Re: Cocoon Forms 2.1 - One of two fields required
Merico Raffaele wrote: Tobia Conforto wrote: In Cocoon 2.1, Forms, how do I require the user to fill at least one out of two (or more) fields? For example, I might have the 'phone' and 'mobile' fields. I want the user to fill at least one of them. Where should I put such a validator? I use the following construct to do that: After all relevant form fields I add a form field (in my case id=form-validation) that is not displayed in the form but invokes a javascript validation function. Thank you. Using a separate widget just to validate the collective requirements of other widgets does in fact work. I'm wondering if it is the only solution. It doesn't seem very clean. -Tobia - To unsubscribe, e-mail: users-unsubscr...@cocoon.apache.org For additional commands, e-mail: users-h...@cocoon.apache.org
Re: Cocoon Forms 2.1 - One of two fields required
Wolfgang Ruelfing wrote:
Hello,
you can put the general form validation code straight into the form
element
Thank you! This is exactly what I was looking for.
I had tried to do something like this on my own, but I made an
critical mistake: I set the validation error on the form itself,
instead of one of its children, and it didn't work. Your solution
works very well.
Here are a couple of improvements: I don't set the composite error
if any of the fields already have a more basic error (because I don't
want to override it with mine); and whenever the state of this
composite validation error changes, I push both fields on to the
user (otherwise I would get stale error icons on the browser, using
Ajax.) Can you please comment on my implementation? Is there a
simpler solution?
form xmlns=http://apache.org/cocoon/forms/1.0#definition;
on-create
javascript
widget.setAttribute('phonemobile_valid', true);
/javascript
/on-create
validation
javascript![CDATA[
var phone = widget.lookupWidget('/phone'),
mobile = widget.lookupWidget('/mobile'),
old_valid = widget.getAttribute('phonemobile_valid'),
new_valid = true;
/* we require the user to enter his landline and/or mobile phone;
* if both fields are without a valid value AND without a
temporary,
* invalid value, set an error on both
*/
if (! phone.getValue() ! mobile.getValue()
! phone.getValidationError()
! mobile.getValidationError()) {
var message = Please enter your landline and/or mobile phone;
phone.setValidationError(new ValidationError(message, false));
mobile.setValidationError(new ValidationError(message, false));
new_valid = false;
}
/* if this composite validation state has changed since last time,
* push both fields to the browser
*/
if (old_valid != new_valid) {
widget.addWidgetUpdate(phone);
widget.addWidgetUpdate(mobile);
}
widget.setAttribute('phonemobile_valid', new_valid);
return new_valid;
]]/javascript
/validation
widgets
...
-Tobia
-
To unsubscribe, e-mail: users-unsubscr...@cocoon.apache.org
For additional commands, e-mail: users-h...@cocoon.apache.org
Cocoon Forms 2.1 - One of two fields required
Hello In Cocoon 2.1, Forms, how do I require the user to fill at least one out of two (or more) fields? For example, I might have the 'phone' and 'mobile' fields. I want the user to fill at least one of them, possibly both, and I don't want the form to validate if they're both empty. I've tried to add the validator to one or the other, but it seems the wrong place to put it (Cocoon forms have a strict hierarchical structure) and in fact the behaviour with ajax and browser-update is buggy. I've tried to add the validator to the form widget itself, but then the error message is not picked up by the jx template, and the user is stuck without knowing what is wrong. Where should I put such a validator? Tobia - To unsubscribe, e-mail: users-unsubscr...@cocoon.apache.org For additional commands, e-mail: users-h...@cocoon.apache.org
Re: Failing Mysql JDBC connections
I wrote: As soon as this happens, all pages being worked on at the same time fail with such messages as: SQLTransformer.Query: Failed to execute query [...] com.mysql.jdbc.exceptions.MySQLNonTransientConnectionException: No operations allowed after connection closed. For the record, this was caused by a deadlock problem, where many simultaneous requests would get a connection for their SQLTransformer, exhausting the connection pool, and then get stuck requesting more connections through an input module I wrote myself, before the SQLTransformers had a chance to return the connections to the pool. I worked around the problem by uncapping the database connections: pool-controller min=100 max=150 max-strict=false/ Now I'd like to know how to avoid the deadlock condition in the first place, with capped connections, but I'm not sure it can be done. I've asked the dev list. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Failing Mysql JDBC connections
Hello I have a big problem right now: my Cocoon MySql connections have started to fail randomly, and when they do they bring down all pages currently being generated. My gut feeling is that it's a synchronization problem, some race condition in Cocoon or in the Mysql connector. This is because the failures happen on random pages, that usually render correctly, and also because the higher the concurrency of http requests, the higher the rate of failure. Apparently I didn't stress-test the server configuration with a high enough concurrency, before going online. Here is what I see in the logs: SQLTransformer.Query: Failed to execute query [...] com.mysql.jdbc.CommunicationsException: Communications link failure due to underlying exception: java.net.SocketException: Socket closed at java.net.SocketInputStream.socketRead0(Native Method) at java.net.SocketInputStream.read(SocketInputStream.java:129) at com .mysql.jdbc.util.ReadAheadInputStream.fill(ReadAheadInputStream.java: 113) ... (other mysql jdbc classes follow, then excalibur, then the SQLTransformer) Looking at a network dump, I can see a series of Mysl protocol QUIT messages that tear down all tcp connections to the Mysql server. It happens simultaneously on ALL tcp connections and I can see the Jdbc connector sending appropriate QUIT messages to the server, so it's not a hard connection reset, coming from the network stack. The Mysql server is on localhost, so it's not a network problem. As soon as this happens, all pages being worked on at the same time fail with such messages as: SQLTransformer.Query: Failed to execute query [...] com.mysql.jdbc.exceptions.MySQLNonTransientConnectionException: No operations allowed after connection closed. or: ErrorHandlerHelper: Failed to process pipeline java.lang.NullPointerException at com.mysql.jdbc.ResultSet.findColumn(ResultSet.java:966) ... or other such NPE inside the JDBC driver, on lines where the driver is trying to use its own iternal connection object (which now appears to be null.) After that, the database connections are reopened and Cocoon proceeds normally, but after a few of these error iterations it all hangs. (Memory leak? I don't know.) I'm using Cocoon 2.1.10 and 2.1.11 (same results) and MySql connector 3.1.14 and 5.0.6 (same results.) I'm using SQLTransformer with prepared statements (sql:in-parameter) and a few custom actions of mine which use JDBI prepared statements. The server runs Debian stable with its latest Sun Java 1.5.0 and Mysql 5.0.32 packages. I cannot use the 5.0.8 or 5.1.6 connectors because they don't work with Cocoon 2.1 (a strange bug with column aliases, aka. SELECT column AS alias, worthy of another thread.) Has anybody seen anything like this before? Do you want me to try and build a minimal example that shows the error? Could it be the prepared statements' fault? Any help will be appreciated. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Old Cocoon query
On 17 Apr 2008, Derek Hohls wrote: in this mail: http://marc.info/?l=xml-cocoon-usersm=119996793507680w=2 you gave guidance on how to post-process escaped tags from a database. However, when I implement it, it also strips out all the sql:field tags as well, leaving just text inside the sql:row tags. How can the code be modified not to lose the nested tags? My fault for posting untested code! Try modifying the first XSLT, unescape-pre.xsl, like this: (still untested) xsl:template match=sql:row/*[not(self::sql:rowset) and (contains(., 'amp') or contains(., 'lt;'))] xsl:copy unescape xsl:copy-of select=node()/ /unescape /xsl:copy /xsl:template xsl:template match=@*|node() priority=-1 xsl:copyxsl:apply-templates select=@*|node()//xsl:copy /xsl:template Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Redirect to GET, strange bug
[EMAIL PROTECTED] wrote:
FYI: Cocoon's redirects usually escape from the current processing.
The redirect is sent to the browser, but the request may not have
been processed. That does not appear to be what you need.
I don't quite understand what you mean here.
The Redirect from submitting a Form cannot be internal to Cocoon (or
any web engine).
Of course.
In fact I'm issuing a global, client-side redirect (no cocoon:
protocol and a true isGlobal argument to cocoon.redirectTo) and that
part works flawlessly, regardless of the errors I get in my logs. The
problem is, what is causing those errors?
Vadim Gritsenko wrote:
This redirect URL does not seem to be correct: should start with '?'
if passing continuation id as parameter or have '/' instead of '='
if passing continuation id in URL path.
We all know URLs can be pretty arbitrary in Cocoon... this is not even
what my url looks like, it was just an example.
Hm, what is this exit() method? I don't see it defined neither in
Cocoon 2.1 nor 2.2. To terminate script processing and send
redirect to browser you should be using suicide method.
AFAIK cocoon.exit() is the recommended way to call FOM_Cocoon.suicide()
It's mentioned here http://cocoon.apache.org/2.1/userdocs/flow/api.html#exit
(while suicide isn't) and its implementation is:
FOM_Cocoon.prototype.exit = function() {
FOM_Cocoon.suicide();
}
In fact, using FOM_Cocoon.suicide() yields the same results as
cocoon.exit(), that is, ProcessingException: Attempted to process
incomplete pipeline.
Do you have any idea what's causing this message?
Who is attempting to process which incomplete pipeline? I'm issuing a
global, client-side redirect, and stopping the flow. Why is Cocoon
complaining?
Also, it is often a good idea to invalidate the form processing
continuation tree, to free up memory and--even if user manages to
submit the form again--he will receive 404.
Interesting. How do I do this?
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Multiple or single cocoon instances
Hi In our (limited) experience, here are pros and cons of using a single instance: + better use of server resources (RAM, CPU) this is quite important; + single global configuration (servlet container, Java VM, Apache or other frontend...) + being forced to deploy global updates (to Cocoon configuration, custom components, etc.) on all apps (projects), this is good for security and performance updates; - slower, more expensive global updates (you have to test all apps on the same Cocoon) - sometimes you have to go through loops to introduce new features while maintaining compatibility with older apps. All in all, we are quite happy with a hybrid approach: we keep all apps in the same Cocoon, until we need such new features or such a different Cocoon configuration to warrant a branch to a new major version of our configuration. So we have something like 30 apps split among 3 Cocoons. The next update to one of the older projects will trigger a switch to the newest Cocoon configuration, and so on. HTH Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
NullPointerException without a stacktrace
Does anybody occasionally get a NullPointerException in the logs,
without a matching stacktrace?
