Just to give you an update...
I managed to work around this issue using a hacked version of
org.apache.hivemind.util.ClasspathResource. Since I already had to hack
another part of Hivemind in order for 'it' to work on
OC4J9.0.4/10.1.2.x.x, I thought it best to centralize these 'hacks'.
public
Checking the (Tapestry) code:
public IAsset createAbsoluteAsset(String path, Locale locale, Location
location)
{
Resource base = new ClasspathResource(_classResolver, path);
Resource localized = _localizer.findLocalization(base, locale);
if (localized == null)
Digging deeper, and deeper...
In DefaultClassResolver this method:
public URL getResource(String name)
{
String stripped = removeLeadingSlash(name);
URL result = _loader.getResource(stripped);
return result;
}
returns null for 'dojo/' ...
smells like a class
Yep. Sounds like oracle is messing with your classloader. I love oracle
java software. I love oracle java software.
On 5/30/07, Jan Vissers [EMAIL PROTECTED] wrote:
Digging deeper, and deeper...
In DefaultClassResolver this method:
public URL getResource(String name)
{
String
Well - okay then...
let me rephrase my question - is there any known record of people using
T4.1.1 on OC4J 9.0.4 or OC4J 10.1.2.xxx
Thx,
-J.
Jesse Kuhnert wrote:
Yep. Sounds like oracle is messing with your classloader. I love
oracle
java software. I love oracle java software.
On
The ClassLoader held by org.apache.hivemind.impl.DefaultClassResolver
has this:
[ClassLoader:
[[D:\ora\ocj10_1_2_0_2\j2ee\home\applications\eva\eva-web-1.0.0\WEB-INF\classes],
[D:\ora\ocj10_1_2_0_2\j2ee\home\applications\eva\eva-web-1.0.0\WEB-INF\lib\backport-util-concurrent-4.1.1.jar
I'm not sure. Only the wonderful oracle java developers know for sure I'm
guessing.
You can try messing with that or searching oracle related things for
configuring / changing the classloader settings. Tapestry will ultimately
defer class (and class resource) resolution up to whatever class
Reading specification of getResource() ...
http://java.sun.com/j2se/1.4.2/docs/api/java/lang/ClassLoader.html#getResource(java.lang.String)
Finds the resource with the given name. A resource is some data
(images, audio, text, etc) that can be accessed by class code in a way
that is
I would say that oracle hasn't interpreted the specification strictly but
incorrectly. Ie they are wrong and fuc#$ up - again.
I've been doing this kind of resource loading for years - loading icons and
such from swing applications as well as variety of others. I've managed to
keep oracle out
I've been doing this kind of resource loading for years - loading
icons and
such from swing applications as well as variety of others.
Ok - but these are in fact examples of individual files,
could it be that 'directories' have been interpreted differently by this
implementation?
And... is
10.1.3 returns
code-source:/D:/ora/ocj10_1_3_2_0/j2ee/home/applications/KLPD-Eva/eva-web-1.0.0/WEB-INF/lib/tapestry-framework-4.1.1.jar!/dojo/
for '/dojo' - which in plain english means, it works ok.
I notice that this version of OC4J uses a different classloader:
Good luck. ;)
p.s. You can always just switch to software not written by bad programmers -
like jboss or something. It must work reasonably well if it's good enough
for google.
On 5/30/07, Jan Vissers [EMAIL PROTECTED] wrote:
snipped/
I'm thinking about filing a service request/bug with
You can make the Shell use a custom AjaxShellDelegate, here's the default
implementation
http://svn.apache.org/viewvc/tapestry/tapestry4/tags/4.1.1/tapestry-framework/src/java/org/apache/tapestry/dojo/AjaxShellDelegate.java?view=markup
btw, i'd go for the 4.1.2 version
On 5/30/07, Jan Vissers
Hi,
Thanks for your suggestion. I am wondering whether a custom
AjaxShellDelegate will effectively change the fact that the getResource()
to dojoPath and tapestryPath are called - as this is what causing me
problems at the moment.
Or to put in other words - will this be sufficient to make the
14 matches
Mail list logo
