Re: Problem with empty key word on a collection.
On Thursday, February 13, 2003, at 05:04 PM, Pierre Delisle wrote: The change that Jeff is proposing, although of interest, would unfortunately break compatibility with the spec. (sorry guys, but the Expert Group ain't perfect) Why would extending the empty keyword to support Collections break compatibility with the spec? - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: Problem with empty key word on a collection.
-Original Message- From: James Cook [mailto:[EMAIL PROTECTED]] Sent: Friday, February 14, 2003 8:17 AM To: Tag Libraries Users List Subject: Re: Problem with empty key word on a collection. On Thursday, February 13, 2003, at 05:04 PM, Pierre Delisle wrote: The change that Jeff is proposing, although of interest, would unfortunately break compatibility with the spec. (sorry guys, but the Expert Group ain't perfect) Why would extending the empty keyword to support Collections break compatibility with the spec? Modifying the empty keyword to support Collections doesn't break anything per se, but it does exceed the definition of empty in Section A.3.8 in the JSTL 1.0 specification. I believe that the JSPA talks about adhering to the spec in independent implementations of a JSR. I'm very uninterested in starting a flamewar about JCP, but it is important to note that adhering to the spec allows for portability over multiple implementations of JSTL and EL. I've already submitted a patch to commons-jexl to make empty work with Collection, but it won't be of any help to anyone using the current Standard Taglib and JSTL 1.0. I think this is what the nascent Unstandard Taglib what meant for, providing functionality for little bits of functionality that might have been inadvertently overlooked by the first rev of JSTL in a way that doesn't conflict with existing JCP specs. Tim O'Brien - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: Problem with empty key word on a collection.
I'll bite. What's the Unstandard Taglib? Is there a website for this? -Original Message- From: O'brien, Tim [mailto:[EMAIL PROTECTED]] Sent: Friday, February 14, 2003 11:45 AM To: 'Tag Libraries Users List' Subject: RE: Problem with empty key word on a collection. -Original Message- From: James Cook [mailto:[EMAIL PROTECTED]] Sent: Friday, February 14, 2003 8:17 AM To: Tag Libraries Users List Subject: Re: Problem with empty key word on a collection. On Thursday, February 13, 2003, at 05:04 PM, Pierre Delisle wrote: The change that Jeff is proposing, although of interest, would unfortunately break compatibility with the spec. (sorry guys, but the Expert Group ain't perfect) Why would extending the empty keyword to support Collections break compatibility with the spec? Modifying the empty keyword to support Collections doesn't break anything per se, but it does exceed the definition of empty in Section A.3.8 in the JSTL 1.0 specification. I believe that the JSPA talks about adhering to the spec in independent implementations of a JSR. I'm very uninterested in starting a flamewar about JCP, but it is important to note that adhering to the spec allows for portability over multiple implementations of JSTL and EL. I've already submitted a patch to commons-jexl to make empty work with Collection, but it won't be of any help to anyone using the current Standard Taglib and JSTL 1.0. I think this is what the nascent Unstandard Taglib what meant for, providing functionality for little bits of functionality that might have been inadvertently overlooked by the first rev of JSTL in a way that doesn't conflict with existing JCP specs. Tim O'Brien - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: Problem with empty key word on a collection.
It's mainly a set of ideas at the moment. People come to this list a lot asking for features in JSTL. Rather than send them away unsatisified, the idea is to implement them in the Unstandard taglib and have happy users. Additionally, it would provide an interesting research ground for the JSTL people and provide some way in which Users can really force home the fact that the c:systemExit/ tag is essential. We set things up for it last week, the developer list has a chunk about it in the mail archives. Tim's put a bit of code in there and Glenn and I have setup the infrastructure for it, but I'm focusing on String 1.0.1 first at the moment. Website to come. Hen On Fri, 14 Feb 2003, Jerome Jacobsen wrote: I'll bite. What's the Unstandard Taglib? Is there a website for this? -Original Message- From: O'brien, Tim [mailto:[EMAIL PROTECTED]] Sent: Friday, February 14, 2003 11:45 AM To: 'Tag Libraries Users List' Subject: RE: Problem with empty key word on a collection. -Original Message- From: James Cook [mailto:[EMAIL PROTECTED]] Sent: Friday, February 14, 2003 8:17 AM To: Tag Libraries Users List Subject: Re: Problem with empty key word on a collection. On Thursday, February 13, 2003, at 05:04 PM, Pierre Delisle wrote: The change that Jeff is proposing, although of interest, would unfortunately break compatibility with the spec. (sorry guys, but the Expert Group ain't perfect) Why would extending the empty keyword to support Collections break compatibility with the spec? Modifying the empty keyword to support Collections doesn't break anything per se, but it does exceed the definition of empty in Section A.3.8 in the JSTL 1.0 specification. I believe that the JSPA talks about adhering to the spec in independent implementations of a JSR. I'm very uninterested in starting a flamewar about JCP, but it is important to note that adhering to the spec allows for portability over multiple implementations of JSTL and EL. I've already submitted a patch to commons-jexl to make empty work with Collection, but it won't be of any help to anyone using the current Standard Taglib and JSTL 1.0. I think this is what the nascent Unstandard Taglib what meant for, providing functionality for little bits of functionality that might have been inadvertently overlooked by the first rev of JSTL in a way that doesn't conflict with existing JCP specs. Tim O'Brien - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Problem with empty key word on a collection.
