Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Fixed
-
2.1.8
-
None
-
None
-
Operating System: other
Platform: Other
-
35228
Description
In [1] in we discussed an XSP expression syntax for attribute value and text
expressions. This patch makes expressions available.
In XSP, you now can write:
<elem attrib="{#expression}" which will be expanded to:
<elem><xsp:attribute
name="attrib"><xsp:expr>expression</xsp:expr><xsp:attribute></elem>
or
<elem>Hello #user.getFullName</elem>, which will be expanded to <elem>Hello
<xsp:expr>user.getFullName</xsp:expr></elem>
Writing {##text} will prevent expansion and will be replaced by the text
"{#text}". Inside expressions, write "##" to get "#" and "#}" to get "}".
This works in XSPs as well as logicsheets. The '#' was chosen out of the
discussed characters because it is IMO least likely to occur in an expression in
the used languages.
This feature is turned on by default and can be turned off by setting
"attribute-value-interpolation" or "text-interpolation" to "false" in cocon.xconf:
<markup-languages>
<xsp-language ... attribute-value-interpolation="false"
text-interpolation="false">
...
</xsp-language>
</markup-language>
It can be turned on or off on a per-XSP-/logicsheet-basis by setting attributes
"attribute-value-interpolation" or "text-interpolation" of the top level element
to true or false. Note that these attributes must belong to the
"http://apache.org/xsp" namespace.
How it works:
New class XSPExpressionParser is a parser for the expressions. New class
XSPExpressionFilter is a filter that gets SAX events with embedded expressions
and generates SAX events for expanded expressions. This is used in LogicSheet to
filter read logicsheers and in XSPMarkupLanguage to filter the XSP.
Changes to existing code.
LogicSheets need to know the namespace and uri of the markup language in order
to replace expressions. Therefor AbstractMarkupLanguage needs to know this when
reading logicsheets. This meant that I had to move this configuration
information from parametrize to configure. It is unclear to me anyway, why
AbstractMarkupLanguage used both methods at the same time (which are described
as "incompatible" in the Avalon documentation).
The old PreProcessFilter wraps text() nodes in <xsp:text> elements inside some
tags (See [2]). It is unclear to me why this was done, and all XSPs I've tested
worked without this. I got no reponse on the list, so I left this feature out of
the new PreProcessFilter.
If any of the above changes need further discussion, clarification or change,
please tell me and I'll update the patch.
This patch also should be applied to 2.2. If it does not work, again please tell
me and I'll make a 2.2 patch available.
[1] http://marc.theaimsgroup.com/?l=xml-cocoon-dev&m=111693513631888&w=2
[2] http://marc.theaimsgroup.com/?l=xml-cocoon-dev&m=111778627925208&w=2
expressions. This patch makes expressions available.
In XSP, you now can write:
<elem attrib="{#expression}" which will be expanded to:
<elem><xsp:attribute
name="attrib"><xsp:expr>expression</xsp:expr><xsp:attribute></elem>
or
<elem>Hello #user.getFullName</elem>, which will be expanded to <elem>Hello
<xsp:expr>user.getFullName</xsp:expr></elem>
Writing {##text} will prevent expansion and will be replaced by the text
"{#text}". Inside expressions, write "##" to get "#" and "#}" to get "}".
This works in XSPs as well as logicsheets. The '#' was chosen out of the
discussed characters because it is IMO least likely to occur in an expression in
the used languages.
This feature is turned on by default and can be turned off by setting
"attribute-value-interpolation" or "text-interpolation" to "false" in cocon.xconf:
<markup-languages>
<xsp-language ... attribute-value-interpolation="false"
text-interpolation="false">
...
</xsp-language>
</markup-language>
It can be turned on or off on a per-XSP-/logicsheet-basis by setting attributes
"attribute-value-interpolation" or "text-interpolation" of the top level element
to true or false. Note that these attributes must belong to the
"http://apache.org/xsp" namespace.
How it works:
New class XSPExpressionParser is a parser for the expressions. New class
XSPExpressionFilter is a filter that gets SAX events with embedded expressions
and generates SAX events for expanded expressions. This is used in LogicSheet to
filter read logicsheers and in XSPMarkupLanguage to filter the XSP.
Changes to existing code.
LogicSheets need to know the namespace and uri of the markup language in order
to replace expressions. Therefor AbstractMarkupLanguage needs to know this when
reading logicsheets. This meant that I had to move this configuration
information from parametrize to configure. It is unclear to me anyway, why
AbstractMarkupLanguage used both methods at the same time (which are described
as "incompatible" in the Avalon documentation).
The old PreProcessFilter wraps text() nodes in <xsp:text> elements inside some
tags (See [2]). It is unclear to me why this was done, and all XSPs I've tested
worked without this. I got no reponse on the list, so I left this feature out of
the new PreProcessFilter.
If any of the above changes need further discussion, clarification or change,
please tell me and I'll update the patch.
This patch also should be applied to 2.2. If it does not work, again please tell
me and I'll make a 2.2 patch available.
[1] http://marc.theaimsgroup.com/?l=xml-cocoon-dev&m=111693513631888&w=2
[2] http://marc.theaimsgroup.com/?l=xml-cocoon-dev&m=111778627925208&w=2