be canonicalized to
<?include href="<?var name="dat/header"?>"?>
or whatever and expanded accordingly. I can see how nested expressions
would be complicated though :-)

On 3/3/06, Jonathan Vanasco <jon@2xlp.com> wrote:
> Usually for things like that, i do:
>
> <tal tal:omit-tag="string:1" tal:define="dynVar foo/${dat/bar}/
> baz.xml"/>
> <span tal:attributes="class var"/>
>
> I'm not sure if that will work in your example.
>
> btw, you can just
>         $processed =~ s/&amp;/&/g ;
>
> i do that for a bunch of petal symbols when doing emails/etc
>
>
> On Mar 3, 2006, at 10:54 AM, adam wrote:
>
> > Howdy.
> >
> > I'm fairly new to petal, and have been battling through the
> > standard problems (overcoming my aversion to &amp;s in hrefs,
> > javascript stuff etc etc), but am stumped on the latest one.
> >
> > Basically I want to choose which metal template to use at run-time,
> > using something along the lines of
> >
> > metal:use-macro="dat/header" or
> > metal:use-macro="foo/${dat/bar}/baz.xml" or
> > metal:use-macro="string: dat/header"
> >
> > Is this supported? Should I be doing something entirely different?
> >
> > Thanks for your help,
> > --
> > Adam
--
Adam


--
Adam
From rabbit at rabbit.us  Sun Mar 19 17:35:38 2006
From: rabbit at rabbit.us (Peter Rabbitson)
Date: Sun Mar 19 17:35:56 2006
Subject: [Petal] Accessing coderefs
Message-ID: <20060319173538.GB30545@rabbit.us>

Greetings,

Is it a bug or a feature that if I have:

$petal->process ( { coderef => \&useful_sub } );

I can not do

...petal:content="coderef --var1 --var2"...

?

Thanks 
From bruno at mkdoc.com  Mon Mar 20 11:21:49 2006
From: bruno at mkdoc.com (Bruno Postle)
Date: Mon Mar 20 11:21:53 2006
Subject: [Petal] Accessing coderefs
In-Reply-To: <20060319173538.GB30545@rabbit.us>
References: <20060319173538.GB30545@rabbit.us>
Message-ID: <441E904D.2050702@mkdoc.com>

Peter Rabbitson wrote:
> 
> Is it a bug or a feature that if I have:
> 
> $petal->process ( { coderef => \&useful_sub } );
> 
> I can not do
> 
> ...petal:content="coderef --var1 --var2"...

This would be ok as an object, what about it doesn't work?

-- 
Bruno
From rabbit at rabbit.us  Mon Mar 20 14:09:37 2006
From: rabbit at rabbit.us (Peter Rabbitson)
Date: Mon Mar 20 14:09:49 2006
Subject: [Petal] Accessing coderefs
In-Reply-To: <441E904D.2050702@mkdoc.com>
References: <20060319173538.GB30545@rabbit.us> <441E904D.2050702@mkdoc.com>
Message-ID: <20060320140937.GA27377@rabbit.us>

On Mon, Mar 20, 2006 at 11:21:49AM +0000, Bruno Postle wrote:
> Peter Rabbitson wrote:
> >
> >Is it a bug or a feature that if I have:
> >
> >$petal->process ( { coderef => \&useful_sub } );
> >
> >I can not do
> >
> >...petal:content="coderef --var1 --var2"...
> 
> This would be ok as an object, what about it doesn't work?
> 
> -- 
> Bruno

Petal gives me no problems if I am to use object/method and supply
parameters. In my case it is very problematic to supply an object to
process(), so I was wondering if there is a way to call a coderef.

Peter

From bruno at mkdoc.com  Thu Mar 23 17:25:37 2006
From: bruno at mkdoc.com (Bruno Postle)
Date: Thu Mar 23 17:25:41 2006
Subject: [Petal] Accessing coderefs
In-Reply-To: <20060320140937.GA27377@rabbit.us>
References: <20060319173538.GB30545@rabbit.us> <441E904D.2050702@mkdoc.com>
	<20060320140937.GA27377@rabbit.us>
Message-ID: <20060323172536.GA12450@mkdoc.com>

On Mon 20-Mar-2006 at 09:09 -0500, Peter Rabbitson wrote:
>
> Petal gives me no problems if I am to use object/method and supply
> parameters. In my case it is very problematic to supply an object to
> process(), so I was wondering if there is a way to call a coderef.