Not surprised at that - a prior company I worked at, where I first learned Java, they had consultants and (i thought..) smart engineers working together to develop a quite large web application. The company folded a year later, and I moved on. It wasn't till after I got into jsp/servlets that I realized that they had horribly mangled the MVC.. namely they wrote the entire thing in a couple massive behemoth JSP page and not a single servlet in sight. I remember one particular .jsp being over 4000 lines. I've since come to the conclusion that so-called consultants get their $150/hr jobs either by being experts, or by sheer incompetence. Which happens more often, I don't know. :) On Friday, February 14, 2003, at 01:33 PM, O'brien, Tim wrote: You might laugh at this, but I've worked with high priced consultants from well respected companies who always managed to do something like System.exit(0); in a Servlet. :-) Or, better yet, the certified web developer from a 3-letter acronym company (not IBM) who liked to write all content to a temporary file before sending it back to the client. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Problem with empty key word on a collection.
On Thu, 13 Feb 2003, Leon Doud wrote: It appears that the empty keyword doesn't always function on a collection. I must be missing something obvious.. The folder object contains a collection of content. The useBean tag isn't initializing the Folder object. Its there so I can use scriptlet debugging code. What exactly does empty check to determine if a collection is empty or not? It looks like that empty is checking if the folder.content is null or not, not if its size is zero. Folder.getContent() returns a type of java.util.Collection. The 'empty' operator is specified to work with java.util.List and java.util.Map (as well as arrays and strings), not java.util.Collection in general. -- Shawn Bayern JSTL in Action http://www.manning.com/bayern - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: Problem with empty key word on a collection.
Wow. That is extremely unintuitive. Why not work on a Collection (which automatically gives you List)? -Original Message- From: Shawn Bayern [mailto:[EMAIL PROTECTED]] Sent: Thursday, February 13, 2003 11:13 AM To: Tag Libraries Users List Subject: Re: Problem with empty key word on a collection. On Thu, 13 Feb 2003, Leon Doud wrote: It appears that the empty keyword doesn't always function on a collection. I must be missing something obvious.. The folder object contains a collection of content. The useBean tag isn't initializing the Folder object. Its there so I can use scriptlet debugging code. What exactly does empty check to determine if a collection is empty or not? It looks like that empty is checking if the folder.content is null or not, not if its size is zero. Folder.getContent() returns a type of java.util.Collection. The 'empty' operator is specified to work with java.util.List and java.util.Map (as well as arrays and strings), not java.util.Collection in general. -- Shawn Bayern JSTL in Action http://www.manning.com/bayern - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Problem with empty key word on a collection.
Well that would be fine if it was possible to call the isEmpty() method on a collection. But since empty is a key word the parser has a problem with collection.empty. Why not implement the empty keyword on all collections? The method isEmpty is available in the Collection interface. Was there some sort of argument against that when the spec was defined? --- Shawn Bayern [EMAIL PROTECTED] wrote: On Thu, 13 Feb 2003, Leon Doud wrote: It appears that the empty keyword doesn't always function on a collection. I must be missing something obvious.. The folder object contains a collection of content. The useBean tag isn't initializing the Folder object. Its there so I can use scriptlet debugging code. What exactly does empty check to determine if a collection is empty or not? It looks like that empty is checking if the folder.content is null or not, not if its size is zero. Folder.getContent() returns a type of java.util.Collection. The 'empty' operator is specified to work with java.util.List and java.util.Map (as well as arrays and strings), not java.util.Collection in general. -- Shawn Bayern JSTL in Action http://www.manning.com/bayern - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] __ Do you Yahoo!? Yahoo! Shopping - Send Flowers for Valentine's Day http://shopping.yahoo.com - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Problem with empty key word on a collection.