Like this:
$date ERROR $url [sitemap.handled-errors] ErrorHandlerHelper: Sitemap:
error calling function $flowscript_func
at map:call - $sitemap_path
java.lang.NullPointerException
And that's it, no Java stacktrace, no flowscript line number,
nothing. I checked Jetty's standard error too, nothing there.
I'm using Cocoon 2.1.10, and I have %{rootThrowable} in logkit.xconf
(which is the default.) Do you think %{throwable} would make any
difference? I should mention that this only happens on the production
server (of course...) so I can't debug as much as I'd like to.
Could this be a bug in one of my components? I mean, can a bug in a
custom component (I've written actions, readers, transformers, and a
input module) cause a bare NPE, without a stacktrace? Can a bug in
flowscript (such as calling null Java objects, or calling them the
wrong way) cause it?
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: AJAX forms: returning to previous page after submit using Apples
Victor Oomens wrote: How can I return to another page after a user completes an ajax form? Put a cocoon.redirectTo(...) after your showForm See http://cocoon.apache.org/2.1/userdocs/flow/api.html Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Sitemap pipelines
Alec Bickerton wrote: I need the pipeline to return to the pipeline in the global sitemap if nothing matches so that the default matchers will get hit. You need to issue an internal redirect to the parent sitemap. Assuming /default is a (fake) path that in the parent sitemap falls through to the default matcher, without entering ANY sub-pipeline, then you can put this at the end of your sub-sitemaps (last item after all matchers, before closing the last pipeline) map:redirect-to uri=cocoon://default/ The heart of this is the double-slash cocoon: protocol, that will redirect internally (without sending a redirect to the browser) to the root sitemap. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Is there an url wildcard selector (equivalent)
Rainer Pruy wrote: wildcard *matcher* is one of the most used components with cocoon, I'd reckon. But what to do, when one needs to handle X/c/** and X/** (excluding X/c/**) different? At least if the contained components are not final (serialize, read) Cocoon has no URI selector, wildcard or otherwise, probably as some kind of deliberate design choice. This kind of problem is easily solved with Perl-compatible regexps, but Cocoon uses the inefficient and less powerful Jakarta Regexp package (this is a big gripe of mine, btw) probably for hysterical raisins. If it used the native java.util.regex you could match on ^X/c/ and on ^X/(?!c/), but alas you're limited to egrep-style regular expressions, which lack negative lookaheads. If you don't like the nested simple selector solution, or the resources solution, you will have to write your own matcher or selector. Which is not hard at all, by the way. Just extend AbstractPreparableMatcher or AbstractSwitchSelector, put your class in WEB-INF/classes and give it a short name in your sitemap.xconf. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Using xsl variables in javascript
Lincoln Mitchell wrote:
I simple need to convert this inline javascript:
script type=text/javascript
function myFunction(){
xsl:choose
xsl:when test= $myParam='x'
alert('x')
/xsl:when
/xsl:choose
}
/script
…to an external javascript file like:
function myFunction(){
if ($myParam='x'){
alert('x')
}
}
I don't think you have execution times very clear. XSLT code runs on
your server at pipeline execution time, before or during page
generation. Client javascript code runs in the client browser at page
rendering time, after page generation and transmission. This means
that whatever comes out of your pipeline must be well-formed HTML and/
or Javascript that will run on the browser, without access to any XSLT
or Cocoon variables. In fact, by the time the generated HTML reaches
the browser, the variables used in your XSLT won't exist anymore.
Nevertheless, it might be that what you're after is simply this:
script type=text/javascript
var myParam = xsl:value-of select=$myParam/;
/script
Notice the extra quotation marks around the value-of: this XSLT code
is generating a HTML script tag containing Javascript code that, when
run by the browser, will declare a global string variable called
myParam. From that moment on, every piece of Javascript code run on
the browser will have access to that global variable.
Also, this is generic web programming matter, not exactly related to
Cocoon.
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Doing string operations over sitemap values
Nacho (Derecho.com) wrote:
* I have this URL http://localhost:8080/b/menores-de-edad;
* In sitemap i have a match like b/**
* I need to do replace - in {1} to spaces
* I do this using an input module inheriting from
AbstractJXPathModule, and using a xpath like expression,
{request:translate('{1}','-','')}
I would write a custom input module that can be called safely like this:
{translate:-: :{1}}
It can be implemented using basic java.lang.String methods.
I think your solution doesn't quote the argument correctly and is
susceptible to JX code injection or other problems.
For example the user might go to:
http://localhost:8080/b/hello',nasty.java.call(),'world
If I'm not mistaken, the ' after hello would close the Jx string and
damage would ensue.
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Doing string operations over sitemap values
Joerg Heinicke wrote:
Nacho (Derecho.com) wrote:
{request:translate('{1}','-','')}
the user might go to: http://localhost:8080/b/hello',nasty.java.call(),'world
It's JXPath, not JXTemplate. Does it evaluate Java calls at all?
If your example really works, http://localhost:8080/b/nasty.java.call()
will as well.
To be honest I don't know what JXPath can and cannot evaluate
(Java calls? any extensions to XPath? Java objects? request and
session objects?)
I just saw some bad/unusual quotation
{request:translate('{1}','-','')} and I thought I'd alert against
possible problems. But maybe I'm wrong.
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: REPOST: Losing request parameters during form error
Derek Hohls wrote: upgrading on the servers is, unfortunately, a much longer and time- consuming process. I guess I was hoping it was not a bug. Maybe it's not. Can you post a minimal example (simple form with maybe 1 widget) and minimal sitemap that exhibits the problem? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: MySQL Year vs CForms Date
Derek Hohls wrote: if I use MySQL Year column and Cocoon Forms integer type together, then the forms validator complains that the widget is not a date type. This part I don't understand. Are you trying to apply some sort of date widget or date formatting or date validation to the integer field? (You shouldn't.) What component complains that the widget is not a date type? Does a simple integer text field work? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cache Browser
[EMAIL PROTECTED] wrote: The reload (Mozilla) or refresh (MSIE) command should force browsers to ignore cache and retrieve the entire page from the server. This is not always effective. This reminds me, Shift+Reload forces the browser to refresh the entire page, including all dependencies, I believe on both browsers. I use it all the time while developing, so I forgot to mention it. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cache Browser
alexandre mazouz wrote: I have 2 web applcations A and B in localhost with the same adress and ports (127.0.0.1:8080\) in 2 server jetty (J1 and J2). i close A and launch B i have A. The only way to have B is to clean the cache of the browser. You might try to define an action that sets the relevant HTTP headers to prevent caching in the browser: map:action name=no-cache src=org.apache.cocoon.acting.HttpCacheAction/ and call it at the beginning of your pipelines (even before any map:match, if you want it to affect every request): map:act type=no-cache/ Although in your case I would simply change the port of one of your servers, as the browser cache has its uses. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Attempting to Install Cocoon-2.2
[EMAIL PROTECTED] wrote: FILE: mvn.bat == Fix Bad command or filename REM if %HOME% == (set HOME=%HOMEDRIVE%%HOMEPATH%) SET HOME=C:\apache\maven-2.0.8 == Fix Invalid switch REM exit /B %ERROR_CODE% I haven't used Windows much (fortunately) but I have the feeling this script was written for NT's cmd.exe instead of 98's command.com. In any case it seems to be an installation and configuration issue. As you pointed out, Maven is pure Java and should work fine once properly installed. I am teaching HTML, CSS, and XSL (=Cocoon) to people using Win98 PCs. If Maven2 will not run on Win98, neither will Cocoon-2.2. I'll use Cocoon-2.1 for the class. I'm sure I've seen people mention on this mailing list that Maven is by no means necessary to compile, install, or use Cocoon 2.2. It's merely a convenience. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cforms fd:assert validation
Carlos Martínez wrote: I use the language from the page http://cocoondev.org/xreporter/docs/core/exprlang.html and it seems that work properly. Yes, that link comes straight from the Most obscure and valuable subfolder of my Cocoon bookmarks folder :-) Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cforms fd:assert validation
Carlos Martínez wrote: I don't know how the expression of the assert should be. I believe this is the language used in fd:assert, at least in Cocoon 2.1: http://cocoondev.org/xreporter/docs/core/exprlang.html Don't ask me why... Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cforms and Character Encoding
Peter Sparkes wrote: The problem turned out to be in web.xml I amended param-nameform-encoding/param-name param-valueISO-8859-1/param-value To param-nameform-encoding/param-name param-valueUTF-8/param-value and now everything is in UTF-8 and works properly Just a word of caution: make sure you test your forms on every possible web browser. I've had trouble with this setting in the past. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: keypress events
Edward Elhauge wrote: I would like some guidance on what the best way is to attach an 'onkeypress' event handler to a Cocoon CForm Widget? My use case is that I want to set a 'form dirty' flag as users type data into my forms. I've been using the 'fd:on-value-changed' handler for that purpose but it doesn't catch the case where a user types data into a field and then directly clicks on a button that would exit the Web page. I've been having a similar problem: - text field (product quantity) with on-change that triggers update of the totals - on-change fires only when the user leaves the text field - if the user jumps directly to the submit button, the two ajax requests collide and bad things happen (mostly the change to the text field is lost, but I've seen errors too) I haven't solved it yet. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: REPOST: HTML Tags from a database field
Derek Hohls wrote: Tobia wrote: I use the HTMLTransformer for that purpose, with a couple of XSLT stylesheets before and after it. The first one writes unescape tags around each piece of escaped HTML. Then the HTMLTransformer is instructed to only transform text nodes inside those tags. Finally a second piece of XSLT gets rid of the unescape tags along with the fake body and html inserted by the transformer, and does a bit of magic with amp; and lt; characters Whew! Sounds like a lot of work for what I thought would be a simple config issue. Any chance you could make the stylesheets available - maybe upload them to the wiki? It's not much work really. First identify the text nodes that need unescaping and wrap them in unescape tags. For me that was all text nodes under a field tag, containing either an or a : xsl:template match=sql:row/*[not(self::sql:rowset)] xsl:choose xsl:when test=contains(., 'amp') or contains(., 'lt;') unescape xsl:copy-of select=node()/ /unescape /xsl:when xsl:otherwise xsl:copy-of select=node()/ /xsl:otherwise /xsl:choose /xsl:template xsl:template match=@*|node() priority=-1 xsl:copyxsl:apply-templates select=@*|node()//xsl:copy /xsl:template Then put it in the pipeline followed by the HTML transformer: map:transform src=unescape-pre.xsl map:transform type=html map:parameter name=tags value=unescape/ /map:transform map:transform src=unescape-post.xsl The latter XSLT cleans up unnecessary elements: xsl:template match=unescape xsl:apply-templates select=html/body/node()/ /xsl:template xsl:template match=@*|node() priority=-1 xsl:copyxsl:apply-templates select=@*|node()//xsl:copy /xsl:template In the last step I also needed to replace with amp; and with lt; (Java string notation) in all text nodes under unescape, by calling a utility Java function from within XSLT. But that's not an elegant solution and is probably related to my particular pipeline setup, so I won't bother you with the details unless you have the same problem. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Proper way to cache pipelines with SQL Transformer?
