better is to upload it in our rfe/patch tracker!johanOn 10/12/06, Matej Knopp <[EMAIL PROTECTED]> wrote:
Hi.Send the patch to the dev list, we will look at it and if we like what
we see we will commit it. :)Thanks.-Matej[EMAIL PROTECTED] wrote:> Hi,>>I have made some enhancement to the wick
unit test?On 9/29/06, Martijn Dashorst <[EMAIL PROTECTED]> wrote:
(the mailinglist seems borked again, so I'm sending this again).
I think I stumbled upon a bug in Wicket's fragments + ajax.
Here's the exception I get:
Caused by: java.lang.IndexOutOfBoundsException: Index: 294, Size: 294 at ja
yes i already did know that that this was the problem.This is already discussed in another thread.We need to make it a bit better (that attribute map)johanOn 10/2/06,
Juergen Donnerstag <[EMAIL PROTECTED]> wrote:
Found the problem. Johan added markupattributes to Component which ofcourse is not up
wohoooOn 10/3/06, Igor Vaynberg <[EMAIL PROTECTED]> wrote:
test
-Take Surveys. Earn Cash. Influence the Future of ITJoin SourceForge.net's Techsay panel and you'll get the chance to share youropinions on IT & business topics t
Are you saying that you have an application that runs on 1.2.2 and when you have the current wicket-1.x branch compiled you get that error?That shouldn't happen.If you upgrade to trunk, thats 2.0, then this could happen and you won't be able to just shutdown tomcat (saving all the sessions)
and th
is your app deployed under the context helloworld?Do you have an index.html in the root that points to the helloworld/app (app == servletname) with a meta redirect?johan
On 10/3/06, Lurlay, Patrick <[EMAIL PROTECTED]> wrote:
All, I am a wicket nubie. I just wrote a sample app deploy the war unde
yes this can't be done. I am looking how to over come this but we need the request to calculatethe rootpath if it isn't calculated yet.That is because of the servlet spec. You can't for example ask or get the context path
in your Servlet.init method.. That is just not possible. (they solve this in
re you walk the model object and create a property map so that we can have working model object sharing.-Igor
On 9/26/06,
Johan Compagner <[EMAIL PROTECTED]> wrote:
do remind me, what is that model walking thing?I did fix a few things. But what things are now still open things? You
r Vaynberg <[EMAIL PROTECTED]> wrote:
do you think you will have the new model walking thing done soon? would be nice to have that in the alpha and let it get lots of exposure.-Igor
On 9/26/06,
Johan Compagner <[EMAIL PROTECTED]> wrote:
Yes put an alpha as soon as possible out of the d
Yes put an alpha as soon as possible out of the doorThings change after that but that is the idea.johanOn 9/26/06, Igor Vaynberg <
[EMAIL PROTECTED]> wrote:after juergen and i (but mostly juergen) figure out this border+ialternateparent thing i think all the major functionalities will be there.
-Ig
If you just point out to use the places we do cache to much and that you can't replace somehowplease let us know.johanOn 9/24/06, Niclas Hedhman
<[EMAIL PROTECTED]> wrote:
On 9/24/06, Eelco Hillenius <[EMAIL PROTECTED]
> wrote:
Niclas, as you can see, we would be open for it, but not too soon. Iop
we still have the problem of looking up the message at render time everytime.I just want an option to do it once. Just with the markup loading. No components what so ever are touched/added or needed. As long as we also have that option it is fine by me.