No, it seems that Petal explicitly stops you passing parameters to 
anything that isn't an object.

The following patch seems to make this work, though it hasn't had 
any thorough testing.  Can you try it out?

Index: lib/Petal/Hash/Var.pm
===================================================================
RCS file: /var/spool/cvs/Petal/lib/Petal/Hash/Var.pm,v
retrieving revision 1.17
diff -r1.17 Var.pm
98a99,100
>           unless (isa $current->{$next}, 'CODE')
>           {
100a103
>           }
130c133
<       $current = (isa ($current, 'CODE')) ? $current->() : $current;
---
>       $current = (isa ($current, 'CODE')) ? $current->(@args) : $current;

From jon at 2xlp.com  Wed Mar 29 02:35:18 2006
From: jon at 2xlp.com (Jonathan Vanasco)
Date: Wed Mar 29 02:35:40 2006
Subject: [Petal] using $ in petal 
In-Reply-To: <20060319173538.GB30545@rabbit.us>
References: <20060319173538.GB30545@rabbit.us>
Message-ID: <CD51B806-091E-4255-95B2-2C8F72C9119F@2xlp.com>

i can't figure this out

if $a is in a document, petal shows me '$a' -- good.

if $aa is in a Petal template, its stripped out.  any $ followed by 2  
or more chars is stripped out.

does anyone know why?  bug , feature , workaround ?  i was trying to  
use petal to generate some downloadable customized php snippets, and  
i've tried just about every escape sequence with no luck.


From fergal at esatclear.ie  Wed Mar 29 09:55:05 2006
From: fergal at esatclear.ie (Fergal Daly)
Date: Wed Mar 29 10:02:28 2006
Subject: [Petal] using $ in petal
In-Reply-To: <CD51B806-091E-4255-95B2-2C8F72C9119F@2xlp.com>
References: <20060319173538.GB30545@rabbit.us>
	<CD51B806-091E-4255-95B2-2C8F72C9119F@2xlp.com>
Message-ID: <875029960603290055i79f59e0co8fbd7ff9511c9297@mail.gmail.com>

Petal is not strict TAL, it also includes some "features" like
$variable being the same as <span petal:replace="variable" />. So I
guess $aa is being replaced it's value (ie blank), my guess is that
something in the parser is wrong so it doesn't correctly recognise
single letter variables with $,

F

On 3/29/06, Jonathan Vanasco <jon@2xlp.com> wrote:
> i can't figure this out
>
> if $a is in a document, petal shows me '$a' -- good.
>
> if $aa is in a Petal template, its stripped out.  any $ followed by 2
> or more chars is stripped out.
>
> does anyone know why?  bug , feature , workaround ?  i was trying to
> use petal to generate some downloadable customized php snippets, and
> i've tried just about every escape sequence with no luck.
>
>
>
From jon at 2xlp.com  Wed Mar 29 17:37:47 2006
From: jon at 2xlp.com (Jonathan Vanasco)
Date: Wed Mar 29 17:37:57 2006
Subject: [Petal] using $ in petal
In-Reply-To: <875029960603290055i79f59e0co8fbd7ff9511c9297@mail.gmail.com>
References: <20060319173538.GB30545@rabbit.us>
	<CD51B806-091E-4255-95B2-2C8F72C9119F@2xlp.com>
	<875029960603290055i79f59e0co8fbd7ff9511c9297@mail.gmail.com>
Message-ID: <00FF9E9B-A51E-4719-BA05-AF16667CF43D@2xlp.com>


On Mar 29, 2006, at 3:55 AM, Fergal Daly wrote:

> Petal is not strict TAL, it also includes some "features" like
> $variable being the same as <span petal:replace="variable" />. So I
> guess $aa is being replaced it's value (ie blank), my guess is that
> something in the parser is wrong so it doesn't correctly recognise
> single letter variables with $,

Are there any known workarounds?  I can't seem to figure out any way  
to put the string $abc into a document

\$var didn't work
$$var didn't work

$ isn't a standard html entity - its a 'special character' and not  
widely supported
&dollar; is on some spec sheets, but i haven't found a browser that  
shows it
&#036; is the equiv, but Petal seems to parse  it as a $ for  
interpolation ( so $a works, but $a[a]* doesn't )