Peter Wyngaard wrote: The external pipelines that invoked these internal pipelines with a cocoon: URL passed their request parameters along. This happens because a non-raw internal request (cocoon: schema) does not create a new Request object (or RequestWrapper or something like that) and it is actually quite a useful property. For example, suppose you need to gather some data about the request not explicitly present in the request itself. Maybe you need to issue a set of queries based on the URI path, to obtain one or more IDs of the objects being referred to, or of the user making the request, or some such. Then you would like to store those IDs and misc attributes somewhere and access them during the rest of the request processing, however complex that might be. I find request attributes (don't confuse with parameters) to be an excellent place for this sort of thing. They live in Cocoon's request object, so they have the same lifetime of the initial HTTP request. Which happens to be less than that of the session, but more than the internal sub-request, as an external request might generate many internal redirects and Cincludes, all using the cocoon: schema. Unless you purposefully make a raw sub- request, and until the initial request processing is over, all sub- requests will be able to see the request attributes, because they share the request object. HTH Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: REPOST: HTML Tags from a database field
Derek Hohls wrote: I am using the SQLTransformer from Cocoon 2.1.8 to extract data from text fields in a database. At present, tags are being returned by the transformer (i.e. before any downstream alteration) as lt;tagsgt; I use the HTMLTransformer for that purpose, with a couple of XSLT stylesheets before and after it. The first one writes unescape tags around each piece of escaped HTML. Then the HTMLTransformer is instructed to only transform text nodes inside those tags (by passing it a parameter name=tags value=unescape/.) Finally a second piece of XSLT gets rid of the unescape tags along with the fake body and html inserted by the transformer, and does a bit of magic with amp; and lt; characters. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cached AbstractReader
Jon Evans wrote:
How would I have made it work as a Configurable reader? I had this in
my sitemap:
map:read type=conceptimage mime-type=image/gif src={1}/
map:parameter name=defaultFilename
value=context://images/default.gif/
/map:read
I'm not sure what Configurable is supposed to do (I haven't got around
to studying Avalon yet!) but every Cocoon Reader must implement setup()
(see: http://tinyurl.com/ynkwv2 ) and if the 4th argument to setup() is
called pars, you can just do: pars.getParameter(defaultFilename);
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Converting CRLF to br/ in pipeline
stevecam wrote:
I have text that I am taking from a database that has formatting in
it, mainly as CR and LF characters. I want these to appear as br/
tags in my XHTML output. Is there a standard mechanism to achieve
this, hopefully simply by adding a standard transformer into my
pipeline?
I don't think there's a standard component to do that, but you can embed
it into a XSLT stylesheet, using for example EXSLT string extensions:
xsl:stylesheet ... xmlns:str=http://exslt.org/strings;
...
xsl:for-each select=str:tokenize($your-string, '#xA;#xD;')
xsl:value-of select=./
br/
/xsl:for-each
I haven't much experience using pre tags in html but that may be
another option.
pre, or the CSS equivalent { white-space: pre }, does output a newline
for each CR and/or LF character, but on the other hand won't wrap the
text when there is no CR/LF, so usually it's not an option.
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cached AbstractReader
Jon Evans wrote: I've implemented a class which extends AbstractReader. It uses some SOAP calls to pull some binary gif image data from a remote system. My problem is that the resource is not cached if I throw a MissingResourceException. At the moment, products with no icons cause the remote system to be hit every time, and the SOAP calls are quite slow. The only thing I can think of is to stream the default image data out of my reader if the product has no image, by directly opening the default image file from disk inside the reader. Yes, I think this is the best course of action. You might want to make the path to the default image a configuration parameter, of course, so that you won't have to recompile your reader when the path changes, or if you want different default images for different products. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: JXTemplate: blank lines in source code
Hello Hans-Christian Meier wrote: I have noticed that the use of JXTemplate produces blank lines in the HTML source code. For example the import of the macros to use CForms (jx-macros.xml). In the HTML output the jx:import line is replaced with many blank lines. Is there a way to prevent this? I'd like to know this as well. I'm thinking of replacing the src attribute of every JX transformer and generator with an internal cocoon:/ URL that will strip whitespace-only text nodes from the JX file. But I'm not sure I'll be able to pull it off for JX imports too, and I'd like to know whether this feature is already there in the JX components and I'm just missing a configuration parameter. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Regular Expression Help
Derek Hohls wrote:
I need to check that a valid web address (grammatically correct
anyway) is being entered in a form field. I am trying to check for
the address starting with a valid prefix - the regexp I am using is:
pattern=^(http|https|ftp)://
The regexp validator anchors the pattern by default, as if you had
^ and $ at the ends of the pattern.
For reference, here is a very strict field that only accepts a valid,
non-numeric http URL, without any custom port, username, password or
query string, and checks that it actually resolves and points to an
existing, public web page.
You might want to loosen some of the requirements, depending on your
application :-)
field id=url required=true
datatype base=string/
labelURL:/label
initial-valuehttp:///initial-value
validation
regexp pattern=http://(?:[a-z0-9-]+\.)+[a-z]{2,}(?:/[^?]*)?
failmessagePlease insert the URL of a web page/failmessage
/regexp
javascript
!-- check that the URL exists --
try {
var conn = new java.net.URL(widget.getValue()).openConnection()
conn.setInstanceFollowRedirects(true)
conn.setConnectTimeout(1)
conn.setReadTimeout(5000)
conn.connect()
if (conn.getResponseCode() != 200) throw 'error'
} catch (e) {
widget.setValidationError(
new ValidationError(Please insert an existing URL, false))
return false
}
return true
/javascript
/validation
/field
HTH
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: What is COB-INF?
dperez wrote: Jeroen Reijn wrote: You can find that on the website: http://cocoon.apache.org/798_1_1.html Arrrgh, another web page I don't know how to reach through the menus! ??? It's the first page shown when you click on Getting involved / Committer's and contributor's section. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: different Stylesheets to be applied within a folder
[EMAIL PROTECTED] wrote: Within a Document a stylesheet reference was set as: Document ... stylesheetmyStyleA/stylesheet ... /Document How could I apply a certain stylesheet ? This seems to be another example of content-based pipeline, which is not directly supported in Cocoon. See this recent thread for a few ideas: http://marc.info/?l=xml-cocoon-usersm=119607094225772 Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cocoon and content-negotiation
Severin Gehwolf wrote: select a particular XML file with the best-matching language extension according to users Accept-Language HTTP headers. This is usually achieved in Cocoon using the LocaleAction: http://cocoon.apache.org/2.1/apidocs/org/apache/cocoon/acting/LocaleAction.html Other kinds of content negotiation, such as detecting patterns in the HTTP Accept header, can be performed with the RegexpHeaderSelector: http://cocoon.apache.org/2.1/apidocs/org/apache/cocoon/selection/RegexpHeaderSelector.html or similar selectors and matchers that access the HTTP headers. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: xpath condition on xml data in sitemap
Sébastien Geindre wrote: i need to make a conditional treatment in my site generator transformer A if condition on xml from transformerA -- transformer B1 else -- transformer B2 transformer C serializer the condition is computed on xml data and global variables What you ask for is not compatible with the way Cocoon works. It is my understanding that Cocoon pipelines work like this: 1. a HTTP request reaches Cocoon; 2. the request is dispatched to the right (sub-)sitemap; 3. the tree of all matchers, selectors and actions in that sitemap is executed, meaning that each of them is asked (in hierarchical order) to make its decision and optionally perform any side effects. What comes out of step 3. is a linear sequence of components, comprising a generator, zero or more transformers and a serializer. (Readers and aggregators are chosen at this time too.) 4. This chain of components is then executed in a streaming fashion, meaning that the serializer can (in theory) start outputting the result document well before the generator is finished*; 5. the result of the serialization is sent as a HTTP response, again in a streaming, progressive way, at least in theory. In any case, streaming or not, it should be clear that you cannot choose what components to use in the pipeline (step 3.) based on the intermediate results of the component chain itself (step 4.) I suggest you either: - make the choice between B1 and B2 based on the request data only, or - merge transformers B1 and B2 in a single transformer, or - always execute both transformers B1 and B2, controlling the function of each of them through appropriate elements in the XML stream, or - delegate the work of B1 and B2 (and possibly of C and serialization) to appropriate sub-pipelines, using either the CInclude transformer, a piece of Flowscript, map:redirect-to, or similar means; remember to use cocoon:/ URLs in the place of http:// whenever you can, as they are optimized for internal sub-requests. It's difficult to offer practical suggestions without knowing what your pipeline does. By the way, this whole email applies to Cocoon 2.1. I don't know how much of it is still true in 2.2. Tobia * A notable exception to this streaming behaviour is the XSL transformer, in the general case anyway, as has been pointed out in other threads. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: New server recommendations?
Peter Flynn wrote: I think the last time I asked, the recommendation was to make httpd proxy Tomcat [...] Does anyone have any specific Dos or Donts about this platform or configuration? My personal experience with Tomcat for use with Cocoon 2.1 has been terrible, both on RHEL and Debian. We ended up using the (quite old) version of Jetty provided with Cocoon itself, along with the cocoon.sh wrapper script and Apache reverse proxy functionality, and it's been working like a charm ever since. The only gotcha has to do with the AJP13 protocol, see here: http://marc.info/?l=xml-cocoon-usersm=119201336509954 We haven't switched to Cocoon 2.2 yet, so I can't say much about it. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Flowscripts - are they cached?