johanOn 9/23/06, Juergen Donnerstag <[EMAIL P
I think behaviours/attribute modifiers could be stateless or reusedas long as it doesn't do something or depends in this method:IBehaviour void bind(Component component);johan
On 9/22/06, Eelco Hillenius <[EMAIL PROTECTED]> wrote:
On 9/22/06, Juergen Donnerstag <[EMAIL PROTECTED]> wrote:> What i
icket:message attributes.You do have that option already. It is trivial to add a flag toWicketMessageHandler or create different ones to control it defaultbehavior: resolve or attach AM
JuergenOn 9/22/06, Johan Compagner <[EMAIL PROTECTED]> wrote:> there is one other problem> Our localizer o
A Page is stateless just as before in wicket 1.2, there are no callbacks to the page.But we have now a nice option to reduce those callbacks: /** * Returns whether the component can be stateless. Being able to be
* stateless doesn't necessary mean, that the component should be stateless
teless attributemodifier tothe component tag and that AM gets automatically attached to thecomponent at render time. That would mean basically no overhead at
runtime.Because the very same AM gets re-used for all compent tags withwicket:message attribute it must be stateless.JuergenOn 9/19/06, Johan Com
As i said before. we can apply this patch, but there is one problemand that is OSGI wicket solutions.. That do jar updates by just dropping them in a running systemHow are we going to test that?Can you test something else? ( i also will try to test that)
comment out this code:try
so every component will go through a Implementation of thatthat can be registered in the settings?phhOn 9/21/06, Eelco Hillenius <
[EMAIL PROTECTED]> wrote:Regarding the discussion here:
http://www.nabble.com/localized-attributes-tf2288202.html#a6393222What do you think about an interface like IOnC
erful.So, if we can find a reasonable way to support this, I'd be all for it.
EelcoOn 9/19/06, Johan Compagner <[EMAIL PROTECTED]> wrote:> what we currently do is over complex and in all my usecases i have totally> not needed
> but maybe in my cases i had where real i18n was needed it wa
ue.The point Eelco was making earlier is the inconsistency between and wicket:message. The former one *is* resolved at
render time and hence has all the component information at hand.wicket:message in contrast has not.JuergenOn 9/19/06, Johan Compagner <[EMAIL PROTECTED]
> wrote:> i think i
point Eelco was making earlier is the inconsistency between and wicket:message. The former one *is* resolved at
render time and hence has all the component information at hand.wicket:message in contrast has not.JuergenOn 9/19/06, Johan Compagner <[EMAIL PROTECTED]
> wrote:> i think i st
omething very basic with />. I don't want to explain this to users, and I don't like theinconsistency. I'm for the best solution, not the easiest for us toimplement or maintain.EelcoOn 9/19/06, Johan Compagner <
[EMAIL PROTECTED]> wrote:> i think i still prefer the doin
i think i still prefer the doing it once by loading the markup.The maybe we loose some advanced localization but i find this acceptable.What options do we then loose?If these options work then it is fine:
Xxxx.htmlgets its messages from:Xxxx.propertiesMyApplicaiton.propertiesand there own Localizat
Then we need to look at all the FormComponent visitors (validation/input conversion finding the submit button)because we need to stop add the form we could then encounter.On 9/17/06,
SourceForge.net <[EMAIL PROTECTED]> wrote:
Feature Requests item #1560074, was opened at 2006-09-17 11:56Message ge
wicket 1.x is still 1.2.3 as far as i know.On 9/16/06, Juergen Donnerstag <[EMAIL PROTECTED]
> wrote:forget about it, everything is okJuergenOn 9/16/06, Juergen Donnerstag <
[EMAIL PROTECTED]> wrote:> Provided I didn't misunderstand it, than wicket-1.x is the development> branch for 1.3. I checked
serve the new file.So just not testing any kind of jar will break stuff.johanOn 9/16/06,
Jean-Baptiste Quenot <[EMAIL PROTECTED]> wrote:
* Johan Compagner:> First of all you also should do that in the lastModifiedTime> call of the URLResourceStream class.Right, thanks for pointing
debyt openConnection()So i guess we should do that in 2 steps.. First openConnection or self. Keep that in an internal reference just like
inputstream. And also try to close that if we can by testing if it is a httpurlconnection or a jarurlconnection..can you test that if that works?johanOn 9/16/06,
Let m know if that helps for your situation.if really a jar url connection is returned in your case. Because this doesn't have to be the case. Any kind of UrlConnection can be returned (WebLogic does have its own ZipUrlConnection or something like that)
In the end this is a problem of the ClassLoad
ld become:public void wrapOnAssignment(Component component) {setComponent(component);}This saves the extra model (== memory). But I might be wrong of course.
