ViewSchema#getTypeForName():
         if (name.equals("text") ||
             name.equals("html"))
-            name = "string";
+            name = "text";

A bit simpler:
if ("html".equals(name)) {
    // [insert explanatory comment here]
    name = "text";
}


+    html: StringPresentationType,
+    text: TextPresentationType,

1) If text and html are treated as synonyms for now, they should have the same presentation type. 2) test/smoke/presentation-types.lzl needs to be updated to test the new types, too.


In TextPresentationType.present():
+          i = e + 1;

i = e? e is the position of the ';', e+1 is the next character which needs to be processed, but as the for-loop increments i another time, the next character which is processed is at position e+2, so the char value[e+1] is left out.


NodeModel#compileAttribute():
+                // Immediate string, token, and id types are
+                // auto-quoted, but must _first_ be parsed as ES
+                // strings!

I'm not sure whether this is valid for IDs and tokens. Have you tested this?
For example this test doesn't compile at all, that means the compiler doesn't not parse the values as ES strings:
<canvas>
<!-- name's type is token, id's type is ID -->
<view name="\x60" id="\x61"/>
</canvas>


+            } else if (type == ViewSchema.XML_CDATA_TYPE
+                       || type == ViewSchema.XML_CONTENT_TYPE) {
+                // Immediate text and html types are auto-quoted; They
+                // will be treated differently in the runtime
+                // PresentationType system

This code is a bit misleading, because right know html-type is auto-converted to text-type (ViewSchema#getTypeForName() !). An additional comment should be added for clarification.


Some components need to be updated to change the type from string to text for backward compatibility. I'd suggest to do this in a subsequent change set. For example in lz/windowpanel.lzx, the "title" attribute is set to string, but it's only used in a constraint for a <text> element. I think in this case the type for "title" should be changed to text.


And it'd be nice if docs/src/developers/methods-events-attributes.dbk was updated to reflect the new understanding of string/text/html.



On 6/12/2010 9:48 PM, P T Withington wrote:
[UPDATED:

Per André's last suggestion, I am not trying to sort out the text/html issues 
here.  I am only splitting string from text/html, so that attributes with type 
string can have literal values expressed as ECMAScript String literals.

]

Change 20100525-ptw-1 by [email protected] on 2010-05-25 09:56:35 EDT
     in /Users/ptw/OpenLaszlo/trunk
     for http://svn.openlaszlo.org/openlaszlo/trunk

Summary: Correct quoting of string-typed attributes

Bugs Fixed: LPP-9027 pattern attribute on text doesn't work as expected in trunk

Technical Reviewer: [email protected] (pending)
QA Reviewer: [email protected] (pending)

Release Notes:

     The LZX `string` attribute type is no longer a synonym for `text`
     and `html`.  The `string` type implies an ECMAScript string.
     Literal values assigned to attributes of that type will be
     interpreted as ECMAScript `String` literals, meaning that the
     ECMAScript single-character (\t, \n, etc.) and short (\xNN) and
     long (\uNNNN) escapes will be interpreted as defined in the
     ECMAScript standard.  The `text` and `html` types imply XML
     content.  Literal values assigned to attributes of that type will
     (continue to) be interpreted as verbatim text.  ECMAScript escapes
     will not be interpreted.

     The `<text>/text` attribute is now declared to be of type `text`,
     to indicate that it represents HTML content, not an ECMAScript
     string.  If you redeclare the `text` attribute in a subclass of
     `<text>` you should omit any type declaration as it is redundant.
     (If you had declared the type as `string`, the compiler will issue
     a warning.)

Overview:

     Split the `string` LZX type from the `text` and `html` types
     (leaving the latter two as synonyms for now).  Declare the type of
     `<text>/text` to be `text` (which is the popular concensus in
     components, although perhaps incorrect because<text>  behaves as
     if it were `html`; that is, it interprets markup).

     With this split, we are able to have the LZX compiler interpret
     literal `string` values as ECMAScript `String`s and solve the
     original bug.  In a future change we will address the
     `<text>/text` `text`/`html` issue.

Details:
     lztest-node:  Added tests to verify that setting a string
     attribute using a literal or expression is the same as setting an
     expression attribute to a literal string using the failing pattern
     from the bug (which verifies the use of \u escapes to specify
     unicode characters works).

     LzNode:  Added mapping for text and html presentation types..

     PresentationTypes:  Added text presentation type which
     escapes/unescapes the 5 XML entities.  This presentation type
     should eventually replace the many copies of xml escaping code.

     LzText:  Change the @lzxtype of the `text` attribute from `string`
     to `text`.

     LzDataset:  Correct spelling of @lzxtype noticed in passing

     SchemaBuilder:  Handle `text` and `html` types

     Schema, ViewSchema, NodeModel: Added real types for 'text' and
     'html', distinct from 'string', which affects how immediate values
     of those types are interpreted by the tag compiler: 'string'
     implies an ECMAScript string (i.e., interpreted as an ECMAScript
     String literal), 'text' implies XML CDATA and 'html' implies XML
     content.  In this change, the latter two remain synonyms (as they
     currently are), and are _not_ interpreted as ECMAScript strings,
     but as XML content.  In a later change the implied distinction
     between `text` and `html` will be addressed

     debugger:  Remove unused `escapeText` noticed in passing.

     basecombobox: Remove redundant (conflicting) attribute
     declaration.

     photo, classes: Remove redundant (conflicting) attrbute type
     specification.

Tests:
     ant lztest with new node attribute tests, smokecheck, various
     combinations of demos and runtimes.

Files:
M       test/lztest/lztest-node.lzx
M       WEB-INF/lps/lfc/core/LzNode.lzs
M       WEB-INF/lps/lfc/core/PresentationTypes.lzs
M       WEB-INF/lps/lfc/views/LzText.lzs
M       WEB-INF/lps/lfc/data/LzDataset.lzs
M       WEB-INF/lps/server/src/org/openlaszlo/js2doc/SchemaBuilder.java
M       WEB-INF/lps/server/src/org/openlaszlo/xml/internal/Schema.java
M       WEB-INF/lps/server/src/org/openlaszlo/compiler/ViewSchema.java
M       WEB-INF/lps/server/src/org/openlaszlo/compiler/NodeModel.java
M       lps/components/debugger/debugger.lzx
M       lps/components/base/basecombobox.lzx
M       demos/lzpix/classes/photo.lzx
M       demos/lzpix/classes/classes.lzx
M       demos/lzpix/classes/clipboard.lzx

Changeset: http://svn.openlaszlo.org/openlaszlo/patches/20100525-ptw-1.tar

Reply via email to