Derek Hohls wrote: on this server it did not seem to have any effect, so I was wondering if there was some other setting I had overlooked. There is a setting in cocoon.xconf that might be related: /cocoon/flow-interpreters/component-instance/reload-scripts Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Flowscripts - are they cached?
Derek Hohls wrote: I already have a setting looking like this: reload-scriptstrue/reload-scripts Is this what you meant? Yes. Well I guess you can try and use the ?cocoon-reload=true request parameter, which is faster than restarting the whole container. You can enable it in WEB-INF/web.xml: init-param param-nameallow-reload/param-name param-valueyes/param-value /init-param Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Flowscripts - are they cached?
Another thing: sometimes I run into missing reloads caused by differences in the system time on servers or workstations. That is, a newer file is not reloaded if Cocoon doesn't see it as newer, based on the system clock. Depending on your setup, you might want to investigate that. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
document() and xsltc
I think I stumbled upon a bug in the XSLTC transfomer regarding the
document() function. Here is a minimal example:
--- sitemap fragment: --
match pattern=test
generate src=foo.xml/
transform src=foo-bar.xsl type=xsltc/
serialize type=xml/
/match
--- foo.xml: ---
foo/
--- bar.xml: ---
bar/
--- foo-bar.xsl: ---
xsl:stylesheet version=1.0 xmlns:xsl=http://www.w3.org/1999/XSL/Transform;
xsl:variable name=foo select=//
xsl:variable name=bar select=document('bar.xml')/
xsl:template match=/
output
fooxsl:value-of select=name($foo/*)//foo
barxsl:value-of select=name($bar/*)//bar
/output
/xsl:template
/xsl:stylesheet
The output I would expect, which is also the output I get using the
default xslt (Xalan-J interpretive) transformer is:
output
foofoo/foo
barbar/bar
/output
But if I use the xsltc (Xalan-J compiled) transformer the output is:
output
foobar/foo
barbar/bar
/output
It would appear that the document() call overwrites the source document
of the transformation, including all existing references to it.
Is this a known bug? Can anybody confirm/reproduce it?
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: limited number of loops with call-template?
Jens Reufsteck wrote:
I'm using a template, which once called from outside starts calling
itself as long as a certain condition is true. But after a certain
amount of calls (1.100 times) it seems stopping to call itself.
I wouldn't write a template that calls itself recursively that many
times, unless I was sure that the implementation is tail-recursive and
that I am writing a tail-recursive template.
Said another way, try and traverse the document using matching (not
named) templates and apply-templates, along with modes if needed.
Based on your scarce snippet, you could try something like this:
xsl:template match=ebene
xsl:if test=text() != preceding-sibling::ebene[1]/text()
!-- we are looking at the first ebene of its kind:
output all the ebene with that value inside a baum --
baum ebene={text()}
xsl:apply-templates select=../ebene[text() = current()/text()]
mode=output/
/baum
/xsl:if
/xsl:template
xsl:template match=ebene mode=output
ebene
...
/ebene
/xsl:template
Of course, I don't have the slightest idea what baums and ebenes are, so
this is probably not what you need :-)
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Compiled XSLT
Ard Schrijvers wrote:
I'm wondering if there is a way to pre-compile XSLT into Java
classes, using the command-line xsltc compiler, put those classes
somewhere in Cocoon's classpath and use them as transformers.
Actually, since Cocoon can already run the cached compiled
stylesheets it shouldn't be that hard to get it to use them in the
first place.
Yes. You can take a look at the excalibur xmlutil in
org.apache.excalibur.xml.xslt.XSLTProcessorImpl and write your own
that for example first looks at compiled version at some place.
Thank you for your suggestion.
I was able to understand how XSLTProcessor, TraxProcessor, and
TraxTransformer relate to one another and I could write my own
xslt-processor-role that loads pre-compiled stylesheets from the
classpath.
Here it is, if anybody is interested or wants to take it from here.
Suggestions are welcome.
Tobia
=== cocoon.xconf ===
component class=ClassLoadingTraxProcessor logger=core.xslt
role=org.apache.excalibur.xml.xslt.XSLTProcessor/clxsltc
parameter name=use-store value=true/
parameter name=transformer-factory
value=org.apache.xalan.xsltc.trax.TransformerFactoryImpl/
/component
=== sitemap.xmap ===
transformer name=clxsltc
src=org.apache.cocoon.transformation.TraxTransformer
xslt-processor-roleclxsltc/xslt-processor-role
/transformer
...
transform src=test.class type=clxsltc/
=== Java patch =
--- cocoon-2.1.10/src/java/org/apache/cocoon/components/xslt/TraxProcessor.java
2006-12-19 12:03:37.0 +0100
+++ ClassLoadingTraxProcessor.java 2007-10-23 18:03:34.0 +0200
@@ -14,7 +14,7 @@
* See the License for the specific language governing permissions and
* limitations under the License.
*/
-package org.apache.cocoon.components.xslt;
+import org.apache.cocoon.components.xslt.*;
import java.io.File;
import java.io.IOException;
@@ -72,6 +72,14 @@
-public class TraxProcessor extends AbstractLogEnabled implements
XSLTProcessor, Serviceable, Initializable, Disposable, Parameterizable,
+public class ClassLoadingTraxProcessor extends AbstractLogEnabled implements
XSLTProcessor, Serviceable, Initializable, Disposable, Parameterizable,
Recyclable, URIResolver {
+
+ private class SourceWrapper implements javax.xml.transform.Source {
+ private String id;
+ public SourceWrapper(String id) { this.id = id; }
+ public String getSystemId() { return id; }
+ public void setSystemId(String id) {}
+ }
+
/** The store service instance */
protected Store m_store;
@@ -208,6 +216,36 @@
TraxErrorListener errorListener = new TraxErrorListener(getLogger(),
stylesheet.getURI());
try{
+
+ if (m_factory instanceof
org.apache.xalan.xsltc.trax.TransformerFactoryImpl
+id.endsWith(.class)) {
+
+ // custom code to load pre-compiled translet
from the classpath
+
+ if(getLogger().isDebugEnabled())
+ getLogger().debug(Loading Templates
from classpath + id);
+
+ m_factory.setAttribute(use-classpath,
Boolean.TRUE);
+
+ final Templates template =
m_factory.newTemplates(new SourceWrapper(id));
+
+ putTemplates(template, stylesheet, id);
+
+ final TransformerHandler handler =
m_factory.newTransformerHandler(template);
+
handler.getTransformer().setErrorListener(errorListener);
+ handler.getTransformer().setURIResolver(this);
+
+ m_factory.setAttribute(use-classpath,
Boolean.FALSE);
+
+ final SourceValidity alwaysValid = new
SourceValidity() {
+ public int isValid() { return 1; }
+ public int isValid(SourceValidity
newValidity) { return 1; }
+ };
+
+ return new
MyTransformerHandlerAndValidity(handler, alwaysValid);
+ // end custom code
+ }
+
if (getLogger().isDebugEnabled()) {
getLogger().debug(Creating new Templates for + id);
}
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Compiled XSLT
Hello, I'm wondering if there is a way to pre-compile XSLT into Java classes, using the command-line xsltc compiler, put those classes somewhere in Cocoon's classpath and use them as transformers. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Compiled XSLT
Ard Schrijvers wrote: I'm wondering if there is a way to pre-compile XSLT into Java classes, using the command-line xsltc compiler, put those classes somewhere in Cocoon's classpath and use them as transformers. your xsl stylesheets are compiled only once in cocoon and then held in memory cache. Unless you set the maxobjects of your default transient store (from top of my head, so might be a little different) extremely low, stylesheets won't have to be recompiled Alas, my problem is not recompilation, but secrecy. I'm working on a proprietary application whose source code my client doesn't wish to release. I'd like to write parts of it in XSLT, pre-compile them into classes and somehow use the classes (translets in Xalan-speak) as transformers. Should I write a custom transformer from scratch that uses the Xalan XSLTC runtime to operate the translets, or is there a better way? Do you foresee problems in integrating the pre-compiled translets with Cocoon's source resolver? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Calling an InputModule from flowscript
I wrote: I'm saving the input module into a global flowscript variable, because it's supposed to be request-independent. Nevermind that. I realized I must acquire the input module from the selector every time, lest I end up with IllegalStateExceptions. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Newbie question: jx:choose/
Marcus Wejderot wrote:
I have tried all kinds of things in @test. What I really want to do
is to check if the variable ${userid} is empty. How to do this?
I usually write: test=${empty(userid)}
empty() is native JX language for testing against most kinds of empty,
such as (IIRC) 0, false, undefined, null, empty string, empty array...
See: http://commons.apache.org/jexl/reference/syntax.html
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Calling an InputModule from flowscript
Joerg Heinicke wrote:
Tobia Conforto wrote:
But I cannot pass null to my own input module in place of the third
parameter (Map objectModel) because I use that objectModel in my
class
Another workaround is to create a map in your flowscript with the
necessary data.
Yes, I had arrived at the same conclusion:
function foo(attr) {
// call an input module that needs access to the request object
var tmp = new java.util.HashMap();
tmp.put('request', cocoon.request);
return fooInputModule.getAttribute(attr, null, tmp);
}
the sample fails to call cocoon.releaseComponent() on the
InputModule selector, after calling getComponent() on it.
the using code should not be aware of the component's implementation
details - and that's why it should always release the component.
Here's how I did it, for the archives. By the way, I'm saving the input
module into a global flowscript variable, because it's supposed to be
request-independent. The fake objectModel map, on the other hand, is
created each time in the foo() function.