MartijnOn 9/15/06, Johan Compagner <[EMAIL PROTECTED]> wrote:> i don't think we can remove it completely. At specific points i have to kno
johan, you are back?>> my gut tells me we should start with IModel and then if we run into a> situation that realy really requires us to know about the wrapper we should
> introduce iwrapmodel back.>> -Igor>>>> On 6/22/06, Johan Compagner <[EMAIL PROTECTED]> wrote:
the CompoundPropertyModel impl/name is right as far as i can seebecause that is a Compound model based on property access.and that is an InheritableModel.ICompoundModel in the javadoc should be renamed.
On 9/14/06, Martijn Dashorst <[EMAIL PROTECTED]> wrote:
I see two different names for the model
and what did stop you to run the test in 2.0 before you commit?johan 'couldn't resist' pain in the ass.On 9/13/06, Igor Vaynberg <
[EMAIL PROTECTED]> wrote:sorrybutton test was there only in 2.0
and i tested the change in 1.0 before porting it to 2.0.-IgorOn 9/13/06,
Juergen Donnerstag <
[EMAIL P
i will look at it this week.On 9/12/06, Igor Vaynberg <[EMAIL PROTECTED]> wrote:
i think there is already a bug reported against it that suggests doing resource.tostring().getbytes().length which should fix this. is someone going to look into this? this isnt exactly my forte
-Igor
On 9/12/06, jan_b
t;[EMAIL PROTECTED]> wrote:
nevermind-IgorOn 9/11/06, Igor Vaynberg <
[EMAIL PROTECTED]> wrote:
if we dont sync on session then the user will have to himself, and that would really suck.-Igor
On 9/11/06, Johan Compagner <
[EMAIL PROTECTED]> wrote:
I am now synchronizing around the pagema
just call:getRequestCycle().urlFor(pageMap, pageClass, parameters);if you want a bookmarkable url for a page class.johanOn 9/11/06,
jan_bar <[EMAIL PROTECTED]> wrote:
Maybe the public getURL() is not needed, because it doesn't work as Iexpected, see user list "Get url of BookmarkablePageLink".Rega
)at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests
(RemoteTestRunner.java:673)at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.run(RemoteTestRunner.java:386)at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.main(RemoteTestRunner.java
:196)JuergenOn 9/11/06, Johan Compagner <[E
7Submitted By: Eelco Hillenius (eelco12)Assigned to: Johan Compagner (joco01)Summary: Infinite Loop on
Session.getPageInitial Comment:From thread:http://www.nabble.com/Infinite-Loop-on-Session.getPage%28%29-tf2233107.html
Here's a trace of what happens that causes the problem :During the RESOL
fine by meOn 9/11/06, Igor Vaynberg <[EMAIL PROTECTED]> wrote:
oh yeah, i also want to move ivalidator and friends out of wicket.markup.html.form.validation into wicket.validation, anyone against? 2.0 only of course.-Igor
On 9/11/06,
Igor Vaynberg <[EMAIL PROTECTED]> wrote:
first pass of this is
I am curious that if it is really needed that a context path just always must end with "/" If that is the case then that patch is not correct because it is very funny that when the servlet path = ""
that you add a "/" to the context path.. That is strange.If the context path just always have to end
Hi,The question did popup again how to redirect to an url (external or internal doesn't matter)We already had a RedirectPage that does a meta tag refresh (can be a delayed one)But i also added a RedirectRequestTarget that can be used now.
Now the question is should we "hide" this target as we do wi
egory: coreGroup: 1.2Status: OpenResolution: NonePriority: 7Submitted By: Eelco Hillenius (eelco12)Assigned to: Johan Compagner (joco01)Summary: Infinite Loop on
Session.getPageInitial Comment:From thread:http://www.nabble.com/Infinite-Loop-on-Session.getPage%28%29-tf2233107.html
Here's a trace of
essage(). if we provide a class then we are really glued to how the class does it - resource keys and interpolation. i am hoping to make this as generic as possible, but do tell me if you think im going overboard unjustifiably.
