Jesse Costello-Good

There is a logical problem in the structure of the code. I'm not sure it can be solved simply. The problem is only the edit mask (DatePicker) contains the logic to parse all acceptable formats (epoch seconds, new Date(), and DateFormat.parse()). The matrix column / message format used to format the data does not share the same parsing logic. Although, it's unclear to me whether it makes sense logically to use the message format to parse.

The following change in Matrix.js seems to resolve the issue but I'm hesitant to commit it because it seems like it could have side effects and I'm not sure whether it is logically correct.

  MessageFormat_prototype.format = function(objDiv, strCDFKey, objMatrix, objMatrixColumn, intRowNumber, objServer) {
    //var objNode = objMatrix.getRecordNode(strCDFKey);
    var strValue = objMatrixColumn.getValueForRecord(strCDFKey);

    if (strValue) {
      try {
        var loc = objServer.getLocale(); // setLocale should publish locale change event.
        if (loc != this._format._locale) // Need to pick up new locale if it changed since this instance was created.
          this._format.setLocale(loc);

        try { strValue = this._format.format(strValue); } catch (e) {}

        objDiv.innerHTML = strValue;
      } catch (e) {
        jsx3.util.Logger.GLOBAL.error(this._format, jsx3.NativeError.wrap(e));
      }
    }
  };

Jesse Costello-Good

committed 949 to GI

13 Apr 11

refs GI-931: from Jim Mann, onload attribute of <script> conflicting with onload setting of config.xml

modified JSX30.js

trunk/gi/JSX/js/JSX30.js (+1 -1)
- Diff
- View
- File History
- Download

Jesse Costello-Good

closed GI-937

12 Apr 11

Jesse Costello-Good

commented on GI-937

12 Apr 11

My hunch is that the following works in IE but not Fx: new Date("01-01-2011"). If it works in IE it's by accident rather than by JavaScript spec. GI's matrix format handler will handle either epoch seconds (as an integer) or a string value that can be parsed by the Date constructor. Here's the code from DateFormat:

  DateFormat_prototype.format = function(objDate) {
    if (!(objDate instanceof Date)) {
      if (!isNaN(objDate)) {
        objDate = new Date(Number(objDate));
      } else {
        objDate = new Date(objDate);
      }

      if (isNaN(objDate)) throw new jsx3.IllegalArgumentException("objDate", objDate);
    }

Jesse Costello-Good

resolved GI-931, GI-936

06 Apr 11

Jesse Costello-Good

changed the status to Open of GI-936

06 Apr 11

Jesse Costello-Good

updated 5 fields of GI-936

06 Apr 11

I guess ImageAlpha CSS thingy doesn't work with IE8 and a "?" in the URL of the page. Way to go IE! So, I need to go back to the old way of determining the browser version.

Jesse Costello-Good

committed 948 to GI

06 Apr 11

refs GI-931: IE9

modified JSX30.js

trunk/gi/JSX/js/JSX30.js (+5 -3)
- Diff
- View
- File History
- Download

Jesse Costello-Good

committed 947 to GI

06 Apr 11

refs GI-931: IE9

modified JSX30.js

trunk/gi/JSX/js/JSX30.js (+1 -2)
- Diff
- View
- File History
- Download

Jesse Costello-Good

changed the status to Needs Info of GI-936

06 Apr 11

Please attach a screen shot of it working and not working.

Jesse Costello-Good

commented on GI-936

06 Apr 11

Ed, is it just a resolved path issue? Can you use Firebug or similar to see whether the URL to the background images is being resolved correctly? When I hear a problem only shows up running outside of Builder it makes me think of URL resolution issues. The relative paths between the deployment page, JSX, and app are usually different when you deploy an app.

FishEye and Crucible

About

Reviews

0 active reviews

Line History

Commit Activity

52 week commits volume

Commits by day

Commits by hour

Commit calendar

Committer Mappings

Jesse Costello-Good is mapped to 1 committer over 1 repository