Custom tile urls - hotspot bug

  • When using custom-tile-urls, hotspots that are smaller than some critical size are no longer active.
    Two similar hotspots are presented in the example. The difference is only in the value of the scale.
    A spot with a larger size is pressed, while a smaller one is not pressed for some reason.
    https://sferitus.com/?xml=osmmap/maps.xml

    This hotspots does not work:

    Code
    <hotspot name="dis" onhover="showtext(test,STYLE4);"
    onclick="onclickB();"
    onclickA="set(hotspot[dis].alpha,1); set (onclick, onclickB() );"
    onclickB="set(hotspot[dis].alpha,0.5); set(onclick, onclickA() );"
    	url="%HTMLPATH%/drone_map512.png" width="2048" height="2048"
    	scale="0.0000001" zoom="true" capture="false"
    	ath="0.0658058" atv="-0.1701285" />

    This hotspots works:

    Code
    <hotspot name="en" onhover="showtext(test,STYLE4);"
    onclick="onclickB();"
    onclickA="set(hotspot[en].alpha,1); set (onclick, onclickB() );"
    onclickB="set(hotspot[en].alpha,0.5); set(onclick, onclickA() );"
    	url="%HTMLPATH%/drone_map512.png" width="2048" height="2048"
    	scale="0.00000025" zoom="true" capture="false"
    	ath="0.0658500" atv="-0.1701285" />
  • in your example you use a 512px image,
    first scale it up to 2048px,
    and then scale it down again with 0.0000001 / 0.00000025

    these are crazy! *blink* scale values

    i avoid to use scale values under 0.5 or over 2.0 *smile*

    something is odd with how you implement that map,
    it should not be necessary to use such scaling

  • Hi,

    thanks for reporting!

    That's numerical bug in the internal hotspot-hittesting code, at very small scales a value in the calculation gets very near to zero and a check for zero there is a bit inaccurate...

    That will be fixed soon in the next release (1.20.6).

    As workaround for the moment you could try using renderer="css3d" for the hotspots. That a different hotspot rendering and hit-testing will be used.


    Zitat

    these are crazy! *blink* scale values

    That's a very large 'pano' (134.217.728 x 134.217.728px ;-)), therefore these scales, and normally (when there are no unexpected bugs) they shouldn't causes problems and are fully valid.

    Best regards,
    Klaus

Jetzt mitmachen!

Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!