-IgorOn 9/7/06, Johan Compagner <
[EMAIL PROTECTED]> wrote:
Is IVal
i still think (we need that method anyway) thatparent.switchParent(IMoveable child) is better. (or what ever you wanna call it)It is compile save, instead of runtime and then it is just a marker and they don't have to implement moveTo()
(which is an internal thing in the end anyway)And people can
the attributes yes but also the markupid (when constructed by the completely hierarchy)johanOn 9/7/06, Eelco Hillenius <
[EMAIL PROTECTED]> wrote:On 9/7/06, Igor Vaynberg <
[EMAIL PROTECTED]> wrote:> none of the code youve written so far counts unless are are writing an app> against 2.0 because the
that would be bad. That will nobody implement. Nobody is going to think about that..Then just let it be moveable and be done with it and warn people about the possible dangerous things(like attributes and id/name changes becaues of the different hierarchy)
johanOn 9/7/06, Eelco Hillenius <[EMAIL PR
Is IValidationError really needed?Can't that be a normal class what would the interface really give us?johanOn 9/7/06, Igor Vaynberg <
[EMAIL PROTECTED]> wrote:ok, so
IValidatable { T getValue(); error(IValidationError error); }IValidationError { String getMessage(ILocalizer localizer); }ILocal
[0] IndifferentOn 9/7/06, Eelco Hillenius <[EMAIL PROTECTED]> wrote:
[0] IndifferentEelcoOn 9/7/06, Nick Heudecker <[EMAIL PROTECTED]> wrote:> I'd like to start a vote on this.>> [+1] Change StringValueConversionException to an unchecked exception
> [0] Indifferent> [-1] Leave it as is.>> Here's my
what about the option:parent.addMovedChild(IMoveable)then it is always compile save and you only have to have moved() on the interface.johanOn 9/7/06,
Igor Vaynberg <[EMAIL PROTECTED]> wrote:
because components have access to a lot more information on creation it is not always safe to move them ar
oleans into a masked value with constants, just havent had a chance yet.-Igor
On 9/6/06, Eelco Hillenius
<[EMAIL PROTECTED]> wrote:
Yeah. Looks like a typo to me.EelcoOn 9/6/06, Johan Compagner <[EMAIL PROTECTED]> wrote:
> Why is that this:>> /**> * Gets the ur
i don't like such a onDetached. Then people has to learn to implement that method for that specific component.Just learn to use the super call.Eclipse if you generate the stub auto generates it for you the super call.
So to keep introducing all those methods is horrible. But don't we have a interna
Why is that this: /** * Gets the url that references this handler. * * @return the url that references this handler */ public CharSequence getCallbackUrl() { return getCallbackUrl(true, false);
}calling this method:public final CharSequence getCallbackUrl(final
n 9/6/06, Martijn Dashorst <
[EMAIL PROTECTED]> wrote:And don't add the @OverrideMartijn
On 9/6/06, Johan Compagner <[EMAIL PROTECTED]> wrote:> no doesn't break any binary thing, it should be a drop in replacement
> only the behaviour is slightly difference that the strea
wasn't there talk about a common validation framework?i think it ended up being just that - talk :)start it up again ;)Everybody start the vote:
[X] Igor leading common validation framework[ ] can't choose anything else.
you can do what you want now also pretty easy.Only one extra adapterclass inb
the same for me.The thing is if it is a checked how would you handle it now? Can you handle it more gracefully?If it is unchecked people are getting the normal error page much sooner, because it is very unlikely
that it will be handled.johanOn 9/6/06, Eelco Hillenius <[EMAIL PROTECTED]
> wrote:I'm
:) that wouldn't compile i guess On 9/6/06, Martijn Dashorst <[EMAIL PROTECTED]> wrote:
And don't add the @OverrideMartijnOn 9/6/06, Johan Compagner <
[EMAIL PROTECTED]> wrote:> no doesn't break any binary thing, it should be a drop in replacement> only the beha
all that protection.. who needs it make it public!johanOn 9/5/06, Frank Bille <[EMAIL PROTECTED]
> wrote:[x] Ay[ ] Nai
On 9/5/06, Igor Vaynberg <[EMAIL PROTECTED]
> wrote:
currently form.findsubmittingbutton() is protected final - anyone against making it public final?-IgorOn 9/5/06, Jaime De La J
no doesn't break any binary thing, it should be a drop in replacementonly the behaviour is slightly difference that the streams are getting to be closed on the end of a request (as it should!)i am +1Only this should be changed for the patch:
+ @Override+ protected final void onDetach()+ {+
code(), ctx); if (ctx.hasError()) { throw new ValidationException(ctx); }}-Igor
On 9/5/06,
Johan Compagner <[EMAIL PROTECTED]> wrote:
But you want to sent the FormComponent instances to the service layer? Is the service layer always on the same machine?
(if that is seperated like 1 or
But you want to sent the FormComponent instances to the service layer? Is the service layer always on the same machine?
(if that is seperated like 1 or more Web servers that handle http load and one service server that does the db handling and so on) you have a problemwhen sending over the formcom
fine by me. I altered PropertyResolver and added some test i had to change some behaviour now because testMapSizeLookup() was failing nowbecause findField does find much more then then before.But all test are running fine. Is this wanted in
1.x or just in 2.0?johanOn 9/5/06, Juergen Donnerstag <[E
hmm i am not to happy with yet another getValueXXX method on FormComponent. We should clean that up!So i would map to an existing method (i guess getConvertedValue())But you want to sent the FormComponent instances to the service layer? Is the service layer always on the same machine?
(if that is s
fine by meOn 9/3/06, Igor Vaynberg <[EMAIL PROTECTED]> wrote:
getdefaultchoice actually generates markup, but the name does not express that. renaming to getDefaultChoiceMarkup() will make things much more understandable. we can do only in 2.0 or create a deprecated delegate in
1.3 as well.
-Igor
yeah just wait a week and see where we end before starting to do things in 3 branches.johanOn 8/30/06, Igor Vaynberg <
[EMAIL PROTECTED]> wrote:wicket-1.2.2 is pretty much feature complete, and its time to concentrate on
wicket-1.3working on 1.3 means committing to wicket-1.x but if we do that now
> and that map is created upfront by the page.doRender() or what every render
> call it is (ajax)> So that when streaming the resolving is never be done anymore?>I'd much rather attach the markupfragment to each component in theconstructor (but allow for lazy loading) and at the beginning of
render
hmm i think we should reverse it then (what igor also proposes)we iterate over the components and let the component handle the markup rendering (which it already should have)johan
On 8/29/06, Juergen Donnerstag <[EMAIL PROTECTED]> wrote:
On 8/29/06, Johan Compagner <[EMAIL PROTECTED]>
xcept auto components or something like that)johanOn 8/29/06, Juergen Donnerstag <
[EMAIL PROTECTED]> wrote:On 8/29/06, Johan Compagner <
[EMAIL PROTECTED]> wrote:> yeah from a purely code view or what ever view at it> but in the end, the end result is the same.>> 1> co
> > > > more often than you might think, than first the container and all its
> > > > parents are checked if they implement IComponentResolver. If yes, it> > > > is called and only if not successful it will continue with the search.> > > &g
ent hierarchy all IComponentResolvers registered with the> > application are called as well, until a resolver returned with> > success. Basically the resolvers can do whatever they want and they
> > are not at all limited to finding a component and render it.> >> &
thats fine by me.On 8/28/06, Igor Vaynberg <[EMAIL PROTECTED]> wrote:
why not refactor the icomponentresolver to do what it says - resolve the index of the component in the markup - the rendering can be factored out to places that call resolve()-Igor
On 8/28/06, Johan Compagner <[EMAIL
I have the ajax auto complete behaviour attached to a fieldBut if that field is set to disabled or readonly i want the auto complete also not to be rendered.With a AttributeModifier we have isEnabled() but that isn't in the basic IBehaviour interface is there a reason for this?
johan
--
i think this is the discussion we had (juergen) in another thread.If you write a IComponentResovler you should also write something else to be able to look up the childs components markup.Can't we add another method on IComponentResolver? That the markupfinder can call on the parent for a child?
jo
all fine by me but just curious for one thing:Who is doing that commit to releses/wicket-1.2.2 (after which it is readonly)?Because i think you can have a working copy of the changes you want in that release branch/tag
and commit that to a tag saying that you want to commit the working copy. I don'
On 8/24/06, Matej Knopp <[EMAIL PROTECTED]> wrote:
Johan Compagner wrote:> thats what i just wanted to say.> Maybe if i have the time on sunday (don't have it on friday and saturday)> i will look once more to the completel list of patches and bugs to see
> what still nee
thats what i just wanted to say.Maybe if i have the time on sunday (don't have it on friday and saturday) i will look once more to the completel list of patches and bugs to see what still needs to be done for 1.2.2
Most of those ajax bugs we had (matej) we should fix if possible (matej!) jo
All of them don't worklook at the pattern that we had:Pattern.compile("\\[if .+\\]>(.|\n|\r)*");and what it is now:Pattern.compile
("\\[if .+\\]>(.|\n|\r)*see the --> is gone.That is because the comments we test, so i take the first of below:is always just this[if IE]>@import url(ie.css);
rd and do it the way I seem fit?
MartijnOn 8/24/06, Johan Compagner <[EMAIL PROTECTED]> wrote:> fine by me except that i don't really like this:>>> Wicket 2.0.1 is released. This means:
> * copy branches/WICKET_2_0 to tags/START_RELEASE_WICKET_2_0> > _1> &
fine by me except that i don't really like this:Wicket 2.0.1 is released. This means: * copy branches/WICKET_2_0 to tags/START_RELEASE_WICKET_2_0
_1 * copy tags/START_RELEASE_WICKET_2_0_1 to releases/WICKET_2_0_1 * check out releases/WICKET_2_0_1 in a clean workspace * build release, fix version nu
; not just the latest update.> Category: None> Group: 1.2> Status: Open> Resolution: None> Priority: 5> Submitted By: Johan Compagner (joco01)> Assigned to: Nobody/Anonymous (nobody)
> Summary: FileUpload must be documented better>> Initial Comment:> because we do this:
onent to control markup loading. It renderscontainer.newMarkupResourceStream() almost useless now. Wouldn't it be
better to lazy load the markup when needed or at least to allow forlazy loading if needed?JuergenOn 8/22/06, Johan Compagner <[EMAIL PROTECTED]> wrote:
> i want to stress out tha
ve.
> but before i do this i would like to know if we are going to rip out the> icomponentresolver and replace it with something else or if its going to> stay for good.>> -Igor>>
>> On 8/8/06, Johan Compagner <[EMAIL PROTECTED]> wrote:> >>> But we reall
erg <[EMAIL PROTECTED]
> wrote:i just dont think its a very common usecase to do that...thats all
-IgorOn 8/8/06, Johan Compagner <
[EMAIL PROTECTED]> wrote:thats why i wanted at first that panels where roots (instead of only pages)
because with servoy we switch a lot of panels (to rea
[X] Yes, let's go Apachelets join the old boys club.On 8/21/06, Eelco Hillenius <[EMAIL PROTECTED]
> wrote:Hi all,Some time ago we announced that we started negotiations with Apache to
start an incubation process. That basically means we would be enteringa 'test period' where we proof ourselves to
#x27;m proposing:trunk/wicket-1.2trunk/wicket-1.3
trunk/wicket-2.0trunk/wicket-2.1I think the former communicates the best intent for the repositorystructure: 1.x and 2.x are both under active development, and boththeir HEAD's can be found in trunk.
MartijnOn 8/21/06, Johan Compag
cket 1.x? When should we do
create that branch then? How can we prevent people from committingsomething to 1_2 instead of 1_3?MartijnOn 8/20/06, Johan Compagner <[EMAIL PROTECTED]
> wrote:> I dont know about this>> Where will 1.3 reside? and 2.1 ?> or 3.x ?>> (or better said
we have still a little discussion about them. But i don't think they will change much.The discussion is purely about IWrapModel and the getNestedModel() method of IModel.What to do with both.I think getting rid of IWrapModel will be a bit hard. Because we need to know on some places
that it is a wr
TagModifierContainer ;)On 8/20/06, Eelco Hillenius <[EMAIL PROTECTED]> wrote:
> the counter point being that you are adding components to a TagModifier -> which is a bit weird since TagModifier doesnt really communicate its a> container, but that is the common usecase here right?
You have a point t
also +1 for meand this sandbox approache:/sandbox/ajax_tryoutOn 8/20/06, Eelco Hillenius <
[EMAIL PROTECTED]> wrote:> We should have:>> /trunk <-
2.0> /branches/wicket-1.2 <- 1.2> /releases/wicket-1.2.0> /releases/wicket-1.2.1> /releases/wicket-1.2.2That makes sense to me: +1 on that.Eelco
e.get() and Session.get()? Don't you dare to hidethose! :-)-MatejJohan Compagner wrote:> ohh forgot to mention. I don't like static access methods. Those if> needed (for threadlocals)
> should be hidden as possible.>> johan>>> On 8/20/06, * Johan Compagner* <[EMAIL PRO
ohh forgot to mention. I don't like static access methods. Those if needed (for threadlocals)should be hidden as possible.johanOn 8/20/06,
Johan Compagner <[EMAIL PROTECTED]> wrote:
No we should hide the thread local access. That is prefectly fine.It is much more clear that all the th
to their classes:public static final void RequestCycle.setResponsePage(Page page) {RequestCycle.get().setResponsePage(page);}This way you can do:
public void onSubmit() {RequestCycle.setResponsePage(new MyPage());}MartijnOn 8/19/06, Johan Compagner <[EMAIL PROTECTED]> wrote:> yes if i
I dont know about thisWhere will 1.3 reside? and 2.1 ? or 3.x ?(or better said 2.0 when we have released it)1.2 is just a a branch in my eyes. Ok not really a maintenance branch but something more.
But it is in my eyes still a branchI just like thistrunk == wicket X (the last wicket where we work
See for an example the FormInputApplication in wicket examples.But the simple thing is in the init of your app:getSharedResouces().add("myimage", new XxxxResource());and thenImageButton("mybutton', new ResourceReference("myimage"))
johanOn 8/19/06, [EMAIL PROTECTED] <[EMAIL PROTECTED]
> wrote:Hi ,c
ConversionException not a runtime exception, which i take in this case could be a good thing.johanOn 8/20/06,
Jean-Baptiste Quenot <[EMAIL PROTECTED]> wrote:
* Johan Compagner:> But i think a converter itself should be even simpler like we have now in> 2.0>> IConverter:> Object convertToObject
+ for puttin it into coreIt falls under the category like StatelessLink/Form which also users can easily create themselfs.Only by having such a thing with the naming you can better read the code so you know what it does.
johanOn 8/20/06, Jonathan Locke <[EMAIL PROTECTED]> wrote:
based on the fact t
+1 for me to backportSo when upgrading the cookies are one time invalid and not recognized? That doesn't seem to bad for me. If it really does fix problems for some languagesjohan
On 8/20/06, Juergen Donnerstag <[EMAIL PROTECTED]> wrote:
Related to: [ 1531861 ] CookieValuePersister support more enc
ple situations like these.
but maybe we should ditch the idea.-IgorOn 8/19/06,
Juergen Donnerstag <[EMAIL PROTECTED]
> wrote:What I haven't understood yet is why this map is modifyable? Why isn't
it just a read-only map. What is wrong with using AttributeModifierinstead.JuergenOn
fine by me but we shouldn't go to far with itThings like RequiredTextField which is now only one property...But again it is widely used i think so it serves it purpose.And we now also have StatelessLink/Form (
2.0) that also really overrides one method But those are really handy and it is directly
-something hopefully a lot of people will do - will be relativelypainless.EelcoOn 8/19/06, Igor Vaynberg <
[EMAIL PROTECTED]> wrote:> go ahead, i dare you!>> -Igor>>>> On 8/19/06, Johan Compagner <[EMAIL PROTECTED]> wrote:
> >> > yes if i have to clean up com
hmmmit seems that also my keyboard is not working correctly;)johanOn 8/19/06, Igor Vaynberg <
[EMAIL PROTECTED]> wrote:you mean a keyboard?
-IgorOn 8/19/06, Johan Compagner <
[EMAIL PROTECTED]> wrote:
really? if you press Q it prints A?Strange keyword indeed.johanOn 8/19/06, Frank Bil
really? if you press Q it prints A?Strange keyword indeed.johanOn 8/19/06, Frank Bille <[EMAIL PROTECTED]
> wrote:... and I need to get out of here... it should have been a "Q" and not an "A".
damn keyboardOn 8/19/06, Frank Bille <
[EMAIL PROTECTED]> wrote:And someone has already come up with a be
1 - 100 of 1020 matches
Mail list logo