// in the global scope: get the input module I will call in foo()
var inputModuleSelector = cocoon.getComponent(InputModule.ROLE + Selector);
try {
var fooInputModule = inputModuleSelector.select(foo);
} finally {
cocoon.releaseComponent(inputModuleSelector);
inputModuleSelector = null;
}
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: .DOC and .PPT modifying
Luiz Antonio Falaguasta Barbosa wrote: Joerg Heinicke wrote: If you tell it to save the files in the XML format instead of the binary format you can use option 2 mentioned below. * since office 2003 : pure xml format, cocoon can handle them since its xml, but I don't know framework base on it Yeah, and use option 2 mentioned below means to use a I don't know framework base on it. The problem is which framework I have to use. Cocoon is the framework. If you save the Office files in XML format, the files are pure XML. In Cocoon you have a lot of tools to parse, transform, combine and serialize pure XML data: I would say that's what Cocoon is all about. I suggest you learn to program in the language XSLT, which you will use to modify the XML data, read the introductory documentation on Cocoon's website and study the samples provided with Cocoon, to learn how to put a simple web-application together. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: .DOC and .PPT modifying
Luiz Antonio Falaguasta Barbosa wrote: using Cocoon includes putting images into the pure XML files (saved by MS Office 2003) too, right? If (and only if) those file formats really are valid XML, there's nothing in them that Cocoon can't handle. You will have to figure out how exactly the images are embedded in the file (I'd wager in CData sections, but I don't have such a file handy) and work to obtain similar results yourself. In the worst case scenario you might have to write a piece of Java code to encode the images as an XML fragment in the same way as the format requires... no big deal. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Calling an InputModule from flowscript
Hello I'm following the advice Grzegorz gave me about coding my custom URL-rewriting logic into an input module and using that input module with the LinkRewriterTransformer, and wherever else I need it. My input module works well in the sitemap, but I don't seem to be able to call it from flowscript. The only references I found are this message from a year ago: http://marc.info/?l=xml-cocoon-usersm=115494972431824 and the file blocks/slide/samples/flow.js They both suggests doing this: importClass(Packages.org.apache.cocoon.components.modules.input.InputModule); var foo = cocoon.getComponent(InputModule.ROLE + Selector).select(foo); var foobar = foo.getAttribute(bar, null, null); where foo is the input module's name as configured in cocoon.xconf. But I cannot pass null to my own input module in place of the third parameter (Map objectModel) because I use that objectModel in my class! Is this the recommended way of accessing input modules from flowscript? If so, how do I pass it the objectModel map? Question #2: Both the post and the sample fail to call cocoon.releaseComponent() on the InputModule selector, after calling getComponent() on it. I was under the impression that you should go to great lengths (try/finally) to release every Avalon component you acquire. Is it not so? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Tabs and CForms
Kamal Bhatt wrote: I would like to create a tab based interface where the user can freely navigate between tabs, without the screen validating and without every field being on the form (only the fields on the tab). Additionally, I would like to validate the form completely validates at the end. Did you figure out what should happen when there is a validation error on a group that's in the background? Just a thought... Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cocoon 2.1 environment objects from a Java XSLT extension
[EMAIL PROTECTED] wrote: XSLT Extensions assist formatting. Data collection belongs in the XML generation phase. Using XSLT Extensions to generate data is a very bad practice. What are the requirements? Let me explain better. I need to translate URIs in a XML document, in the middle of a pipeline, fetching relevant information from a SQL database. In the middle of a pipeline means that there are various other steps before it, so I cannot fetch this SQL data in the initial generator. At first I tried to have a sub-pipeline do the job, using the document() XPath function with a cocoon: schema, but it was too slow. My next (and current) idea is transforming the XML with XSLT, while coding the actual URI transformations (including the SQL queries) as a Java extension function, to be called from XSLT where appropriate. I'm hitting a wall, because I need a DataSourceComponent to be able to make SQL queries, but a Java XSLT extension class doesn't have any access to Cocoon's object model (that I can see.) I cannot pass the DataSourceComponent, or any other interesting object, from the sitemap to my extension function, passing through the XSLT as a parameter, because XSLT parameters get passed from the sitemap as strings. Suggestions are welcome. My last resort is doing everything as a custom transformer, but I'd like to avoid it, as I'm not too keen on coding in Sax what can easily be dealt with in XSLT (changing a bunch of attribute values.) Although it wouldn't be too hard with Sax either... Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: SendMail smart host (was sendmail)
Peter Sparkes wrote: I have to use a Smart host, smtp.bytemark.co.uk. Please is it possible to use a Smart Host with sendmail and if so how. You didn't say whether you're using the Sendmail Transformer, the Sendmail Action, or another thing entirely. Here is how you can configure for the first two. Notice that the transformer wants smtphost, while the action wants smtp-host. map:transform type=sendmail map:parameter name=smtphost value=smtp.bytemark.co.uk/ ... map:act type=sendmail map:parameter name=smtp-host value=smtp.bytemark.co.uk/ ... Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cocoon 2.1 environment objects from a Java XSLT extension
Grzegorz Kossakowski wrote: I'm not sure if it's going to fit in your case but I would suggest to take a look at Cocoon's quite powerful LinkRewriterTransformer[1]. Very interesting, thanks. It seems to fit my case perfectly! Basically all you would need to do is to implement your own input module (quite simple task) and configure transformer so it uses your own input module. So I would put my SQL logic in the input module... makes sense! I strongly agree that using extension functions in your case that access external that is *NOT* a good idea. Very well, lesson taken. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Cocoon 2.1 environment objects from a Java XSLT extension
Hi I'm rewriting some functionality into Java XSLT extensions, for performance reasons. For now I'm just writing a Java class, copying it to WEB-INF/classes and calling its static methods from XSLT as such: xsl:value-of select=myclass:myMethod($whatever) xmlns:myclass=http://xml.apache.org/xalan/java/MyClass/ This works, but I need to access the standard Cocoon environment objects (Request would suffice) from my Java class. Should I pass the whole request object to the stylesheet as a parameter and then pass it on to every invocation of my methods? How do I pass the request object to the stylesheet? I couldn't do it. Is there a better way? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cocoon Installation problems on SunOS
Joerg Heinicke wrote: Derek Hohls wrote: We are trying to install Cocoon 2.1.8 under SunOS (version 5.7). The vanilla version of Cocoon installs and runs fine. However, as soon as we try and add a mySQL driver to the lib directory, and alter the cocoon.xconf file, the Cocoon startup reaches a certain point - and then Tomcat stops Nothing in the archives I can find that addresses this problem except http://marc.info/?l=xml-cocoon-usersm=113356786000252w=2 which was not answered All this sounds really weird. Seems to be only in a very special environment setup where it fails. But since Toby is still around on the list and has not given up with Cocoon he might have figured that out. Uh... this is from two years ago, I don't remember how I solved it. Sorry! Now we are running Cocoon 2.1.10 on the bundled Jetty 4.2.23 (which I recommend over Tomcat, by the way) with MySQL Connector 5.0.6 on the same old server and everything works well. But we did run on Tomcat for a while, so I'm sure that problem was solved, maybe by upgrading Tomcat, Cocoon, the MySQL connector, or by fixing the classpath... Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cocoon Installation problems on SunOS
Derek Hohls wrote:
In the meantime, rather than hacking around with all sorts of
combinations, we are trying to locate and use another server
(running Linux, which seems not to have these sorts of problems)
The server I was having that problem on was running Linux, so I believe
the issue to be OS-independent, but rather caused by some combination of
versions. Or maybe something else entirely which I have forgotten about!
Do you know of any detailed comparison of Tomcat vs Jetty, especially
for configuring and running under Apache?
Sure!
Jetty is much lighter on memory and starts / restarts faster. That's
all that matters to us, especially in development. Which means we are
using it in deployment too, to avoid potential problems.
To connect Apache and Tomcat one usually uses the AJP13 protocol, with
the AJP13 connector, which is supposed to be slightly faster than HTTP.
Unfortunately using Jetty 4.2.23's AJP13 port caused weird problems
(random NullPointerException when using continuations!) so we switched
to Apache's reverse proxy functionality. The configuration is like this:
VirtualHost *:80
ServerName www.yourwebsite.com
UseCanonicalName Off
ProxyPreserveHost On
ProxyPassReverse / http://10.0.0.3:/
ProxyPass / http://10.0.0.3:/
...
/VirtualHost
In this example 10.0.0.3 is the Cocoon server. It could be 127.0.0.1.
Note #1: if the various Proxy* directives don't give you enough of a
fine-grained control over which requests go to Cocoon and which ones
stay in Apache, you can use rewrite rules with the [P] flag.
