krpano 1.22 - Reactive APIs, ThreeJS 3D Plugin, krpano Maps, Improved Javascript APIs, Custom Encryptions

klaus.krpano · November 19, 2024 at 4:01 PM Official Post

Quote from kme

In the 1.22 example, hittest does not work on the object itself, but it does register a hit when one moves the mouse to the left/under of the motorcycle.

Thanks for this example!

This is a bug - internally the hittest-direction-vector wasn't initialized for screen-raycasts when no hotspot had the hittest setting enabled. The example works correctly when adding hittest=true to the threejs hotspot.
This will be fixed in the 1.22.2 release.

Quote from kme

To make the motorcycles appear in the EXACT same location, I had to map the origin values of krpano obj to tx/ty/tz for threejs object.

In your example the obj and gltf files are differently centered.
E.g. try using the .obj file instead of the .gltf for the threejs hotspot.

About mapping - instead of using origin, set the image.ox/oy/oz to the same values as the hotspot tx/ty/tz settings. Then they would be at the same position.

Best regards,
Klaus

kme · November 19, 2024 at 4:06 PM

Thanks, just tested with hittest="true" for the threejs object and - as you said - it works properly in that case.

Also thanks for the hint for differences in the .obj vs the .gltf model, something might have shifted indeed when converting it from one format to another.

I could not use the .obj as it is encrypted, and threejs does not support the encrypted .obj like krpano does.

klaus.krpano · November 20, 2024 at 8:33 PM Official Post

Hi,

another bugfix release: krpano 1.22.2

New: The textfield template supports now also linking to events (and using an expression for the resulting content)
New: The ThreeJS Sun Light supports now an optional custom center position for the shadowmap (atview=false). Additionally there is new shadowmaphelper setting to visualize also the position and range of the shadowmap. And the Sun Light helper is now dragable.
Fix: Static defined linking to attributes of the same element was not working (report).
Fix: Adding a value to an existing krpano array-item, that has the the index 0, wrongly created an new auto-named array-item (report).
Fix: Cubemap panos with objectsize=0 weren't displayed.
Fix: Panos with a 3D-model as depthmap and rendermode="depthmap" had a wrong origin.
Fix: screen-based raycasts were not working correctly with external renderers (e.g. ThreeJS) when no hotspot had the hittest setting enabled (report).
Fix: The <if> xml element didn't work inside non-root-level-elements.
Fix: Depthmap hittestmode="gpu" was not working.
Fix: Using depthmap-hittesting and hotspot-hittesting the same time was not working.
Fix: drag3d.xml - support 3D dragging also when using view offsets.
Fix: Enable the VR position tracking also in the ThreeJS Environment Maps, Tuscany and Open Source Plugin examples (report).
Fix: Spherical images for the ThreeJS environment texture were using the wrong color-space (causing washed-out colors).
Fix: The leveling tool in the krpano Tools app was not working (report).
Fix: <a href="..."> links inside textfields were not working with touch controls.
Fix: Some typos and missing links in the documentation.

Additionally, the ThreeJS Tesla example has been updated:

It supports VR now - if you have a Meta Quest 3, have look , the proportions should look quite realistic.
There are multiple different pano-environments to test different colors and lightnings.
And the materials of the car 3D-model can be interactively changed.

Best regards,
Klaus

Tuur · November 20, 2024 at 11:22 PM

ohhh the new 'Tesla' example is really great!!

Tuur

kme · November 21, 2024 at 8:55 AM

Quote from klaus.krpano

Additionally, the ThreeJS Tesla example has been updated:
It supports VR now - if you have a Meta Quest 3, have look , the proportions should look quite realistic.

Just FYI - Tested with Quest 2, and although it looks good, device can barely keep up with head movement. When looking around away from the car (no threejs objects visible), the scene performs well. But as soon as the car becomes in view, turning the head creates temporary black borders which breaks the VR experience. Maybe on Quest 3 performance is just enough, but I have to test when I get my hands on that device.

klaus.krpano · November 21, 2024 at 9:52 AM Official Post

Hi, right, the Quest 2 is too slow for complex ThreeJS content - the car 3D-model is big and complex and the ThreeJS fragment shaders are also not the fastest ones...

It's not perfect on the Quest 3 either, but it's quite usable there.