On Thu, 13 Feb 2003, Leon Doud wrote:
Well that would be fine if it was possible to call the isEmpty()
method on a collection. But since empty is a key word the parser has
a problem with collection.empty.
You could use ${collection[empty]}.
Why not implement the empty keyword on all collections? The method
isEmpty is available in the Collection interface. Was there some sort
of argument against that when the spec was defined?
I believe so; I don't remember the details of the debate, but I think the
justification is that the EL doesn't deal with Collections in general in
any other situation.
--
Shawn Bayern
JSTL in Action http://www.manning.com/bayern
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Problem with empty key word on a collection.
Thanks! I forgot about that syntax.
--- Shawn Bayern [EMAIL PROTECTED] wrote:
On Thu, 13 Feb 2003, Leon Doud wrote:
Well that would be fine if it was possible to call
the isEmpty()
method on a collection. But since empty is a key
word the parser has
a problem with collection.empty.
You could use ${collection[empty]}.
__
Do you Yahoo!?
Yahoo! Shopping - Send Flowers for Valentine's Day
http://shopping.yahoo.com
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
RE: Problem with empty key word on a collection.
From: Shawn Bayern [mailto:[EMAIL PROTECTED]] Why not implement the empty keyword on all collections? The method isEmpty is available in the Collection interface. Was there some sort of argument against that when the spec was defined? I believe so; I don't remember the details of the debate, but I think the justification is that the EL doesn't deal with Collections in general in any other situation. ?? I'm pretty sure I can do a forEach across a Collection. This seems entirely inane. Would someone patch the Jakarta implementation to allow the empty operator on Collection? We don't use specifications, we use software. Jeff - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: Problem with empty key word on a collection.
On Thu, 13 Feb 2003, Schnitzer, Jeff wrote: From: Shawn Bayern [mailto:[EMAIL PROTECTED]] Why not implement the empty keyword on all collections? The method isEmpty is available in the Collection interface. Was there some sort of argument against that when the spec was defined? I believe so; I don't remember the details of the debate, but I think the justification is that the EL doesn't deal with Collections in general in any other situation. ?? I'm pretty sure I can do a forEach across a Collection. This seems entirely inane. Would someone patch the Jakarta implementation to allow the empty operator on Collection? We don't use specifications, we use software. For Tomcat, I imagine the answer would be a flat no, but Jeff has a point here. Standard Taglib isn't the RI anymore is it? Doesn't that come from Sun? So can't Standard Taglib add patches like this that are natural and right? Hen - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Problem with empty key word on a collection.
Henri Yandell wrote: On Thu, 13 Feb 2003, Schnitzer, Jeff wrote: From: Shawn Bayern [mailto:[EMAIL PROTECTED]] Why not implement the empty keyword on all collections? The method isEmpty is available in the Collection interface. Was there some sort of argument against that when the spec was defined? I believe so; I don't remember the details of the debate, but I think the justification is that the EL doesn't deal with Collections in general in any other situation. ?? I'm pretty sure I can do a forEach across a Collection. This seems entirely inane. Would someone patch the Jakarta implementation to allow the empty operator on Collection? We don't use specifications, we use software. For Tomcat, I imagine the answer would be a flat no, but Jeff has a point here. Standard Taglib isn't the RI anymore is it? Doesn't that come from Sun? For Tomcat, I imagine the answer would be a flat no, but Jeff has a point here. Standard Taglib isn't the RI anymore is it? Doesn't that come from Sun? Correct. 'standard' is the basis for the RI shipped by Sun. So can't Standard Taglib add patches like this that are natural and right? What is important to note is that 'standard' also ships as an implementation of the JSTL specification. Because of this, 'standard' must therefore pass the TCK. The change that Jeff is proposing, although of interest, would unfortunately break compatibility with the spec. (sorry guys, but the Expert Group ain't perfect) Given that the EL is now moving from the JSTL spec to the JSP spec, the best way to handle this would be for Jeff to submit a comment to the JSP spec comments alias at [EMAIL PROTECTED] Hope this helps, -- Pierre - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