Note #2: when using reverse proxy, you won't find the remote IP address
(that of the web browser) in request.getRemoteAddr(), but rather in
request.getHeader('X-Forwarded-For')
HTH
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: SendMail
Peter Sparkes wrote: I have put the Sun mail.jar and activation.jar in my WEB-INF/lib and removed removed geronimo-spec-javamail-*.jar and geronimo-spec-activation-*.jar from WEB-INF/lib. However I still get the error message Could not connect to SMTP host: localhost, port: 25 Stupid question, but... are you sure you have a MTA running on localhost? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: .DOC and .PPT modifying
Luiz Antonio Falaguasta Barbosa wrote: Is it possible to use Cocoon to change keys by values into .DOC and .PPT file types? Does anybody could give me some example? I don't have first-hand experience with it, but here are a few ideas. If the file is not too big, you can load it into memory as a string and perform string replacement on it, then write it back to disk or stream it to the user. You can accomplish this with a number of Cocoon technologies, for example with a custom Reader or with a piece of Flowscript. Actually you can do the same with any other web language or web application framework: string substitution is pretty basic stuff. You should check whether those file formats store strings in plain text, which encoding they use, and how well they tolerate string substitution. For example, any substitution will break the format if it employs a CRC or another integrity check; variable-length substitutions will fail if they break record lengths or offsets. It's a hack in any case, and it should be treated as such. I have had good results with the RTF format, which is well understood by most word processing applications, is text-based and allows arbitrary substitutions. On the other hand, the newer office productivity file formats are all XML-based (both OpenDocument and that other Microsoft equivalent) and Cocoon excels at handling XML. You could create (or convert) your document templates into such a format, using specific tags or element IDs as placeholders, or anything else that can easily be replaced with an XSLT stylesheet. Then have Cocoon perform the substitution, zip the resulting XML with other needed resources (such as images) using the ZipArchiveSerializer, and serve it to users. Just a few ideas. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: global:home param local and remote
Lincoln Mitchell wrote: I am building many website and placing them in folders in webbapp 1st I to this locally: http://localhost:/Asite.com.au/ I then move it to the webapp on the server to go live: http://www.Asite.com.au/ The problem is that when I move to webapp on the server the URLs become www.Asite.com.au/Asite.com.au/... So how do you get around this? I like to use custom domain names. If you can configure a local nameserver, setup a .local (or other) fake domain with a * (wildcard) record pointing to your development server. Then you can call www.Asite.com.au.local and handle the request in Cocoon. You just have to put a bit of logic in an action or somewhere, to strip the .local part from request.getServerName() If you cannot set up a local nameserver, you will have to put each fake domain name in your hosts file*, as the hosts file does not support wildcards, as far as I know. Tobia *: /etc/hosts or c:\windows\system32\drivers\etc\hosts - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Parsing HTML entities
Andrew Stevens wrote: Tobia Conforto writes: I cannot change this data source component, therefore I need a transformer to examine every text node in the stream, split it at the fake br tags, substitute them with xhtml:br/ elements, and replace every escaped HTML entity with the relevant Unicode character. We have something similar in our application; I arrange the early part of the pipeline so that the escaped HTML appears within a unique element e.g. some_escaped_htmlLorem ipsum lt;brgt; dolor/some_escaped_html pass it through the html transformer map:transform type=html map:parameter name=tags value=some_escaped_html/ /map:transform and follow that by a small xsl transformation to strip out the some_escaped_html elements and the html body elements that JTidy inserts. Net result, the same SAX stream but with the HTML unescaped and cleaned up so it's well-formed again. Thank you. After extensive testing, turns out this is the best method. It works for any kind of malformed HTML and is efficient enough, provided I put some_escaped_html tags only where they are needed. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: XSLT Transformation error in C2.2
Sébastien Geindre wrote: I try to do this in a xslt stylesheet : xsl:variable name=top select=if(($type='cb') and ($typeName!='wims:statusweatherproduct')) then 'max_fub' else 'top'/ if (...) then ... else ... is not an XPath 1.0 expression! You have a few options. One is to use standard XSLT constructs: xsl:variable name=top xsl:choose xsl:when test=$type='cb' and $typeName!='wims:...'max_fub/xsl:when xsl:otherwisetop/xsl:otherwise /xsl:choose /xsl:variable Another is to define a custom function: func:function name=my:if xsl:param name=test/ xsl:param name=then/ xsl:param name=else select=false()/ xsl:choose xsl:when test=$test func:result select=$then/ /xsl:when xsl:otherwise func:result select=$else/ /xsl:otherwise /xsl:choose /func:function And then use it like this: xsl:variable name=top select=my:if($type='cb' and $typeName!='wims:...', 'max_fub', 'top')/ See here for custom functions: http://exslt.org/func/index.html it works with C2.1 I can't see how. Maybe it's a custom extension of the XSLT processor you use? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Problem with CInclude and nested exceptions
I'm currently experiencing an annoying problem where Cocoon does not log exceptions generated by sub-pipelines called through CInclude. Here is my setup: a pipeline (already nested in other calls) generates a CInclude include element and passes it through the CInclude transformer. The transformer invokes a sub-pipeline via a cocoon:/... url, which executes a function or continuation call. Now, if the flowscript raises an exception, such as a runtime EcmaError, I can't see it anywhere! Here is what I get instead: Exception page title (notice that EcmaError isn't mentioned anywhere!) javax.xml.transform.TransformerException: javax.xml.transform.TransformerException: javax.xml.transform.TransformerException: Exception in CIncludeTransformer Java full stacktrace (abridged) org.apache.cocoon.ProcessingException: Sitemap: error calling function 'B' Caused by: org.mozilla.javascript.WrappedException: Wrapped org.apache.cocoon.ProcessingException: Sitemap: error calling function 'A' Caused by: org.apache.cocoon.ProcessingException: Sitemap: error calling function 'A' Caused by: org.mozilla.javascript.WrappedException: Wrapped org.apache.cocoon.ProcessingException: Sitemap: error calling function 'C' Caused by: org.apache.cocoon.ProcessingException: Sitemap: error calling function 'C' Caused by: org.mozilla.javascript.WrappedException: Wrapped org.apache.cocoon.ProcessingException: Failed to process pipeline Caused by: org.apache.cocoon.ProcessingException: Failed to process pipeline Caused by: org.apache.commons.lang.exception.NestableRuntimeException: javax.xml.transform.TransformerException: javax.xml.transform.TransformerException: javax.xml.transform.TransformerException: Exception in CIncludeTransformer Caused by: javax.xml.transform.TransformerException: javax.xml.transform.TransformerException: javax.xml.transform.TransformerException: Exception in CIncludeTransformer The included flowscript function that raises the error is neither A, B, nor C. Those are higher-level functions that call one another before getting to the pipeline with the CInclude transformer. Why their presence should cause so much exception noise is beyond me. Am I losing exceptions because of some nesting limit? If that is so, can I increase that limit? Otherwise, how can I get the CInclude transformer to pass along the exception, so that I can see why the inner pipeline is failing, at least in the logs? I don't recall having this problem before, so it could very well be that I've reached a nesting limit. Any help will be appreciated! Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
mod_gzip -like functionality?
Dear all, While trying to optimize execution and load time for a Cocoon web application, I noticed that AJAX forms require a 173K dojo.js file. This file is only transferred once per session or less, and then cached, but I'd like to optimize the first load time among other things. Therefore I would like to serve a compressed version of this file to browsers that support content compression, much like mod_gzip does. Can I enable such a feature in Cocoon? Ad hoc, manual solutions (such as pre-compression) are welcome too, as I will only do this for a few select static files (like dojo.js) Is my only (or best) option to do this on the front-end Apache server that I'm already using? TIA Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: xhtml serialization and form textarea
Peter Sparkes wrote: If I serialize it as html the output is : textarea .../textarea However when I serialize as XHTML 1.0 Strict the xhtml output is: textarea .../ What you say is true and also happens with div/, script/, and a few other elements. My advice is to avoid the xhtml serializer altogether, as defined in the root sitemap. The problem is that it outputs XML (in fact it uses the XML serializer) but with a text/html mime type, and that confuses browsers. You can either use the xhtml11 serializer, which is correctly defined to output an application/xhtml+xml mime type and that means the output will be parsed as XML; or use the html serializer and work with HTML 4. What I do is use xhtml11 on browsers that support XHTML (nearly all) and html on the legacy ones (IE, wget, lynx): selector name=accept src=org.apache.cocoon.selection.RegexpHeaderSelector pattern name=xhtmlapplication/xhtml\+xml/pattern header-nameAccept/header-name /selector ... select type=accept when test=xhtml serialize type=xhtml11/ /when otherwise transform src=html-fix.xsl/ serialize type=html/ /otherwise /select Notice how it parses the Accept header and thus works on browsers old, new, and yet to come, without the need to do user-agent sniffing. html-fix.xsl adds the following element to head/, which is needed in legacy mode but does harm in XHTML mode: meta http-equiv=Content-Type content=text/html; charset=utf-8 / HTH Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Parsing HTML entities
Hello I have a data source from which I get SAX text nodes into my pipeline that contain escaped HTML entities and br tags. In Java syntax: Lorem ipsum mdash; dolor sit amet. br Consectetuer or, in XML syntax: Lorem ipsum amp;mdash; dolor sit amet. lt;brgt; Consectetuer As you can see, the entities and br tags are escaped and part of the text node. I cannot change this data source component, therefore I need a transformer to examine every text node in the stream, split it at the fake br tags, substitute them with xhtml:br/ elements, and replace every escaped entity with the relevant Unicode character. I tried doing it with the Parser transformer, but it's too slow. I tried using the HTML transformer, but I couldn't get it to work. My question is: what do you suggest I use on the Java side? Is there anything like PHP's html_entity_decode() available somewhere in a library that Cocoon is already using, that can parse and convert HTML 4.0 entities with a single pass on the string? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Parsing HTML entities
Never mind, I solved it by hand
I wrote a Python script that takes a list of HTML entities and generates
a huge tree of switch() { case: switch () { case: switch () { case: ...
The generated Java code goes through a char[] in a single pass and when
it recognizes an entity it pushes the associated Unicode char into the
SAX stream, instead of the chars composing the entity.
It's pretty brutal, it produces a 36k class file, but it's the fastest
thing that could possibly solve the job, short of writing a C extension!
The pattern transformer took 800ms on some data, where mine takes 2ms!
If anybody is interested, I can post or email the code.
Joerg Heinicke wrote:
That's one of the rare cases where I consider
xsl:text disable-output-escaping=yes a valid approach
Yes, that was the first thing I tried, but I discarded it as it was
causing more problems than it solved.
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Java strings vs. Javascript strings
Jason Johnston wrote: So digging around in the Rhino API I discovered the WrapFactory class[1], which is what Rhino uses to wrap Java objects returned from methods so that they can be scripted. By default it wraps java.lang.String objects just like it wraps any other Java object, giving your script access to its Java methods. But it does allow you to switch this behavior for primitive values (strings, numbers, booleans), so that it exposes them as native JavaScript primitives Packages.org.mozilla.javascript.Context.getCurrentContext().getWrapFactory().setJavaPrimitiveWrap(false); Thank you! This is going to be a noticeable time saver for us, by eliminating lots of small errors. IMHO dynamic languages such as JavaScript are less expensive to program in than statically typed ones like Java, at least for small- to medium-sized projects. But only if Things Just Work™, and having to deal with only one class of primitive types is part of it. I'd like to suggest making it the default in Cocoon. Should I post an issue to JIRA or a message to the dev mailing list? Tobia [1] http://www.mozilla.org/rhino/apidocs/org/mozilla/javascript/WrapFactory.html - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Java strings vs. Javascript strings
Hello Am I the only Cocoon user who is extremely annoyed by the fact that Java strings are not Javascript strings are not Java strings? Depending on the context, in your Flowscript you have to use either: str.startsWith(...) str.length() str.codePointAt(n) and the rest of java.lang.String methods, or: str.match(/^.../) str.length str.charCodeAt(n) and the rest of Javascript String methods. This is very annoying and a huge waste of time, at least for me. I found that I can wrap every string in a Javascript String(...) constructor, so the following always work: String(str).match(/^.../) String(str).length String(str).charCodeAt(n) But it's ugly, repetitive, and prone to errors. And don't get me started on testing for a null value or void string! This is giving me headaches. Can't Rhino, or whatever piece of software is bridging Javascript to Java, do the Right Thing and present every string as a Javascript string when I'm programming in Javascript? Can it be configured to do so? Am I doing something wrong? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Java strings vs. Javascript strings
Jason Johnston wrote:
Rhino also makes the JavaScript methods available to Java strings if
the java.lang.String class doesn't already define them. For example:
js javaString.match(/a.*/)
Thanks, that clears part of the confusion to me.