Although it should be possible to optimize the example for performance, e.g. a less complex 3D-model, smaller textures and lightmaps, simpler and faster shaders, but then also the visual quality would be a bit more worse.

PanoramicWEB · November 30, 2024 at 2:40 PM

Hi , One of the bugs I noticed, I prepared an example for it.

krpano

Code

<events keep="true"  
ondoubleclick=" zoomfovpositionfunction();"
onsingleclick=""
onmousedown=""
onkeyup=""
onup=""
 />

	 <hotspot name="tesths"  onclick=""  url="arrowhead.png" distorted="true" alpha="1"  width="250" capture="false" handcursor="true" enabled="true" height="250" atv="0" keep="true"   />
 
	
<action name="zoomfovpositionfunction">
screentosphere(mouse.x, mouse.y, mouseath, mouseatv);
adjusthlookat(mouseath);
if(zoomfovposition === true ,  tween(view.fov , 87); tween(view.hlookat , get(mouseath)); tween(view.vlookat , get(mouseatv)); set(zoomfovposition,false);
,  tween(view.fov , 30) ; tween(view.hlookat , get(mouseath)); tween(view.vlookat , get(mouseatv)); set(zoomfovposition,true); );


</action>

Display More

If I set capture='false' on the hotspot, the ondoubleclick event is triggered when I click on the hotspot. However, if I set capture='true', it is not triggered."

jeromebg · November 30, 2024 at 2:45 PM

Quote from PanoramicWEB

If I set capture='false' on the hotspot, the ondoubleclick event is triggered when I click on the hotspot. However, if I set capture='true', it is not triggered."

Already reported here :

Thread

Trigger ondoubleclick event

When layer is clicked, if capture is set to false, even a single click will be considered a double-click. I think it's logical not to trigger the ondoubleclick event.

zhitongjia

November 25, 2024 at 1:37 PM

klaus.krpano · December 19, 2024 at 1:20 PM New Official Post

Hi,

there is a new 1.22 release: krpano 1.22.3

It provides several fixes and also some compatibility improvements:

New: scalemode="sizescale" - a new optional layer scale mode, if used then additionally also the width/height changes of images are used as scale-factor for children layers.
New: krpano Maps - the pantospot() action has now an optional zoomlevel parameter.
Change: To improve compatibility with older krpano or Panotour Pro examples, the default setting of scalemode (for old xml files) is changed from "resize" to "sizescale".
Fix: Improved Panotour Pro compatibility, if in the showlog() action the first parameter is an empty string, then this is evaluated as true, which means the log will be opened.
Fix: For improved compatibility with older examples, support for the 'alturl' (alternative url for html5) setting was added again.
Fix: A layer with capture=false was wrongly triggering a global doubleclick event on a single-click, and the onsingleclick/ondoubleclicks events on the layer were not working (report).
Fix: It's now possible to optionally disable the mouse-wheel capturing in html, iframe and krpano elements. Therefore the capturewheel setting is true by default now for these elements. But if set to false, the elements will pass-through the mouse-wheel events.
Fix: Support scalemode="scale" now also for html, iframe and krpano elements (report).
Fix: Hotspots with type html, iframe or krpano were missing after returning from VR mode.
Fix: The sameorigin setting wrongly affected also layer/hotspot images (report).
Fix: Improved error reporting for layer/hotspot loading errors. If known, the reason for the loading error is also given.
Fix: WebVR - On Android devices, the new userAgentData Browser API will be used to get the real device-name. That name is necessary to look for the correct physical screen-size in the VR-device-databases.
Fix: The setting bgbordermode="inside" didn't work on hotspots (report).
Fix: The textfield textalign was wrong when using bgbordermode="inside".
Fix: Several layout bugs with flowchildren and missing updates were fixed (report).
Fix: When a layer was added to another layer (using the parent setting) and that layer was not ready/created yet, that layer could be missing (not displayed)
Fix: Layout performance improvements when using changing textfields and flowing layers.
Fix: The text alignment could be wrong when using ellipsis.
Fix: A wait(load) call in the first pano load could wrongly cause an infinity waiting.
Fix: The right mouse button tracking was not working correctly on Mac (report).
Fix: Use a time-and-random-based unique hotspot name in the krpano Tools VTour Editor when creating new hotspots to avoid potential naming conflicts (report).
Fix: In some cases html-links inside disabled (enabled=false) textfields were still clickable.
Fix: krpanomaps.xml and maps2 example - use the shortest 360 wraparound path to the target when panning on the map.
Fix: showtext.xml - the text position was wrong when using a custom area or display.safearea.
Fix: Add the missing documentation for the wildcmp(), spliturl() and buildurl() Javascript helper functions.

