tomcat-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mark Thomas <ma...@apache.org>
Subject Re: svn commit: r1379091 - in /tomcat/tc7.0.x/trunk: ./ java/org/apache/el/parser/AstValue.java test/org/apache/el/TestMethodExpressionImpl.java
Date Sat, 01 Sep 2012 09:27:19 GMT
On 01/09/2012 09:28, Mark Thomas wrote:
> On 01/09/2012 01:55, Konstantin Kolinko wrote:
>> 2012/9/1 Mark Thomas <markt@apache.org>:
>>> On 31/08/2012 12:37, Konstantin Kolinko wrote:
>>>> 2012/8/30  <markt@apache.org>:
>>>>> Author: markt
>>>>> Date: Thu Aug 30 19:44:46 2012
>>>>> New Revision: 1379091
>>>>>
>>>>> URL: http://svn.apache.org/viewvc?rev=1379091&view=rev
>>>>> Log:
>>>>> Fix https://issues.apache.org/bugzilla/show_bug.cgi?id=53792
>>>>> Support method expressions that include a method call that is not at
the end of the expression
>>>>>
>>>>> Modified:
>>>>>     tomcat/tc7.0.x/trunk/   (props changed)
>>>>>     tomcat/tc7.0.x/trunk/java/org/apache/el/parser/AstValue.java
>>>>>     tomcat/tc7.0.x/trunk/test/org/apache/el/TestMethodExpressionImpl.java
>>>>>
>>>>> --- tomcat/tc7.0.x/trunk/java/org/apache/el/parser/AstValue.java (original)
>>>>> +++ tomcat/tc7.0.x/trunk/java/org/apache/el/parser/AstValue.java Thu
Aug 30 >>>                  ctx.setPropertyResolved(false);
>>>>
>>>> The above was previously called before calling resolver.getValue(),
>>>> but now it is not called neither before resolver.invoke(), nor before
>>>> resolver.getValue().
>>>
>>> The calls are unnecessary. The CompositeELResolver sets the property to
>>> false at the start of both of those methods and the Resolver
>>> implementations are responsible for setting it if they find a match.
>>
>> This relies on the fact that ELContext.getResolver() is an instance of
>> CompositeELResolver. Other resolvers do not set this flag to false.
>>
>> The API itself (ELContext, ELContextImpl(ELResolver)) allows any resolver type.
>>
>> Though it would seem strange if one uses a non-CompositeELResolver
>> here, as this would omit some features of the EL language.
> 
> The flag is only read by the CompositeELResolver so that shouldn't matter.

Clearly this is not correct. The JSP TCL failures I am seeing in EL
appear to be related to this. I'm digging further.

Mark


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org


Mime
View raw message