It also explains why sometimes I could get away with treating strings
as the language (JavaScript) calls for, while other times I couldn't.
For example match() works the same on both string types, as the example
says, while replace() doesn't. That's just great.
To complicate matters further, I cannot simply wrap a java.lang.String
in a JS String() constructor and call replace() on the resulting object,
because whenever I handle a java.lang.String I must expect to receive a
Java null, but that becomes a JS null, and String(null) == 'null'!
So:
null is false
is false
java.lang.String() is true (!)
String(java.lang.String()) is false
String(null) is true
Uhh...
I guess I'll have to wrap EVERY Java string in something like this:
function str(javaString) {
if (javaString == null)
return '';
else
return String(javaString);
}
Which means every cocoon.request.get*() call, every string coming from
forms... What a mess!
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: I'm losing my faith in SQL Transformer!
Derek Hohls wrote: I have been frustrated before by the fact that it only nests down to one level; its still not clear to me why further levels of nesting cannot be supported? Can you explain this problem and maybe provide an example? I think I've nested queries up to 4 levels deep without problems, using the SQL Transformer. sql:execute-query rpt:cols sql:query ... I don't think you can intersperse arbitrary tags between the SQL Transformer's tags, but I'm not sure. Glossing over this detail, for the moment, I see that your question has a more serious flaw: Expecting this will return: rpt:report ... sql:rowset rpt:cols sql:row (2) sql:nameDerek/sql:name /sql:row /rpt:cols rpt:matrix sql:rowset nrofrows=0/ (1) /rpt:matrix /sql:rowset /rpt:report This expected result does not make sense. The results of a nested query (1) can only be inside the relevant row of the parent query (2). If you need some other sort of correlation between each row and its child rowset, such as an id attribute, then you will need to do it in XSLT. Otherwise, please post the correct version. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: output indent yes?
Lincoln Mitchell wrote: but doesn't indent my html in the xsl file indentyes/indent does not actually indent the code, it just adds newlines here and there. I personally prefer to leave all output unindented (one long line) and just *view* it indented when I want to. This way I don't need to worry about newlines being inserted where they shouldn't, possibly causing strange layout problems. Here is the way I do it, in case you find it useful. I have configured my browser (Opera) to open my text editor (Vim) whenever I choose view source. Then I have configured Vim so that by typing a shortcut the whole file being displayed is passed through xmllint, a tool that can re-indent XML. The Vim config is: map M-f :%!xmllint --format -CR This has the benefit of alerting me of strange errors, as xmllint is very strict about XML syntax. Another way is to use this Firefox extension: https://addons.mozilla.org/en-US/firefox/addon/697 HTH Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: expire setting
Johnson wrote: Is it possible to set up expired time for the generated pages of cocoon Yes. Take a look at pipeline configuration, here: http://cocoon.apache.org/2.1/userdocs/concepts/caching.html and in the root sitemap.xmap from the Cocoon distribution, lines 370–440. You might also find the Http Cache Action useful: http://cocoon.apache.org/2.1/apidocs/org/apache/cocoon/acting/HttpCacheAction.html If so, we can improve the speed of running cocoon. I'm not sure how that would work, but more power to you! Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Database connection problems during upgrade
Johannes Textor wrote: this is probably because you have not included com.mysql.jdbc.Driver (not sure if this is the exact class name) in web.xml. Or maybe he didn't copy the actual driver class. For example, mine is: WEB-INF/lib/mysql-connector-java-5.0.6-bin.jar Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: access remote jpg
Lincoln Mitchell wrote: img src=http://www.absnetwork.com.au/abs_consult-106079.jpg; img src=http://static-r.arttoday.com/thb/thb13/JU/5366_C00347/ 4559477.thb.jpg?dlsmC0034765 only the 1st image appears, but if I save the HTML cocoon generates and open that html file independent of cocoon the 2nd image also appears! That's because static-r.arttoday.com seems to forbid access to their images from unapproved referrers, but allows access with no referrer. Besides, it's considered very rude to embed media from a different site in your web page without permission, even if the copyright would allow it. I suppose you should fetch the image once only, with a blank referrer header, and serve it to your users from your server. You should also cache it in a database or in the filesystem for future requests. I don't have a clue how to do that in Cocoon :-) Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: access remote jpg
Lincoln Mitchell wrote:
You said: What you probably can do is to make the request to your
Cocoon server and let Cocoon grab it for you with a reader from the
remote server. I am not familiar with the terms you are using so am a
bit confused.
He meant: «Generate an HTML page that will make the client request the
image from your Cocoon server; in Cocoon, handle the request with a
reader, that will download the image from the remote server and send it
to the client.»
Example HTML:
img src=remote/123456.jpeg/
and matcher + reader:
map:match pattern=remote/*.jpeg
map:read src=http://remote.server/blah/blah/{1}.jpeg/
/map:match
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: regexpUriMatcher
Boissé Denis wrote: Usually ungreedy matcher are in a form (.*?) and suppose to match something like temp/hello/ or temp/hi/ but not matching something like temp/hello/dir/ (that's what happen to me) As Alfred said, you cannot use the 'greediness' of quantifiers to control the matching of an anchored regexp, because they will always grow as much as needed to match the string, if possible. You tipically use them to give more weight to one piece of the regexp with respect to another piece. Compare: ^temp/(.*?)/(.*)/$ ^temp/(.*)/(.*?)/$ They will match temp/a/b/c/ as (a, b/c) and (a/b, c) resp. What you described is true when the regexp is not anchored: ^temp/(.*?)/ matches temp/hello/ ^temp/(.*)/ matches temp/hello/dir/ Anyways, I believe Cocoon uses Java's builtin regexp classes: http://java.sun.com/j2se/1.5.0/docs/api/java/util/regex/Pattern.html Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [Flowscript] everything converted to number type?
Mark Lundquist wrote: I've got some CForms widgets of datatype base=long/, so I expect that the widget value should be a java.lang.Long, right? Widget's getValue() returns an Object, so that would make sense, yes. it seems like every reference to the widget value gets converted to a JS number type Probably because all Java number types (including their wrapping objects in java.lang) get converted to a JS Number, which sounds reasonable BTW. I suppose you could find notes on the type mapping between Java and JS on Rhino's website. Rhino is the Javascript interpreter used by Cocoon. Now, if it also converted java.lang.String to JS strings automatically, it would have spared me many a headache... unfortunately that's a conversion that Rhino seems reluctant to do. and truncated to an integer value. I believe that a Java long is an integer type. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Problem with SQLTransformer
warrell harries wrote: From the xml snippet it appears that you have missed out the sql: prefix Actually his namespaces seem to be ok, so I'm not sure what the problem might be... maybe the SQLTransformer code is not prefix-independent? The source XML: sql-select-themes.xml ?xml version=1.0? page execute-query xmlns=http://apache.org/cocoon/SQL/v2; ^^^ querySELECT nom FROM Theme/query /execute-query /page Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: error using java method id flowscript
christoph wrote: I get following error: line 6: Can't find method org.apache.cocoon.components.flow.util.PipelineUtil.processToDOM(string). Maybe because it's processToDOM(String uri, Object viewData) See http://cocoon.apache.org/2.1/apidocs/index.html Or maybe it's something else... Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Disable url rewriting
Does anybody know how to disable url encoding in Cocoon on Jetty? For a certain web application I want sessions to be handled only through cookies (and not to work when cookies aren't available.) Is it a Cocoon or a Jetty configuration parameter? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Disable url rewriting
Let me explain the problem a bit more.
I'm still curious about the issue, although I've found a workaround.
According to the docs Cocoon will not rewrite URLs, unless you use the
encodeURL transformer or call the equivalent flowscript function.
What I'm seeing is that the Location header of a client-side redirect
gets rewritten somewhere between my flowscript call and the network.
I call:
cocoon.redirectTo('test')
And on the wire I see:
Location: http://my.server/test;jsessionid=xx
Who is rewriting the URL? Cocoon or Jetty? How can I disable it?
The workaround I've found is calling redirectTo like this:
cocoon.redirectTo('http://my.server/test', true)
With both a full URL as the first parameter and true as the second one,
the Location URL doesn't get rewritten.
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Disable url rewriting
Nevermind, something was (is) creating spurious sessions where it shouldn't. It will take a while to figure out what misguided component or piece of flowscript is doing so. In the meantime I have put this in a couple places to make sure there is no session before calling redirectTo(): if (cocoon.request.getSession(false)) cocoon.session.invalidate() Otherwise redirectTo, which ends up calling response.encodeRedirectURL() (in HttpEnvironment.java:138) will add ;jsessionid to the URL. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Simple Java from flowscript ?
Schmitz, Jeffrey A wrote: First, is there anywhere to see specifics about the compilation error encountered? All it tells me is Compilation produced 1 syntax error on line 1. Check your cocoon.log file (it's in WEB-INF/logs/ if you didn't change its location; you can change its location in WEB-INF/logkit.xconf) Look for the useless error message about Compilation produced 1 syntax error and a couple lines above you will find the real message with error type, line and column. any ideas on what may be wrong? var uri = neutralInstModel / + inputType + In/ + inputFile + ... ^ :-) Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Client-side cache handling
Hi all Is there an established way, recommended way, ad-hoc component or sample about manually controlling the client-side cache? I'm serving the css for some pages as a unified, space-stripped, virtual css file. This speeds up page loading a bit, especially when the original css files are many and small. Unfortunately it means that Cocoon won't send a Last-Modified header anymore (it seems to me that only readers do so), the browser won't put If-Modified-Since in the following requests and the css bundle will be downloaded time and again. I'm looking for a way to tell Cocoon: this pipeline only depends on these resources, so please compute and send a Last-Modified header. Any idea? Should I do it by hand in a piece of flowscript? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Error while evaluating 'cocoon:/{../1}/...not so many levels
Vaduvoiu Tiberiu wrote:
in my parent sitemap I have
map:match pattern=*/*/*/*
map:mount check-reload=yes uri-prefix={1}/{2}/{3}
src={3}/sitemap.xmap /
/map:match
and in my sub-sitemap I have:
map:match pattern=home.htm
map:aggregate element=root label=source
map:part element=menu src=cocoon:/{../1}/{../2}/menu.xml/
/map:aggregate
You can't access sitemap parameters across sitemaps.