Additionally, I wish everyone a Merry Christmas and Happy Holidays!

Thanks and Best regards,
Klaus

Fernando · December 19, 2024 at 1:32 PM New

Merry Christmas to you too Klaus. Thank you so much for making our dreams come true!!

Of course, my congratulations also go to all the kind friends of the group, without whom many of our projects would not be possible.

Tuur · December 19, 2024 at 1:40 PM New

Thank you very much Klaus for all your efforts and time!

And i wish everybody Merry Christmas, peace and enjoyable holidays

Tuur

indexofrefraction · December 19, 2024 at 2:50 PM New

Santa Klaus was here !!!

Code

  .-""-.
 /,..___\
() {_____}
  (/-@-@-\)
  {`-=^=-'}
  {  `-'  } 
   {     }
    `---'

Thank you for all the work and

Merry Christmas to you and yr Family!

luo jian chao · December 20, 2024 at 9:43 AM New

klaus.krpano · December 20, 2024 at 3:46 PM New Official Post

Hi, there is a new 1.22.3 release online (build 2024-12-20).
It fixes a minor problem that was introduced in the previous release. If a parent layer uses flowchildrenexpand=true together with a children layer that uses flowoverflow=true and has overflowing layers, the resulting layout was wrong.

godrifle · December 26, 2024 at 10:20 PM New

Congrats on the release. Incredible work as always. Quick question: Seems we're still bound by Rosetta on a Mac. Is a native Apple Silicon version in the roadmap?

klaus.krpano · December 26, 2024 at 10:24 PM New Official Post

Quote from godrifle

Is a native Apple Silicon version in the roadmap?

Yes

ctt360 · December 30, 2024 at 11:09 AM New

Hi Klaus,

there was an issue with the right mouse button click on Mac, which was fixed in 1.22.3 (report). It still doesn't work correctly for me on Linux (tested with Chromium 131.0.6778.139 and Firefox 133.0.3 on Ubuntu 22.04). "mouse.button" returns 0 and "mouse.rightbutton" returns false on right click. It does work as expected on Windows 11 for me.

Kind regards,
Christian

indexofrefraction · December 31, 2024 at 1:52 PM New

Klaus

I sent you a barebone with a possible 1.22 set3dtransiton issue (bug?) by email

klaus.krpano · January 1, 2025 at 10:49 AM New Official Post

Quote from indexofrefraction

I sent you a barebone with a possible 1.22 set3dtransiton issue (bug?) by email

Thanks - that should be the same issue as here:

Post

RE: sluggish loading of tiles & "shadow" artifacts

Hi,

I think the problem has been found: There was a texture caching issue (only for 3D model cubemaps) where a texture was reused but with the wrong mipmap filter state, and this caused mipmap levels to be temporarily used with incorrect/old texture data.

As workaround try adding this line anywhere in your xml:

(Code, 1 line)

In the next release 1.22.4 this bug will be fixed.

Best regards,
Klaus

klaus.krpano

January 1, 2025 at 10:19 AM

klaus.krpano · January 1, 2025 at 10:33 PM New Official Post

Quote from ctt360

there was an issue with the right mouse button click on Mac, which was fixed in 1.22.3 (report). It still doesn't work correctly for me on Linux (tested with Chromium 131.0.6778.139 and Firefox 133.0.3 on Ubuntu 22.04). "mouse.button" returns 0 and "mouse.rightbutton" returns false on right click. It does work as expected on Windows 11 for me.

Sorry, I can't reproduce that... did the same Linux tests now and in my case all mouse button values are correct...

krpano 1.22 - Reactive APIs, ThreeJS 3D Plugin, krpano Maps, Improved Javascript APIs, Custom Encryptions

Trigger ondoubleclick event

RE: sluggish loading of tiles & "shadow" artifacts

Participate now!

Users Viewing This Thread