Uploaded image for project: 'ZK'
  1. ZK
  2. ZK-4277

Negative values in NumberInputWidgets (JDK >=9)

XMLWordPrintable

    • ZK 8.6.2 S1, ZK 8.6.2 S2
    • None

      Steps to Reproduce

      	<vlayout>
      		No locale: <intbox value="-123"/>
      		en_US: <intbox locale="en_US" value="-123"/>
      		sv_SE: <intbox locale="sv_SE" value="-123"/>
      	</vlayout>
      

      Try typing a minus character into the int boxes.

      Current Result

      For the intbox with swedish locale the MINUS key (numpad and normal keyboard) is ignored

      Expected Result

      Possible to type negative numbers
      Allow using the minus key on a keyboard to type in negative numbers.

      Debug Info

      Looking closely at the swedish Intbox you'll see the MINUS character is not a HYPHEN-MINUS (code 0x2d) but a unicode MINUS (code \u2212)

      Root Cause

      Since JDK 9 the default locale DB was changed to CLDR (http://openjdk.java.net/jeps/252).
      Which defines the unicode MINUS as the minusSign for various locales (including swedish)
      https://www.unicode.org/cldr/charts/33/by_type/numbers.symbols.html#2f08b5ebf85e1e8b
      (most locales still use the HYPHEN-MINUS code 0x2d)

      Workaround

      1. https://bugs.openjdk.java.net/browse/JDK-8214926 mentions a possibility to change the default locale provider to be compatible with JDK8 by specifying a JVM startup argument:
        -Djava.locale.providers=COMPAT,CLDR
        
      2. alternative in ZK: In order to allow both characters override the getAllowedKeys_ function for NumberInputWidgets
      <script><![CDATA[
      /*Patch for ZK-4277*/
      zk.afterLoad('zul.inp', function() {
      	var xNumberInputWidget = {};
      	zk.override(zul.inp.NumberInputWidget.prototype, xNumberInputWidget, {
      		getAllowedKeys_ : function() {
      			return xNumberInputWidget.getAllowedKeys_.apply(this, arguments) + '-';
      		}
      	});//zk.override
      });//zk.afterLoad
      ]]></script>
      

      This happens to work, as the number parser will replace the localized MINUS character with a HYPHEN-MINUS anyway, so that allowing to type it doesn't change the parsed result.

            DevChu DevChu
            cor3000 cor3000
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: