[Logo] Enterprise Client Community
  [Search] Search   [Recent Topics] Recent Topics   [Members]  Member Listing   [Groups] Back to home page 
[Register] Register / 
[Login] Login 
Messages posted by: rspiller  XML
Profile for rspiller -> Messages posted by rspiller [59] Go to Page: 1, 2, 3, 4 Next 
Author Message
Hi,

in a list component, selecting an entry and trying to navigate trough entries with arrow-keys is not working.
The down-arrow-key scrolls the list, but does not change the selection.

Bug or feature request?

Expectation for example: Setting focus in the list and arrow-key will pre-select|hover an entry and space will select it. When pre-selection is at the bottom of the list, scrolling will be done for the next entry.

Make sense? ... Honest question.

greetz
Ruven
Hi,

all right!

But a little bit confusing. Setting the value to 90 or 60 both resulted in an actual heartbeat for every two minutes.

Setting it to 50 now generates a heartbeat every minute.
Maybe updating the template help text?

Why am I so picky? A customer has an apache http-deamon in front of the application and the timeout inside is set to 100 seconds.
So every two minutes does not help and therefore we need the heartbeat every minute as lowest possible value. Which is fine - when you know how to get it ;)

greetz
Ruven
Hi,

am I looking in the wrong place or does the configuration not work for websockets?

I set the keepaliveduration to 90 but in the browser-tools I only see a update message every 120 seconds.
At least if I look into the websocket cclongpolling connection an inside there into the message tab.

Wrong expectations or another issue?

greetz
Ruven
Hi folks,

a classical for your information:

We are using the beanprocessing jshowurl component to trigger downloads. The component sends the browser to an URL with the MediaType of the file or application/octet-stream to trigger the actual download to the hdd of the given file.

Recently we discovered a problem within firefox when using the target _self within the jshowurl component when using it to trigger a file download.
Somehow firefox reloads some page aspects afterwards and closes websocket connections for example.

Since the chrome browser does not have this behavior, we do not really know what is happening and it seems a browser specific problem.

We are solving it by always tell the jshowurl component to open the URL in a new window -> _blank as target.
Finally!! ;)
Hello,

(short answer: read *4*)

we faced a similar problem in the RISC-Client. We refactored the download behavior to be more user friendly in Browser-Environments.

The files are generated in sub threads and saved to the disk. Actually files, not byte-arrays in memory. After the thread finishes it writes its status into an queue (Queue<DownloadInfo>) inside the managing UI-class (DownloadManagerUI).
The DownloadManagerUI will be triggered to do a roundtrip (via longpolling) and in the longpolling action we trigger the download of the next file of the queue. Since we have an actual file and we are in an RISC-Environment, we use the t:jshowurl component. To save the file we use the CC-FileManager.
Is the queue not yet empty, we call another roundtrip (not via longpolling, but a additional t:callserverontrigger) and repeat the step above until the queue is empty.

More explanation:
*1*
We used two different triggers, longpolling and callserverontrigger, to synchronize the Main-UI-Thread and the Sub-Threads who will generate the files. The Sub-Threads are triggering the longpolling (wakeup). If the Sub-Threads are quicker then the roundtrip (two wakeup-calls inside a single roundtrip) we may lose information to trigger the download. Thats why we write into a queue and need the t:callserverontrigger to watch inside the queue if we maybe missed something since the last roundtrip.

*2*
You will likey need to remove the generated files manually from the (server-) disk. Or use the CC-TempFileManager (org.eclnt.jsfserver.util.tempfile.TempFileManager).

*3*
In our case the DownloadManagerUI has a list of download files. And a button to re-download. And you may abort the running threads if you trigger a lot of downloads. So it is a nice little usability tool...

*4*
Before the RISC-Environment (Java-Client, like in your scenario) we generated the PDF inside the Main-UI-Thread, triggered the download and used the feature "withcallback" of the t:download component to trigger the next download.

Actually you can try to use the "withcallback" function in your scenario. In a RISC-Environment this option ist not "safe". Browsers have different behavior on this matter.

greetings
Ruven

Old thread but new information:

we face the same problem. The tiny difference, why the combofield popup is NOT closing, is the attribute "keepfocus".
Since the popup (opened by the combofield action) never gets the focus, it cannot lose it (by moving the outer window/popup) to be closed.

Any chance to change this behavior on SWING-Client?

greetings
Ruven
Hey,

SWING-Client!

We are using a Combofield inside a grid. The Combofield opens a popup containing a <t:textwithlinks>. The text which is displayed inside looks like this:

Code:
<html height='100%'><body height='100%'>asd</body></html>


Now inside the opened popup with the textwithlinks component I right click on the text and use the buildin feature to copy the text. See attached Screenshot.

This text has a linebreak (linefeed). Although the String itself has no linefeed...

Bug in CaptainCasa? Swing?
Hey,

strange behavior: giving a t:radiobutton a clientname, will not render the data-riscclientname attribute, but remove the CSS of the radiobutton.

Without clientname all fine, with its messy.
Seems like a litte bug in javascript processing?

In the attached picture the top radio button is without clientname. The two next ones have a clientname. One is checked/marked.

greetz
Ruven
Hi,

the actual question is: Is there a lazy loading for SUBMENUs.
For Level 2 so to say.

Feature-Request: MENUNode disable. Disable a SUBMENU as a whole. Right now I have to render it as a MENUNODEItem and disable it to indicate there is something, but is is disabled.
Or am I missing something?
The changelog PDF looks strange.
Some characters are ... wrong? Seems like PDF generation did some funny stuff...

e.g.
Code:
Improvement 7 Client perBormance increase due to switcginF oBB loFFinF


And empty pages. Maybe Upload not complete?
Am I too fast? ;)

greetz
Ruven
Within the same scenario, but maybe not related directly.

The RISC-Cllient throws the following errors when working (clicking) inside the combofield:

Code:
 risc.js?version=20180924:1 Uncaught RangeError: Maximum call stack size exceeded
     at RISCField.RISCComponent.removeBackgroundModifier (risc.js?version=20180924:1)
     at RISCComboField.RISCComponent.removeBackgroundModifier (risc.js?version=20180924:1)
     at RISCField.RISCComponent.removeBackgroundModifier (risc.js?version=20180924:1)
     at RISCComboField.RISCComponent.removeBackgroundModifier (risc.js?version=20180924:1)
     at RISCField.RISCComponent.removeBackgroundModifier (risc.js?version=20180924:1)
     at RISCComboField.RISCComponent.removeBackgroundModifier (risc.js?version=20180924:1)
     at RISCField.RISCComponent.removeBackgroundModifier (risc.js?version=20180924:1)
     at RISCComboField.RISCComponent.removeBackgroundModifier (risc.js?version=20180924:1)
     at RISCField.RISCComponent.removeBackgroundModifier (risc.js?version=20180924:1)
     at RISCComboField.RISCComponent.removeBackgroundModifier (risc.js?version=20180924:1)
 
Hi,

the icon for the combofield does not trigger any action in the RISC-Client.
A doubleclick does. In this particular example we have the invokeevent set to doubleclick on top of the icon itself.

Swing-Client works fine.
The code for the combofield. It is used inside a PBC.
Code:
 <t:combofield id="g_5" actionListener="#{TextblockUIC.onCombofieldAction}" activationhotkey="40" bgpaint="#{TextblockUIC.bgPaint}" border="0" flush="true" flushtimer="800" image="/images/fugue/block.png" invokeevent="doubleclick" keepfocus="true" requestfocus="#{TextblockUIC.focus}" selectallwhenfocussed="false" text="#{TextblockUIC.contentText}" width="100%" />
 


greetz
Ruven
Hi,

maven and sources repo is not up to date yet.

greetz
Ruven
In the RISC-Client the icon to quickly set the divider location to "0" is not present.

In the demo workplace it is described in "Containers -> Split Pane - Extended Functions -> Quick Navigation".
There are just "normal" splitpanes.
 
Profile for rspiller -> Messages posted by rspiller [59] Go to Page: 1, 2, 3, 4 Next 
Go to:   
Powered by JForum 2.1.6 © JForum Team