I can see two solutions to your problem:
1) take off that uri-prefix={1}/{2}/{3} from the mount command and
modify all matchers of all sub-sitemaps to deal with the extra URL
components;
or
2) modify that particular matcher in the sub-sitemap to match agaist the
full request URI, which you can get for example with {request:requestURI}
The first solution seems more correct, but of course it depends on
your particular sitemap layout.
Also I don't know exactly how to do the latter point.
I'd wager there is a matcher somewhere that you can use with input
modules (the {request:...} thing is an input module.)
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cocoon sessions and cookies
blogger wrote: I am pipelining a b2evolution blog through cocoon, and I am having a problem maintaining the sessions and cookies. I suggest you fire a packet sniffer (my favorites are tcpick in a terminal, followed by Wireshark in a GUI or on Windows) to capture both the traffic between your browser and Cocoon, and between Cocoon and the blog server. This way you can see what's happening on the wire and you can easily spot any problems. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Namespace: behaviour differences
Let's recap. The following two statements are equivalent and are both prefix-independent: xsl:apply-templates select=sq2:[EMAIL PROTECTED] = 'program-view']/ xsl:apply-templates select=*[local-name() = 'rowset' and namespace-uri() = 'http://apache.org/cocoon/SQL/2.0' and @name = 'program-view']/ The only way to make an XPath expression prefix-dependant (which is usually bad) is to use the name() function. Thus the name() function should be avoided whenever possible, using local-name() in its stead, or better yet direct expressions such as self::sq2:rowset. The two statements above should apply to both XML files (which are two equivalent forms of the same XML data) yielding the same results. The former statement is preferable, since it's shorter, it's more easily optimizable (as somebody else pointed out) and you are defining the SQL2 namespace only once, but they should both work on both XML files. Indeed they do: I just tested all four combinations with xsltproc and the result is always the same, with both found page and found rowset It looks like something is broken in your server installation. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Namespace: behaviour differences
Here's an interesting question: is one of the following more efficient than the other? xsl:value-of select=//wh:thing xmlns:wh=http://whatever/ xsl:value-of select=//*[local-name() = 'thing' and namespace-uri() = 'http://whatever']/ I'm talking about runtime efficiency, not compilation speed (assuming stylesheets are compiled in some way, which I believe is the case.) Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Access to bizData in fd:on-actionfd:javascript
Dev at weitling wrote:
in an on-action-javascript in my forms definition I want to get access
to the bizData provided with the showForm()-function.
In the end I want to be able to change the entries in a selection list
on an ActionEvent.
It might be just personal taste, but I tend to write all event handlers
and widget manipulation in the flowscript, so I would write it as such:
--- flowscript -
var selist = ...
var form = new Form(definitions/test2.xml);
form.getChild('mySelectionList').setSelectionList(selist, 'value', 'label');
form.getChild('changeSelectionListButton').addActionListener(function() {
// selist.blah() ...
// remember to call setSelectionList again here
});
form.showForm(templates/test2.xml);
--- definition -
fd:field id=mySelectionList
fd:datatype base=string/
/fd:field
fd:action id=changeSelectionListButton
fd:labelDo it!/fd:label
/fd:action
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: The continuation ID is invalid
Christian Sengstock wrote: What happens is: the ID is not changing anymore! It's the same for all users requesting the form with a given browser. We restarted tomcat .. still the same problem We cleared the work dir .. still the same problem We restared the machine .. the form is working again So the problem lies in another process, external to Tomcat... a process which is wrongly caching the pages generated by Cocoon and serving them to the users, with stale continuation ids. Apache? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: howto call xsp from xsp
robf wrote: At some point in an xsp I want to include the resultset from calling another xsp, just like it is possible in xslt's This should work: xsp:page language=java xmlns:xsp=http://apache.org/xsp; xmlns:util=http://apache.org/xsp/util/2.0; . . . util:get-source util:hrefcocoon://someurl?esql:get-string column=XYZ//tagx/util:href /util:get-source The util logicsheet has a few useful functions, but I don't think it's well documented. You can look at the sources though (look for util.xsl) Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Calling map:handle-errors from flowscript
Joerg Heinicke wrote: Tobia Conforto wrote: throw new Packages.org.apache.cocoon.ResourceNotFoundException() produces the following Java stacktrace: org.mozilla.javascript.JavaScriptException Caused by: org.mozilla.javascript.WrappedException Caused by: org.apache.cocoon.ProcessingException Caused by: org.mozilla.javascript.JavaScriptException This looks familiar: https://issues.apache.org/jira/browse/COCOON-1205. The exceptions are not kept, but swallowed. Can you confirm this? It doesn't look like the same thing to me, but I'm not sure. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Calling map:handle-errors from flowscript
Joerg Heinicke wrote: Tobia wrote: cocoon.sendStatus(404) cocoon.response.setStatus(404) cocoon.response.sendError(404) throw new Packages.java.io.FileNotFoundException() throw new Packages.org.apache.cocoon.ResourceNotFoundException() The problem with the (send|set)(Status|Error) is that they don't invoke the exception handlers defined in the sitemap. The problem with throwing new exceptions is that they are wrapped in a JavaScriptException and thus are reported as a 500 error, while I need a 404. Did you have a look at the documentation? It talks about unrolling an exception. Does this work? It doesn't seem to work, or I can't get it to work. Throwing a new ResourceNotFoundException produces the following Java stacktrace: org.mozilla.javascript.JavaScriptException Caused by: org.mozilla.javascript.WrappedException Caused by: org.apache.cocoon.ProcessingException Caused by: org.mozilla.javascript.JavaScriptException The exception selector is configured to unroll as much as possible, but the JavaScriptException seems to be unrollable. Anyways, I solved the problem using the XPathExceptionSelector. Flowscript: throw 'not-found' Sitemap: selector name=exception src=org.apache.cocoon.selection.XPathExceptionSelector exception name=not-found class=org.apache.cocoon.ResourceNotFoundException/ exception name=flowscript class=org.mozilla.javascript.JavaScriptException xpath name=not-found test=starts-with(message, 'not-found ')/ /exception exception class=java.lang.Throwable unroll=true/ /selector Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Basic error handling
Derek Hohls wrote: right at the end of the sitemap, after *all* of the map:pipeline sections I have: map:handle-errors map:select type=exception map:when test=not-found map:generate type=file src=docs/missing.xml/ map:call resource=simple-page2html/ /map:when /map:select /map:handle-errors /map:pipelines /map:sitemap I believe handle-errors goes inside a pipeline. Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [Jexl] linebreaks in jexl expressions
Christoph Gaffga (triplemind.com) wrote:
But the encoding of the linebreak in JX Generator expressions seems to
be a problem:
p${value.indexOf('\n')}/p
p${value.indexOf('#10;')}/p
Did you try all the naive workarounds?
${value.indexOf('\n')}
${value.indexOf('\\n')}
${value.indexOf('\\\n')}
${value.indexOf('n')}
${value.indexOf('#10;')}
${value.indexOf('amp;#10;')}
${value.indexOf('amp;amp;#10;')}
${value.indexOf('\#10;')}
${value.indexOf('\\#10;')}
${value.indexOf('\amp;#10;')}
...
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: serialize type question
Vaduvoiu Tiberiu wrote: my serializer: map:serializer name=xhtml ... Here you are defining a serializer called xhtml. I have a question: if my serialize type is the type xml, my resulted xml file has encoding ISO-8859-1. How can I change this to utf-8? ... map:serialize type=xml/ Here you are using a serializer called xml, that is probably defined in the root sitemap. They are two different serializers, defined in different places, that do different things. Did you try Jasha's suggestion? Tobia - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cocoon database access strategy
Here's a followup for the fellow users who build their queries with XSLT.
I found the question-mark + sql:in-parameter syntax a bit inconvenient,
especially when the query is composed out of building blocks in XSLT:
sql:query
SELECT ...
FROM ...
WHERE ...
xsl:if test=$name and $filter-name
AND name = ?
/xsl:if
/sql:query
xsl:if test=$name and $filter-name
sql:in-parameter nr=1 value={$name}/
/xsl:if
As you can see, this approach is a maintainability nightmare: think of a
50 line query with dozens of parameters.
So I came up with this solution, which works very well for me.
It's somewhat similar in usage to esql:parameter:
sql:query
SELECT ...
FROM ...
WHERE ...
xsl:if test=$name and $filter-name
AND name = sql:param value={$name}/
/xsl:if
/sql:query
This is processed by an additional XSLT transformer before being fed to
the SQL transformer:
xsl:template match=sql:query
xsl:copy
xsl:apply-templates select=@*|node()/
/xsl:copy
xsl:for-each select=.//sql:param
sql:in-parameter value=[EMAIL PROTECTED] nr={position()}/
/xsl:for-each
/xsl:template
xsl:template match=sql:param
xsl:text?/xsl:text
/xsl:template
I think it would be useful to include this functionality in the SQL
transformer itself.
HTH
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Cocoon database access strategy
[EMAIL PROTECTED] wrote:
xsl:template match=sql:query
xsl:copy
xsl:apply-templates select=@*|node()/
/xsl:copy
xsl:for-each select=.//sql:param
sql:in-parameter value=[EMAIL PROTECTED] nr={position()}/
/xsl:for-each
/xsl:template
xsl:template match=sql:param
xsl:text?/xsl:text
/xsl:template
I like it. Add priority=2 to the first match to make certain
sql:in-parameter elements are created before the sql:param are changed
to question marks.
From my understanding of XSLT it doesn't change sql:param's to ?s in
the source tree, but only in the result tree, so the processing order
is irrelevant. Actually there is no predefined processing order: XSLT
is (mostly) a declarative programming language free from side effects.
Tobia
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
